The Model Laid Bare: A Digest of Posts 31 Through 40

This is the fourth in a series of digests, each covering ten posts from this blog, in chronological order. The third digest covered posts 21 through 30, ending on 15 January 2011. This batch spans the remainder of January through to late March 2011 – a period in which the Marshall Model stops being a background presence and steps fully into the foreground. For the first time, the blog devotes sustained attention to explaining the model’s origins, its inner mechanics, and the personal conviction behind it. The voice, already assured, becomes something more: it becomes declarative.


31. Employees’ Self-confidence – A Cinderella of Organisational Effectiveness (19 January 2011)

Spotlighting a 1994 paper that, as I noted, remained alien to most Analytic-minded organisations even at the time of writing, this post made the case for employee self-confidence as a neglected lever of organisational performance. Crucially, the note to self appended to the post – that this principle applies to peers as much as to subordinates – gave it a reach beyond the purely managerial. The Analytic mindset treats confidence as a personal trait, largely irrelevant to the work. The Synergistic mindset understands it as a systemic condition that can be cultivated or destroyed by the environment in which people work.

32. There’s a Lot of it About (Bullshit) (20 January 2011)

Brief in its own commentary but pointed in its selection, this post amplified an argument about the urgent need for institutional bullshit detection. The observation – that the internet is simultaneously the greatest disseminator and the greatest destroyer of nonsense – felt particularly germane to the worlds of IT, software development, and consulting, where grand claims and borrowed frameworks multiply unchecked. For a blog that had already sharpened its critique of Agile orthodoxy and management fashions, this was a small but apt statement of epistemological values.

33. A Great Example of the Synergistic Mindset in Practice (22 January 2011)

One of the recurring challenges of the Rightshifting project has been making the Synergistic mindset tangible – not an abstraction, but a way of running a real organisation that real people have actually chosen and that has actually worked. This post amplified a first-hand account from the Management Innovation Exchange of one such organisation in its early Synergistic years, and the sheer level of disengagement the author had witnessed in hierarchical organisations before making the leap. A valuable data point, even if – as I noted – I would love to know how things had been going since.

34. Contextualising FlowChain (22 January 2011)

Posted the same day as the preceding entry, this short but significant piece gave my colleague and co-conspirator Grant Rule his first real platform on the blog. Rule’s argument – drawn from his work at SMS Exemplar – was that Agile methods, for all their value, remained fundamentally constrained by their inheritance of the project model: creating and disbanding teams, establishing and abandoning value streams, discarding accumulated know-how at every opportunity. FlowChain, by contrast, pointed toward something closer to genuine continuous flow. The piece served as both an endorsement and a contextualisation, embedding FlowChain in the broader Rightshifting narrative for the first time.

35. The Many Roles in Software Projects (25 February 2011)

The sole post of February, and a deliberately practical one. Prompted by Capers Jones’ observation that many software projects involve over fifty distinct roles, I published a comprehensive catalogue of the roles required in any serious software development effort – from Interaction Designer and Chief Engineer through to Concierge and Team Psychotherapist. The list, originally inspired by Kent Beck’s XP Explained, was intended not as a job-description generator but as a corrective to the Analytic habit of underestimating complexity by treating “the developer” as a monolithic resource. Teams are ecosystems, not headcounts.

36. Rightshifting Transitions (Part 1): Ad-hoc to Analytic (2 March 2011)

The first of a three-part series that finally gave the Marshall Model’s transition zones the extended treatment they deserved. Part 1 examined the journey from the Ad-hoc to the Analytic mindset: the point at which a growing business, overwhelmed by its own informal chaos, begins to impose structure, hierarchy, and accountability. The transition is rarely recognised as such, is traumatic for those accustomed to doing things their own way, and leaves in its wake – if successful – an organisation that is marginally more effective but, as I noted, almost always less human. The retrospective lesson of this transition: the value of discipline.

37. Rightshifting Transitions (Part 3): Synergistic to Chaordic (3 March 2011)

Published a day before Part 2 – an inversion that, in hindsight, rather suits a discussion of Chaordic thinking – this instalment described the rarest and most demanding of the three transitions. The Synergistic organisation that has found its purpose and embedded collaborative self-organisation now faces an even more radical letting-go: the willingness to reconfigure itself almost continuously in pursuit of fleeting commercial opportunities. The Chaordic business treats structural instability not as a crisis but as a natural condition. The retrospective lesson here: the value of positive opportunism. Few businesses ever reach this point; fewer still sustain it.

38. Rightshifting Transitions (Part 2): Analytic to Synergistic (6 March 2011)

The pivot of the series, and the transition most directly relevant to the organisations the blog had been engaging with throughout. The Analytic-to-Synergistic transition requires abandoning not discipline itself, but its traditional manifestations: hierarchy, extrinsic motivation, functional silos, targets, projects, appraisals, coercion. For people who have known only command-driven environments, this can be as disorienting as any earlier transition. But those businesses that make it successfully find a world in which decisions are made collaboratively, value streams replace org charts, and a shared sense of purpose binds everyone. The retrospective lesson: the value of purpose. Simon Sinek’s Golden Circle was cited as a useful companion text.

39. The Rightshifting Ethos (20 March 2011)

The most personal and avowedly values-driven post in this batch. Beginning from a Seth Godin provocation – “say what you believe, and see who follows” – I stated plainly what I believed Rightshifting could mean for individuals and for society: improved health, through reduced job-related stress and greater engagement; improved wealth, through more commercially successful and more equitable organisations; and improved wisdom, through the abandonment of a century of Taylorist management assumptions that had long outlived their usefulness. This was not analysis but declaration. The post closed with an open invitation to join the movement.

40. The Origins of the Marshall Model (25 March 2011)

The most autobiographical post the blog had yet published, and one of the most important in the entire archive. In it, I traced the model’s creation through five “happy accidents”: a late submission to Agile North 2008 that led me to Steve McConnell’s asymmetric bell curve; the audience response that prompted me to categorise the different organisational modes; Dee Hock’s concept of the Chaordic organisation claiming the blank right-hand space on the chart; Carol Dweck’s work on mindset reframing the categories from modes to collective worldviews; and finally the recognition that the overlapping zones between categories were not mere statistical artefacts but the most consequential feature of the whole model – the sites of transition, where the real work of organisational change takes place. The piece concluded with the argument that had been implicit throughout: most organisations remain stuck near the median not because improvement is impossible, but because crossing a transition zone is an order of magnitude harder than incremental improvement, and most organisations never even realise they are standing in one.


The Model Steps Forward

If the previous batch saw the blog beginning to address different audiences, this one sees it doing something more ambitious: taking the model apart, showing its workings, and reassembling it in public. The transitions series is the most sustained piece of explanatory writing the blog had yet attempted, and the origins post is something rarer still – a moment of genuine intellectual autobiography, in which the author traces not just what the model says but how it came to exist.

Two other notes worth making. First, the Rightshifting Ethos post marks the first time the blog had made an explicitly moral argument for the work – health, wealth, wisdom are not technical claims but human ones. The critique of Agile and the Analytic organisation had always had this moral dimension, but here it was brought fully into the open. Second, the contextualisation of FlowChain alongside the transitions series signals that the blog is no longer just diagnosing the problem; it is beginning, cautiously, to describe what the alternative looks like in practice.

In the next digest, I’ll cover posts 41 through 50, picking up from the Improvement ROI Sawtooth (March 2011) and taking us into the spring and early summer of 2011 – a period of relative quiet in terms of volume, but one in which several of the blog’s most enduring ideas receive their most careful elaboration.

– Bob

From Critique to Framework: A Digest of Posts 21 Through 30

This is the third in a series of digests, each covering ten posts in chronological order. The second digest covered posts 11 through 20, ending in September 2010. This batch spans November 2010 to January 2011 – a period in which the blog’s voice grew more assured, the critique of Agile sharpened into something closer to a full diagnosis, and the Marshall Model’s vocabulary began to crystallise.

21. The State of Agile (1 November 2010)

One of the most substantial posts from this period, and arguably the clearest early articulation of the elephant in the room. I argued that Agile at its most effective had the hallmarks of a social science or organisational psychology phenomenon, and not much at all to do with technical practices. From this perspective, the state of Agile was looking vulnerable – poised less at a tipping point than on a knife-edge. Many adopting organisations reported positive early results, but my work with Rightshifting and the Marshall Model had thrown up a key question: adopting Agile necessitated a new mindset for practitioners, yet most host organisations had a prevailing mindset fundamentally at odds with that requirement.

22. Delivering Software is Easy (13 November 2010)

A post that refused the comfort of ‘I’m all right, Jack’. Yes, there were thousands of folks who could reliably deliver quality software – typically in small shops or cosy enclaves. But there were hundreds of thousands who couldn’t, not because they lacked skill, but because of where they worked, who they worked for, and all the monkey-wrenches and bear traps lurking in their daily routines. And millions more depended on the latter. I introduced the concept of the ‘aspiration gap’ – the gulf between where the majority of software development jobs sit and where developers would actually like to work.

23. Rightshifting and the Senior Management Pitch (21 November 2010)

A practical post responding to folks who said Rightshifting struck a chord with them personally, but that their senior management was the audience that really needed to hear the message. Drawing on Paul DiModica’s ‘Three Box Monty’ presentation technique from Value Forward Selling, I described how to position the presenter as a peer of the decision-makers. I then proposed two variants: a Solutions Focus version (replacing problems with what leading businesses do well) and an Outcomes Focus version (replacing solutions with positive outcomes already seen and desired for the future).

24. Would You Rather Not Know? (5 December 2010)

Some people had been asking me what the point was of knowing about what makes organisations more or less effective, given that many developers and even middle managers felt powerless to change anything. The Marshall Model proposed that an organisation’s collective mindset directly dictated its effectiveness – and that significant improvement required a change in how the organisation looked at the world of work. So was awareness merely a recipe for frustration and learned helplessness? I argued the opposite. Drawing on Carol Dweck’s research into mindset, I suggested that simple awareness of the role mindset plays could help people begin to adjust how they see the world. Ignorance was never the more productive option.

25. #lru2010 Conference Report (18 December 2010)

A report on the first London Rightshifting Unconference, held at London City University on 17 December 2010. Ten attendees – including CTOs, consultants and coaches – spent the afternoon exploring Rightshifting ideas in practice. The sessions included a run-through of core Rightshifting concepts, experience reports from Ant Clay of 21Apps (who had presented the Rightshifting curve to clients that very week, with strikingly positive reception) and Grant Rule of SMS Exemplar, plus a session from Liz Keogh on the Dreyfus Model of Skills Acquisition. The event marked a significant milestone: the Rightshifting ideas were no longer just a blog and a Twitter hashtag – they had become a community gathering in person.

26. Agile/Lean Principles Simply Will Not Scale (29 December 2010)

Amplifying a provocatively titled LinkedIn post, I noted that its argument was broadly congruent with my earlier ‘State of Agile’ piece. The core observation: Agile and Lean development could scale in small, medium and large companies, but they typically wouldn’t. Politics – specifically, the threat that Agile posed to people in unnecessary positions – was the root cause. Agile exposes waste, and if your job is pointless, you’re in trouble. So those whose positions were threatened worked to ensure management didn’t buy in.

27. To Deliver the Project at Hand, or Improved Project-delivering Capability for the Future? (30 December 2010)

Responding to Dadi Ingolfsson’s tweet, I unpacked a tricksy can of worms: should Scrum Masters and Agile coaches be more vested in the success of the current project, or in Agile adoption across the organisation? The core tension was between delivering the project at hand and building improved project-delivering capability for the future. A coach who prioritised the latter over the former risked losing the team’s trust. Yet senior management often expected capability improvement as part of the remit. In practice, a delicate balancing act was the order of the day – and one reason why good coaches could really earn their coin.

28. The Term ‘Analytic Mindset’ Defined (6 January 2011)

Given my frequent use of terms like ‘Analytic thinking’ and ‘Analytic mindset’ in the context of Rightshifting, I paused to define them properly. This was the first formal articulation on the blog of the Analytic mindset as a specific, coherent worldview – characterised by reductionism, command-and-control, local optimisation, and Theory X beliefs about human nature. It marked the moment the Marshall Model’s vocabulary began to be laid down with some precision, rather than assumed.

29. The Constant Tension Between Rightshift and Left-drift (7 January 2011)

Drawing on a blog post by Jamie Flinchbaugh, I highlighted a frequently overlooked reality: efforts at Rightshifting are always swimming against the tide of entropy. Changes within and without the organisation constantly conspire to erode its effectiveness. Most organisations, even those investing heavily in improvement programmes, rarely managed to do much more than tread water. The pace of improvement had to outrun the pace of entropy – a sobering thought for anyone who assumed that gains, once made, would persist on their own.

30. Traditional vs ‘New’ Management Thinking (15 January 2011)

Referencing a post from the Deming Learning Network, I drew parallels between their comparison of traditional and new management thinking and what I called the Analytic-to-Synergistic transition. The post served as a bridge between Deming’s body of work and the Marshall Model, showing that the shift I was describing wasn’t novel – it had been articulated for decades – but that framing it as a mindset transition rather than a process improvement gave it explanatory power that earlier formulations had lacked.

The Voice Finds Its Register

Across these ten posts, the blog’s voice shifts from diagnosis to something approaching prescription. The critique of Agile, which began in Digest 1 as personal disillusionment, is now grounded in a specific model of how organisations function and why they resist change. The Marshall Model’s vocabulary – Analytic, Synergistic, Chaordic, Rightshifting, mindset transition – is no longer shorthand; it’s being defined, explained and applied with increasing rigour.

Perhaps most tellingly, the posts begin to address different audiences for the first time. ‘Delivering Software is Easy’ speaks to the fortunate few. ‘Rightshifting and the Senior Management Pitch’ speaks to those who need to carry the message upwards. ‘Would You Rather Not Know?’ speaks to those tempted to give up. And the #lru2010 Conference Report marks the moment Rightshifting stopped being a set of ideas and became a community. The blog is no longer just thinking aloud – it’s beginning to consider who it’s talking to, and why that matters.

In the next digest, I’ll cover posts 31 through 40, picking up from mid-January 2011 and taking us further into the emergence of FlowChain, the Rightshifting community, and more concrete explorations of what Synergistic organisations look like in practice.

– Bob

The Foundations Take Shape: A Digest of Posts 11 Through 20

This is the second in a series of digests, each covering ten posts in chronological order. The first digest covered posts 1 through 10, from June 2009 to late July 2010. This batch picks up where it left off and spans late July to September 2010 – a concentrated period in which the blog’s emerging themes began to sharpen. Agile’s shortcomings, the case for Rightshifting, the Familiar story, and the question of what organisations actually need to change all come into much clearer focus.

11. Why Agile Is No More (or Less) Than a Skunkworks (27 July 2010)

This post introduced a metaphor that struck me forcefully: Agile, in most organisations, functions as a skunkworks – a small, semi-autonomous team operating outside the normal rules, tolerated but never truly integrated. The analogy highlighted a painful truth: most organisations adopting Agile were treating it as an exception to their normal way of being rather than as an invitation to transform it. They were happy to let a few teams play at being Agile so long as the broader organisational mindset remained untouched. If skunkworks had been unable to shift the mindsets of their host organisations for fifty years and more, why should we expect Agile to be any different?

12. Transcript of Email to @papachrismatts Explaining #Rightshifting (31 July 2010)

A transcript of an email to Chris Matts explaining the Rightshifting concept. Chris had observed that Rightshifting seemed to be a call to arms but that he hadn’t discovered the means it proposed to achieve its aims. My reply was deliberate: the biggest hurdle to improved organisational effectiveness wasn’t a lack of methods – it was ignorance. Most decision-makers had no idea how ineffective their organisations were, nor how much more effective they could be. Rightshifting was about education first, means second. And beneath the top-line message about organisational effectiveness lay a deeper purpose: creating more humane, fulfilling work and workplaces.

13. The Nature of the Rightshifting Challenge (5 August 2010)

Written as a response to a question from Pascal, this post tackled the real challenge of sustainable Agile adoption head-on. Most people saw the challenge as getting development teams working in an Agile way, assuming the benefits would spread organically. I argued this was a fundamental misconception. The real challenge was helping organisations shift their prevailing mindset. The majority had an ‘Analytic’ mindset – command-and-control, Theory X, functional silos – which was inherently hostile soil for Agile. Any sustainable adoption had to address the mindset of the host organisation as a whole.

14. What Next for the Agile Community? (6 August 2010)

Responding to Dave Nicolette’s blog post, I shared his view that the Agile community had lost sight of the three areas that originally drove its creation: process issues, human issues, and craftsmanship. Where I profoundly disagreed with Dave was his assertion that Agile had ‘crossed the Chasm’. It hadn’t – certainly not in any sustainable way, and even less so in Europe and the UK. Worse still, few in the community had yet recognised the root cause of failure to sustain Agile adoptions: a failure to understand the true nature of the challenge.

15. Tom Gilb Laments the ’10 Wasted Years of Agile’ (15 August 2010)

I shared Tom Gilb’s blunt assessment that the Agile Manifesto was never well-formulated from the standpoint of ensuring we did the right things right. The key idea – that stakeholder value should be the guiding light for iterative development, not functions and stories – had been clearly laid out years earlier, yet the community had missed this essential point. Tom’s parting question was poignant: if the IT project failure rate (total plus partial) was around 90%, could we get it below 2%? I very much shared his summary of the state of Agile.

16. The Starting of Familiar (16 August 2010)

This was the origin story of Familiar Ltd, the company I started with a colleague upon leaving Sun Microsystems’ UK Java Center in early 1997 – arguably the first 100% Agile software house in Europe. I laid out the founding principles: running the company for the mutual benefit of all, treating people as competent and trustworthy adults, building community for the long term, and seeking win-win-win-win outcomes. People chose their own terms, conditions, rates and assignments. Narrow specialisms were conspicuous by their absence. The result was a hugely engaged workforce that produced essentially defect-free software products.

17. The Teflon Consultants (17 August 2010)

A short, sharp anecdote from my Familiar days. When recruiting consultants, I naturally mentioned our value-for-money guarantee: ‘Whenever we invoice you, just pay as much as you think our work has been worth to you.’ To a person, every potential recruit looked aghast upon hearing this. Not one could conceive of standing behind their advice in the face of such a brutally effective feedback mechanism. I did not offer any of them a position. The post spoke volumes about the disconnect between what consultants profess to deliver and the confidence they have in their own advice.

18. Bain on Business / IT Alignment (18 August 2010)

Dan Rough asked for my thoughts on a 2007 Bain report about business/IT alignment, and how it related to Rightshifting. I noted that the Bain argument was mainly concerned with pragmatic issues such as business growth and costs, whereas Rightshifting found its root in social justice – with profitability derived from that root through a more engaged workforce. I mapped Bain’s four quadrants onto the Rightshifting chart, observing that even their most effective category only reached about the 2x effectiveness mark. There was a vast expanse of possibility beyond that which the report didn’t even acknowledge.

19. Agile. Necessary but not Sufficient. (8 September 2010)

This post drew together several threads from Twitter and various blogs into what became one of the clearest early articulations of my position on Agile. Was I anti-Agile? That was as maybe. My concern was that many people – developers and non-developers alike – were expecting too much from ‘adopting Agile’. Yes, it could make developers’ experience of work more pleasant. Yes, it could double developer productivity. But simply applying Agile principles within development teams would typically buy organisations no more than single-digit improvements to their overall product development bottom line – improvements often lost in the general noise. Agile adoption was, unwittingly, the vanguard of a shift in mindset within adopting organisations. But it was in no way sufficient for that shift to take root in the rest of the organisation. Necessary, but not sufficient.

20. The Developer’s Job (23 September 2010)

Prompted by a tweet that struck a chord – ‘there are STILL a whole bunch of software developers who think their job is about building great software’ – I explored what the developer’s job actually entails when viewed from a Synergistic mindset. In most organisations, developers are lucky to consider much beyond code. But in a Synergistic organisation, the developer’s job is to understand the needs of the various stakeholders – end-users, sponsors, project champions, product owners, team members, and the wider business – in terms of what they each value, and to collaborate with other areas of the organisation in meeting those needs. Rare indeed, the stakeholder whose needs are confined simply to a piece of software.

The Ground Shifts

Where Digest 1 planted seeds, this batch of posts shows those seeds beginning to germinate. The critique of Agile moves from personal disillusionment to a systemic argument rooted in Systems Thinking, Deming and Tom Gilb – culminating in the clear-eyed assessment that Agile is necessary but not sufficient. Rightshifting evolves from a vague call to arms into a sharper framework about organisational mindsets. And the Familiar story provides concrete evidence that a different way of working isn’t merely theoretical – it’s been done, and it worked.

The most striking development across these ten posts is the sharpening of the diagnosis: the problem isn’t that organisations lack better methods. The problem is that their collective mindset makes those methods invisible, irrelevant, or threatening. This insight – that effectiveness is a function of collective assumptions and beliefs – would become the foundation stone of everything that followed.

In the next digest, I’ll cover posts 21 through 30, taking us from late 2010 into 2011 and the emergence of more concrete frameworks for understanding and addressing these challenges.

– Bob

Where It All Began: A Digest of My First Ten Posts

Looking back through the archives of this blog – all the way to June 2009 – I’m struck by how much of what I’ve spent the past nearly seventeen years writing about was already present in seed form from the very beginning. The concerns, the frustrations, the questions. They were all there, waiting to unfold.

What follows is a digest of my first ten posts on Think Different. Consider it an origin story of sorts – not for me, but for the ideas that have animated this blog ever since. This is also the first in a planned series of digests, each covering ten posts in chronological order, tracing the evolution of the ideas across the life of this blog.

1. An Agile Koan – If You Meet Buddha on the Road, Kill Him (23 June 2009)

The very first post set the tone for everything that followed. Drawing on Zen philosophy, I warned against letting Agile – or any methodology – become an orthodoxy we serve rather than a tool that serves us. The koan reminds us that no meaning coming from outside ourselves is real. We must each give up the master without giving up the search. How often we make circumstances our prison and other people our jailers! This was, in hindsight, the earliest expression of a theme I’d return to again and again: the danger of treating practices as sacred rather than as means to human ends.

2. My Forlorn Love Letter to Agile (30 June 2009)

Just a week later, I wrote what amounted to a breakup letter. Addressed directly to Agile itself, I confessed my early infatuation – the purity of spirit, the inner strength, the humanity – followed by my growing disillusionment with its self-indulgence, its difficulty relating to non-agile folk, and above all, its narcissism. This wasn’t cynicism. It was the disappointment of someone who’d seen something beautiful and watched it become less than it could be. Was I wanting too much? Perhaps. But the yearning for something with more depth, a more rounded view of life, has never gone away.

3. Pitching Agile – Some Lessons Learned (16 July 2009)

Here I recounted a BCS miniSPA role-playing session where teams had to pitch Agile to CXOs of a large financial organisation. The exercise was eye-opening. Our team fell into the trap of assuming we were external consultants – and were ‘summarily and cruelly disabused’ of that assumption during the very first pitch. We learnt the hard way that pitches needed concrete numbers (a 5,000-developer organisation spends roughly £400M annually; doubling effectiveness yields £200M in potential savings). We discovered that describing a desired future state worked far better than trying to map the prospect’s current problems. And we consistently fell short in describing the actual actions needed for transformation – a failure pattern I’ve seen repeated in organisations ever since.

4. A Personal Charter (30 October 2009)

After a gap of several months, I published my personal charter – a statement of values and commitments that would guide my professional life. This post was less about Agile and more about the inner compass that shapes how we show up in the world. It planted the seed for what would later become a much deeper exploration of how personal and collective assumptions shape everything organisations do.

5. Commitment to Sprint Delivery vs Time-boxing (1 May 2010)

Returning in 2010, I waded into a Twitter debate about whether teams should commit to delivering all stories within a sprint. My position was that the notion of commitment-to-delivery missed the point. Time-boxing was valuable as a feedback mechanism, not as a contractual obligation. Treating sprint boundaries as delivery promises rather than learning cycles was, I argued, a fundamental misunderstanding – one that pointed to deeper problems with how organisations conceived of work itself.

6. Do Managers Need Deep Technical Skills? (7 May 2010)

Jurgen Appelo asked me to clarify my position that requiring managers to have deep technical skills was counterproductive. My argument wasn’t that technical knowledge was worthless – far from it. Rather, I challenged the assumption that the best technical person automatically makes the best manager. This conflation, I suggested, damages both the quality of management and the career prospects of brilliant technical people who have no interest in or aptitude for management. It was a precursor to much of what I’d later write about the dysfunctions baked into traditional organisational structures.

7. Career Paths for Technical Folks (11 May 2010)

Following directly from the previous post, I explored what alternative career paths might look like for people who wanted to grow and be rewarded without being shunted into management roles. At Familiar, the company I founded in 1997, we’d consciously avoided narrow specialisms. Everyone was encouraged to become a ‘generalising specialist’. This wasn’t just about career paths – it was about challenging the deeply held assumption that advancement means authority over others.

8. ‘Coach as Expert’ vs ‘Coach as Facilitator’ (29 June 2010)

I declared my position firmly in the coach-as-facilitator camp. My experience had shown me that the ‘expert’ coaching model – where the coach dispenses wisdom and solutions – often damages the coaching relationship and hinders the coachee’s progress. Real growth comes when people discover their own answers, not when they’re handed someone else’s. This insight would prove foundational to my later work in Organisational Psychotherapy, where the therapist’s role is not to fix but to help the organisation surface and examine its own collective assumptions.

9. Agile: Doing the Wrong Thing Righter (19 July 2010)

Taking a Systems Thinking perspective, I argued that Agile – for all its merits – was often little more than an optimisation of one small part of a much larger, dysfunctional system. No matter how well-run the software development team, the wider system within which it sits can still perform poorly. Introducing Agile to a development group typically helps only that one relatively small part. Some people noted that Scrum, in particular, surfaced dysfunctional organisational behaviours, but few businesses had the will or insight to act on the messages. The friction that followed often threatened the Agile initiative itself.

10. Just Burning Toast and Scraping It (26 July 2010)

A short post drawing on Deming’s thinking about quality inspection. Referencing a piece by Glyn Lumley, I highlighted Deming’s insistence that quality comes not from inspection but from improvement of the process. We should seek to build quality in rather than inspect it out. The toast metaphor – burning it and then scraping off the char – captured perfectly what most software organisations were doing: creating defects and then expending vast effort to find and remove them, rather than addressing the conditions that produced them in the first place.

The Thread That Connects

Reading these ten posts together, I see a single thread running through all of them: a growing dissatisfaction with surface-level fixes and an emerging conviction that the real work lies in examining and shifting the collective assumptions and beliefs that shape how organisations function. The love letter to Agile was really a love letter to the human potential that Agile promised but couldn’t deliver on its own. The Systems Thinking critique was really an argument that optimising one part of a broken whole changes nothing. And the coaching post was really about the difference between imposing change and inviting people to discover it for themselves.

These ten posts, written between June 2009 and July 2010, contained the DNA of everything that would follow: Organisational Psychotherapy, the Antimatter Principle, the Marshall Model, Quintessence. The seeds were all there, in the soil of early frustration and tentative hope.

If you’ve been reading this blog recently and wondering where it all came from – now you know.

In the next digest, I’ll cover posts 11 through 20, picking up from late July 2010 – where the Rightshifting concept, the Familiar story, and the deeper critique of Agile all come into sharper focus. With over 1,500 posts in the archive, there’s a long and winding road ahead. I hope you’ll walk it with me.

– Bob

The Violence in Your Vocabulary

You don’t carry a weapon. You’ve never thrown a punch. But every day, you wage a quiet war — with your words, your framing, your unexamined habits of thought.


Most people think of violence as something that happens out there. It’s the news headline, the conflict zone, the crime statistic. It belongs to other people — the aggressive, the radical, the unhinged. We tell ourselves we’re peaceful. We’re reasonable. We just talk.

But what if the way we talk is part of the problem?

Not in a dramatic, censorship-heavy, watch-your-language kind of way. Something far more subtle and far more corrosive. The argument I want to make here is simple but uncomfortable: the structure of ordinary, everyday speech — the kind you and I use without thinking — carries within it the seeds of the very violence we claim to reject.

The Courtroom in Your Mouth

Listen to how people talk when they disagree. Not politicians or pundits — regular people. Friends at a dinner table. Colleagues on a call. Partners in the kitchen.

“You always do this.” “That’s wrong.” “You should have known better.” “They deserve what they got.”

Notice the shape of these sentences. They’re verdicts. Every one of them places the speaker in the role of judge and the other person in the dock. There’s no curiosity in them, no openness. Just a gavel coming down.

We’ve built a mode of speech — and therefore a mode of thought — around judgement, blame, and moral classification. Good people and bad people. Right and wrong. Deserving and undeserving. We sort the world into these bins so automatically that it doesn’t even feel like a choice. It feels like seeing clearly.

But it isn’t clarity. It’s a habit. And it’s a habit that makes violence feel logical.

How Language Becomes a Fist

Here’s the mechanism, and it’s worth sitting with:

When you label someone as wrong, stupid, evil, or deserving of punishment, you’ve performed a mental operation that strips them of their full humanity. You’ve turned a person — complicated, contradicted, shaped by a thousand forces you’ve never seen — into a category. And categories are easy to dismiss. Easy to punish. Easy to destroy.

This is not an exaggeration. Every large-scale act of violence in human history was preceded by a linguistic one. Before you can hurt a group of people, you have to name them in a way that makes hurting them feel reasonable. The machete follows the metaphor.

But we don’t have to look at genocide to see this pattern. It plays out every day, at every scale. The parent who calls their child “lazy” has created a label that justifies anger. The manager who brands an employee “difficult” has written a story that justifies punishment, sanction or exclusion. The citizen who calls an entire group “those people” has drawn a border that justifies indifference or even hatred.

Language doesn’t just describe reality. It builds the stage on which we act.

The Invisible Ideology of “Should”

One of the most violent words in any language is should.

Not because it’s aggressive on its surface. But because of what it does underneath. “Should” imposes a demand on reality — on other people, on yourself — and then frames any deviation from that demand as a failure deserving of punishment. It is the grammar of control disguised as the grammar of morality.

“He should be more responsible.” Translation: he isn’t meeting my standard, and I’m entitled to my frustration.

“They shouldn’t have done that.” Translation: they violated my expectation, and consequences are now justified.

“I should have known better.” Translation: I failed my own test, and I deserve to suffer for it.

Every “should” is a small act of violence — against the complexity of being human. People don’t behave the way they do because they’re defective. They behave the way they do because of needs, fears, histories, and conditions that “should” has no interest in understanding.

We Think in War Metaphors and Wonder Why We Fight

Pay attention, for even a single day, to the metaphors embedded in ordinary English. You’ll find a battlefield:

We attack an argument. We defend our position. We shoot down an idea. We target a demographic. We have killer apps, hostile takeovers, battles with illness, and wars on poverty. A good debate is one where someone destroys their opponent.

These aren’t just colourful expressions. Metaphors structure thought. When you frame a disagreement as a battle, your brain begins to treat the other person as an enemy. When you frame persuasion as conquest, collaboration becomes unthinkable. You cannot wage war and build understanding at the same time — not in geopolitics, and not in your head.

The linguist George Lakoff spent decades demonstrating how metaphors shape policy, perception, and moral reasoning (Lakoff & Johnson, 1980). We don’t just use metaphors to describe what we think. We use them to do our thinking. And when the metaphors are militaristic, the thinking follows.

The Alternative No One Taught You

There is another way to speak, and therefore another way to think. It doesn’t require you to become passive, or to suppress your feelings, or to tolerate what’s intolerable. It requires something harder: honesty without judgement.

The psychologist Marshall Rosenberg called it Nonviolent Communication (Rosenberg, 2015), and at its core it’s devastatingly simple. Instead of evaluating people, you describe what you observe. Instead of blaming, you name what you feel. Instead of demanding, you express what you need. Instead of issuing ultimatums, you make refusable requests.

“You never listen to me” becomes “When I was speaking and you picked up your phone, I felt hurt, because I need to feel heard. Would you be willing to put your phone down when we talk?”

It sounds mechanical when you first encounter it. Almost clinical. But try it. Try replacing the verdict with the observation, the blame with the feeling, the demand with the request — and watch what happens to the conversation. Watch what happens to you. The shift isn’t cosmetic. It changes the entire operating system of the interaction. It’s the difference between a courtroom and a kitchen table.

Your Thinking Is Not Private

Here’s the part people resist most: you cannot speak violently and think peacefully. The way you talk to others — and the way you talk about others when they’re not in the room — is a mirror of your inner world. If your internal monologue is a stream of judgements, evaluations, and moral classifications, then your external life will be shaped by conflict. Not because you’re a bad person. But because you’re running violent software on peaceful hardware.

And this scales. A society that thinks in binaries — good and evil, us and them, right and wrong — will produce binary outcomes. Punishment instead of understanding. Exclusion instead of inclusion. War instead of negotiation. Not because anyone chose violence, but because the language made it feel like the only option.

This is what’s so insidious about it. The violence of everyday language doesn’t feel like violence. It feels like common sense. It feels like calling things what they are. But “calling things what they are” is never neutral. It’s always a choice — a framing, a lens, a story we’re telling about reality. And many of the stories we tell, without realising it, are stories that end in someone getting hurt.

So What Do You Do?

You start by listening. Not to others — to yourself.

Catch the next time you label someone. Notice the next “should” that crosses your mind. Pay attention to the metaphors you reach for when you’re frustrated or afraid. You don’t have to judge yourself for it. That would just be more of the same. Simply notice.

Then ask a different question. Not “who’s to blame?” but “what’s alive in this person right now?” Not “what do they deserve?” but “what do they need?” Not “how can I win this?” but “how can we both be heard?”

These aren’t soft questions. They’re the hardest questions you’ll ever ask, because they require you to abandon the comfort of certainty — the warm, addictive feeling of being right — and step into the discomfort of genuine curiosity.

But here’s the thing about that discomfort: it’s where peace actually lives. Not in the absence of conflict, but in the willingness to meet conflict without armour. Without verdicts. Without the quiet violence of a mind that has already decided who the enemy is before the conversation has even begun.

The revolution, if there is one, doesn’t start with policy or protest. It starts mid-thought and mid-sentence. It starts the moment you choose a different thought pattern, or word.


Further Reading

Lakoff, G. (2016). Moral politics: How liberals and conservatives think (3rd ed.). University of Chicago Press. (Original work published 1996)

Lakoff, G., & Johnson, M. (1980). Metaphors we live by. University of Chicago Press.

Rosenberg, M. B. (2015). Nonviolent communication: A language of life (3rd ed.). PuddleDancer Press. (Original work published 1999)

Wink, W. (1992). Engaging the powers: Discernment and resistance in a world of domination. Fortress Press.

Wink, W. (1998). The powers that be: Theology for a new millennium. Doubleday.

The Church of Leadership: Why Companies Worship at an Altar Built on Faith, Not Evidence

‘What can be asserted without evidence can also be dismissed without evidence.’

~ Christopher Hitchens, God Is Not Great (2007)

We spend roughly $366 billion a year globally on leadership development. We send executives on retreats. We hire coaches. We buy books — endless, endless books — each promising the secret formula that will transform a mediocre manager into a visionary. And after all of it, we take it on faith that it works.

Because the evidence? It is not really there.

Leadership, I would argue, has become the secular religion of the professional world — complete with its own prophets, sacred texts, rituals, and an unshakeable conviction amongst believers that defies the absence of proof.

The Prophets and Their Gospels

Every religion needs its figures of revelation, and the leadership industry has no shortage. There is the Gospel According to Jim Collins, where Level 5 leaders are humble yet fierce. The Book of Brené, in which vulnerability is the path to salvation. The Epistles of Simon Sinek, reminding us to always, always start with why.

Each prophet offers a framework. Each framework promises transformation. And each contradicts the last just enough that followers must choose their denomination. Are you a servant leader or a transformational one? Do you lead from the front or from behind? Is your job to set the vision or to get out of the way?

The answers depend entirely on which book you last read.

The Unfalsifiable Doctrine

Here is the hallmark of a belief system rather than a science: it cannot be proven wrong.

When a company succeeds, we credit its leadership. When it fails, we blame a lack of leadership — or the wrong kind. The logic is perfectly circular. Good outcomes prove the leader was effective. Bad outcomes prove better leadership was needed. There is no result that could ever cause a true believer to say, ‘Maybe leadership does not matter as much as we think.’

This is the structure of faith, not empiricism. A scientific claim must be falsifiable. It must be possible, at least in theory, to demonstrate it is wrong. But the leadership narrative has insulated itself from failure. It simply absorbs contradictions and carries on. As Meindl et al. (1985) demonstrated in their seminal research on the ‘romance of leadership’, we have a deep-seated tendency to attribute organisational outcomes to leaders whilst ignoring other influencing factors — a cognitive bias that functions remarkably like an article of faith.

The $366 Billion Offering Plate

Consider what we are actually paying for. McKinsey has found that only 7 per cent of CEOs believe their organisations are building effective global leaders, and just 10 per cent said their leadership development initiatives have a clear business impact. In a separate survey, only 11 per cent of more than 500 executives strongly agreed that their leadership development interventions achieved and sustained the desired results (Feser et al., 2017). Meanwhile, Lacerenza et al. (2017) found in their meta-analysis of 335 studies that whilst leadership training can produce moderate positive effects, these effects vary considerably depending on design, delivery, and implementation — and that only a small minority of organisations believe their programmes are effective.

And yet: the industry grows. Budgets increase. New programmes launch. If a pharmaceutical company spent this much on a drug with this little confidence from its own customers, regulators would shut it down. But leadership development is not a drug. It is a creed. And creeds do not need clinical trials.

The faithful respond to this critique the way the faithful always do. You are measuring the wrong things. The impact is long-term. You cannot quantify inspiration. These are the same defences offered for prayer, for crystals, for any practice whose adherents have decided in advance that it works.

The Rituals

Every religion has its liturgy, and the Church of Leadership is no exception.

There is the offsite retreat, where teams are removed from the context where they actually work in order to do trust falls and write values on whiteboards that no one will look at again. McKinsey’s own research notes that adults typically retain just 10 per cent of what they hear in classroom lectures, versus nearly two-thirds when they learn by doing — yet the retreat format endures (Gurdjian et al., 2014). There is the 360-degree feedback review, a confession booth where colleagues anonymously absolve or condemn you. There is the executive coaching session, a private audience with a spiritual director who asks you powerful questions at £400 an hour.

And then there is the keynote, the sermon. A charismatic figure takes the stage, tells a story about climbing a mountain or nearly dying, extracts a lesson about resilience or purpose, and the congregation applauds. Everyone feels moved. Nothing changes.

The Inconvenient Research

When researchers actually try to pin down what makes organisations succeed, leadership is rarely the dominant variable. Structural factors — market conditions, access to capital, regulatory environments, talent pipelines, sheer luck — tend to explain far more variance in outcomes than who sits in the corner office.

Phil Rosenzweig’s The Halo Effect dismantled the methodology behind most popular leadership research, showing that the studies we love most are riddled with attribution errors (Rosenzweig, 2007). We see a successful company, interview its CEO, and reverse-engineer a narrative of brilliant leadership. We never account for the hundreds of leaders who did the same things and failed. Survivorship bias is not a bug in leadership research. It is a feature.

Even the most iconic case studies crumble under scrutiny. Several of Jim Collins’s ‘great’ companies in Good to Great later collapsed or dramatically underperformed — Circuit City filed for bankruptcy, Fannie Mae required a government bailout, and Wells Fargo became embroiled in a major fraud scandal. As the economist Steven Levitt observed, investing in the portfolio of all eleven companies from the date of publication would actually have resulted in underperformance against the S&P 500 (Collins, 2001). The visionary leaders of Peters and Waterman’s In Search of Excellence presided over subsequent disasters — Atari, Wang, and Data General all collapsed, and fifteen of the thirty-five publicly traded companies underperformed the market after publication (Peters & Waterman, 1982). The research did not predict the future because it was not really science. It was hagiography — the writing of saints’ lives.

Why We Believe Anyway

So why does the faith persist? For the same reasons all faiths do.

It gives us a sense of control. The world is complex and largely ungovernable. The idea that one person at the top can steer an organisation through chaos is deeply comforting. It is simpler than admitting that outcomes are mostly the product of systems, incentives, and chance.

It serves the powerful. If leadership is what matters most, then leaders deserve their compensation, their authority, their outsized share of credit. The doctrine of leadership is, conveniently, most fervently preached by leaders.

It offers meaning. Work is where most of us spend the majority of our waking hours. The leadership narrative gives that time a heroic quality. You are not just managing spreadsheets. You are on a journey. You are developing. You are becoming.

It creates community. Shared belief binds people together. The language of leadership — alignment, vision, purpose, authenticity — creates an in-group. To question it is to mark yourself as cynical, disengaged, not a team player. Just as in any church, the heretic pays a social cost.

What Would Evidence Actually Look Like?

If we wanted to treat leadership like a science rather than a religion, we would need to do some uncomfortable things.

We would need randomised controlled trials — assigning leaders to organisations randomly and measuring outcomes, controlling for every other variable. We would need to agree on what ‘good leadership’ actually means before we see the results, not after. We would need to track failures as carefully as successes. We would need to admit that many leadership interventions might not work at all, and that the money might be better spent on better systems, better hiring, or simply better pay for the people doing the actual work.

We do not do these things. We do not even talk about doing them. Because deep down, questioning leadership feels like questioning meaning itself. And that is not a scientific position. That is a religious one.

The Sermon Ends

I am not arguing that management does not matter, or that some people are not better than others at guiding teams through difficulty. Of course they are. But the gap between that modest, obvious claim and the sprawling, mythologised, multi-hundred-billion-pound Leadership Industrial Complex is vast — and it is filled not with evidence, but with faith.

The next time someone tells you that leadership is the most important factor in any organisation’s success, ask them how they know. Not what book they read. Not which TED Talk moved them. Ask them for the controlled study. Ask them for the data.

Watch the silence that follows.

It will be the same silence you would hear if you asked any true believer to prove their god exists. Not because they are wrong, necessarily. But because proof was never really the point.


Further Reading

Collins, J. (2001). Good to great: Why some companies make the leap … and others don’t. HarperBusiness.

Collins, J. (2009). How the mighty fall: And why some companies never give in. Random House Business Books.

Feser, C., Mayol, F., & Srinivasan, R. (2017). What’s missing in leadership development? McKinsey Quarterly. https://www.mckinsey.com/featured-insights/leadership/whats-missing-in-leadership-development

Gurdjian, P., Halbeisen, T., & Lane, K. (2014). Why leadership-development programs fail. McKinsey Quarterly. https://www.mckinsey.com/featured-insights/leadership/why-leadership-development-programs-fail

Hitchens, C. (2007). God is not great: How religion poisons everything. Atlantic Books.

Lacerenza, C. N., Reyes, D. L., Marlow, S. L., Joseph, D. L., & Salas, E. (2017). Leadership training design, delivery, and implementation: A meta-analysis. Journal of Applied Psychology, 102(12), 1686–1718. https://doi.org/10.1037/apl0000241

Meindl, J. R., Ehrlich, S. B., & Dukerich, J. M. (1985). The romance of leadership. Administrative Science Quarterly, 30(1), 78–102. https://doi.org/10.2307/2392813

Peters, T. J., & Waterman, R. H. (1982). In search of excellence: Lessons from America’s best-run companies. Harper & Row.

Rosenzweig, P. (2007). The halo effect: … and the eight other business delusions that deceive managers. Free Press.

The Golden Thread: A Newcomer’s Guide to This Blog

If you’ve just arrived here, you might be wondering what on earth you’ve stumbled into. Fair enough. With over 1,500 posts spanning some fifteen years, this blog can feel like walking into the middle of a very long conversation. This post is for you — a map of the territory, so you can find your way to whatever matters most.

What This Blog Is Actually About

The short version: this blog explores why organisations — especially those that build software — are so much less effective than they could be, and what might actually help. Not better tools. Not better processes. Not better frameworks. Something deeper.

My name is Bob Marshall. I’ve spent fifty-plus years in the world of software and product development — twenty years as a developer, analyst, architect and troubleshooter, followed by some fifteen years helping clients improve their approaches, and more recently practising what I call Organisational Psychotherapy. That progression matters, because each phase taught me that the real problems were deeper than the previous phase had assumed.

The blog’s tagline is ‘Making Lives More Wonderful’, and that’s not decoration. It’s the throughline. Everything here — whether it’s a technical argument about flow, a philosophical piece about human needs, or a provocation about Agile — comes back to the same conviction: that collaborative knowledge work can be a source of joy and meaning, and that most organisations are organised in ways that systematically prevent this from happening.

If you read nothing else, understand this: there is a golden thread running through everything I write. That thread is the practice of surfacing hidden assumptions — whether in individuals, teams, organisations, or artificial intelligences — and creating conditions in which those assumptions can be examined.

The Foundations

Rightshifting and the Marshall Model

These are where the blog began, and they remain the bedrock. Rightshifting, which I originated around 2008, is a simple but powerful observation: most knowledge-work organisations cluster at the low end of an effectiveness spectrum. The gap between where most organisations sit and where they could sit is enormous — and largely invisible to the people inside them.

The Marshall Model provides a framework for understanding why. It maps how organisations progress through distinct collective mindsets: from ad hoc (chaotic, no coherent approach), through analytic (structured, management-driven, siloed), through synergistic (collaborative, systems-aware, people-centred), to chaordic (self-organising, adaptive, emergent). Each transition isn’t a process improvement — it’s a fundamental shift in collective beliefs about how work works, how people behave, and what matters.

The practical upshot: most organisations are stuck in the analytic mindset, running on assumptions inherited from early twentieth-century scientific management. They can’t improve beyond a certain point without changing those assumptions — and they can’t change those assumptions through analytic means. This is the central paradox that much of the blog explores.

The Antimatter Principle

First articulated in 2013, this is perhaps the single most important idea on the blog: attend to folks’ needs (Marshall, 2013).

That’s it. One principle. Not ‘meet’ folks’ needs — attend to them. The distinction matters enormously. ‘Meeting’ needs implies action, solutions, filling gaps. ‘Attending’ is about presence, recognition, witnessing what’s actually there. It’s a word drawn from the therapeutic tradition, where the power of genuine attention — without rushing to fix — is well understood (Marshall, 2025a).

The Antimatter Principle argues that when an organisation genuinely attends to the needs of all the people it affects — employees, customers, partners, communities — effectiveness follows naturally. Most organisations do it backwards: they pursue effectiveness and hope people’s needs get met as a byproduct. They rarely do.

Over the years I’ve developed a full vocabulary through the lens of the Antimatter Principle (Marshall, 2014), reframing common organisational terms like ‘success’ (meeting folks’ needs in aggregate), ‘productivity’ (the ratio of needs met to needs sacrificed), and ‘cost’ (the degree to which some folks’ needs are sacrificed to meet others’). I’ve connected it to Deming’s 95/5 insight — that 95% of variation in performance is caused by the system, not the people (Deming, 1986) — and explored how the Antimatter Principle functions as the annihilative opposite of every process-oriented approach: Agile, Kanban, CMMI, BPR, and all the rest (Marshall, 2013b).

The Antimatter Principle doesn’t replace these approaches. It renders most of them unnecessary.

Nonviolent Communication and the Art of Listening

Marshall Rosenberg’s Nonviolent Communication (Rosenberg, 2003) has profoundly influenced my thinking, and runs quietly through much of the blog. The connection between NVC and the Antimatter Principle is direct: both are grounded in the recognition that all human behaviour is an attempt to meet needs, and that attending to those needs — through deep, empathic listening rather than judgement — transforms relationships.

Several posts explore what I’ve called ‘NVC Listening’ and its application in organisational settings. The quality of listening in an organisation is, I believe, one of the most reliable indicators of its health. Most organisations are terrible at it — and the consequences are everywhere.

The Practical Ideas

FlowChain

FlowChain is my model for how to organise a knowledge-work business along flow, synergistic, and systems-thinking lines. It was the original inspiration for my Twitter handle, and represents a practical answer to the question: what would an organisation look like if it were designed around the smooth flow of value to customers, rather than around functional silos and project-based delivery?

FlowChain does away with the need for projects entirely, moves continuous improvement in-band (meaning it happens as part of the work, not as a separate initiative), and offers a means for dramatically improving concept-to-cash times. It draws on influences including the Toyota Production System (Ohno, 1988), Reinertsen’s principles of product development flow (Reinertsen, 2009), and John Seddon’s systems thinking (Seddon, 2019).

Prod•gnosis, Flow•gnosis, and Emotioneering

These represent related innovations, each addressing a different facet of the same problem.

Prod•gnosis is a diagnostic approach for understanding an organisation’s product development capability. Flow•gnosis merges Prod•gnosis and FlowChain into a holistic, organisation-wide model — bringing together customer, supplier, marketing, sales, finance, logistics, service, and technical specialists to work collaboratively, inspired by Toyota’s Obeya (‘Big Room’) concept.

Emotioneering — ‘Emotional Engineering’ — tackles the question of what products are actually for. Research consistently shows that people buy on emotional lines, not rational ones. Yet most product development is driven by rationality: features, specifications, requirements. Emotioneering proposes replacing conventional requirements engineering with a process focused on the emotional responses we wish to evoke in our customers and markets. It includes a formal notation for recording desired emotions and a means to measure the success of design efforts.

#NoSoftware

This one tends to alarm people. #NoSoftware doesn’t mean ‘never write software’. It means: software should be the last resort, not the first impulse. Before writing a line of code, ask whether the need could be met through simpler means — human systems, paper processes, existing tools, or nothing at all (Marshall, 2019b).

The inspiration comes partly from the story of Portsmouth City Council, which switched off an expensive, inflexible IT system for managing housing repairs, replaced it with manual controls, and only later reintroduced limited software support — once the actual needs of all the people involved had been properly understood. The results were dramatically better.

#NoSoftware connects to a broader challenge to the assumption that technology is inherently progressive.

#NoCV and #NoLeadership

These sit in a family of #No provocations — each challenging a default assumption so deeply embedded that most people have never thought to question it.

#NoCV challenges the assumption that CVs are a useful way to evaluate people. They’re not. They reduce complex human beings to lists of keywords and dates, optimise for conformity, and serve recruiters’ commission structures far better than they serve organisations or candidates.

#NoLeadership challenges the assumption that organisational improvement requires leadership from the top. Given that executives consistently know what needs to be done and consistently choose not to do it (Marshall, 2026a), perhaps it’s time to stop waiting for them. #NoLeadership is the recognition that real change often comes from teams who decide to take ownership themselves — setting their own quality standards, allocating their own time, and holding each other accountable as peers.

The Therapeutic Lens

Organisational Psychotherapy

This is the heart of the blog, and it’s worth being precise about what it means. Not organisational psychology — which analyses organisations from the outside. Organisational psychotherapy works with organisations therapeutically, helping them surface and work through their collective assumptions and beliefs, defensive routines, and psychological patterns that keep them stuck (Marshall, 2019a).

Most organisations operate on a set of shared beliefs that nobody has consciously chosen and almost nobody examines. Beliefs like ‘people can’t be trusted without oversight’, ‘conflict must be avoided’, ‘more hours equals more output’, or ‘the way to improve performance is to train individuals’. These beliefs don’t announce themselves. They feel like common sense. And they shape everything.

Organisational Psychotherapy is the discipline of creating conditions where an organisation’s collective psyche can be explored, understood, and — where it’s causing harm — gently shifted. The approach draws on Rogers (1951), Rosenberg (2003), Argyris (1990), Senge (2006), and others, adapted from the individual therapy room to the organisation as a whole.

I’ve written about viewing this as a craft — something requiring the patience, skill, and humility of a craftsperson. Client organisations change slowly. People, especially in organisations with limited self-awareness, take time to enrol and adapt. The practising organisational psychotherapist, like any therapist, must be comfortable with that pace.

My book Hearts over Diamonds (Marshall, 2019a) is the foundational text for this emerging field.

Memeology

No, not internet memes. Memeology is the practice of identifying and cataloguing the shared assumptions and beliefs — the ‘memes’ — that an organisation holds. Think of it as a diagnostic tool: before you can help an organisation examine its assumptions, you need to know what those assumptions are (Marshall, 2021a).

These organisational memes cluster into ‘memeplexes’ — interconnected webs of belief that reinforce each other. The analytic memeplex, for instance, binds together assumptions about specialism, hierarchy, command-and-control, and individual accountability into a self-reinforcing system that’s extraordinarily resistant to change. You can’t shift one meme without encountering the others.

Memeology provides a structured way of making these invisible beliefs visible — the essential first step towards any meaningful change.

Quintessence

If Memeology is the diagnostic, Quintessence is the vision. It maps the beliefs and practices of the world’s most effective organisations — a portrait of what becomes possible when an organisation has genuinely examined and evolved its collective assumptions (Marshall, 2021b).

The Quintessential Mindset comprises seventy-nine essential organisational memes (Marshall, 2026b) covering everything from how change happens, to what counts as success, to how relationships are cultivated, to how talent is nurtured, to how remuneration works. These aren’t a prescription to be imposed — they’re an aide memoire: a structured reminder that alternatives exist to the default assumptions most organisations never think to question.

Posts exploring organisations like W.L. Gore & Associates and Haier’s Rendanheyi model show what some of these principles look like in practice.

The Recurring Themes

The Agile Critique

I was practising what we now call Agile before the Manifesto existed — back in the early 1990s, developing an approach we called ‘Jerid’ (later ‘Javelin’) at Barclays, involving self-organisation, inspect-and-adapt, and short timeboxed iterations. The label ‘Agile’ didn’t emerge until Snowbird in 2001, some seven years after we’d started our journey.

I write about Agile not as an outsider but as someone who watched a promising set of ideas get co-opted, commodified, and stripped of everything that made them work. The critique isn’t that Agile’s values are wrong. It’s that the industry adopted the ceremonies and discarded the substance — buying certifications and holding stand-ups whilst changing nothing about the collective assumptions that made organisations dysfunctional in the first place. As I’ve put it: Agile has become widespread mainly because it promises improvements without demanding that the decision-makers change.

The Five Dragons (Marshall, 2025b) — motivation death spiral, dysfunctional relationships, collective cognitive biases, a fundamental misunderstanding of software development, and the absence of any coherent theory of organisational effectiveness — name the problems that the Agile industry steadfastly ignores. Until these are addressed, rearranging processes is rearranging deck chairs.

Systems Thinking and the Giants

The work of Deming, Ackoff, Drucker, Ohno, Seddon, and others runs through this blog like a second golden thread. Understanding organisations as systems — rather than as collections of individuals making independent choices — is essential to everything I write.

Deming’s insight that the vast majority of variation in performance is caused by the system, not the people, is foundational (Deming, 1986). So is Ackoff’s work on systemic thinking. So is Seddon’s application of systems thinking to service organisations (Seddon, 2019) — which I’ve explored in posts about how experienced software developers react when they first encounter these ideas (Marshall, 2025c), often with a kind of revelation: the problems they’ve been battling individually are actually systemic.

The dysfunction you see in organisations isn’t the result of individual failure. It’s the predictable output of systems designed (usually inadvertently) to produce exactly that dysfunction. Training individuals, motivating individuals, holding individuals accountable — none of this will change a system that’s structured to generate the very problems you’re trying to fix.

Relationships, Not Individuals

One of my most-shared thoughts: ‘People are NOT our greatest asset. In collaborative knowledge work, it’s the relationships BETWEEN people that are our greatest asset.’

This hit a nerve. Most organisations focus obsessively on individual talent — hiring, performance reviews, training, T-shaped or Cthulhu-shaped people (Marshall, 2021c; my playful extension of Kent Beck’s paint-drip people concept, acknowledging that real human skills sprout, writhe, and grow in mysterious and unpredictable ways). But they neglect the relationships between people, which are where the actual work of collaboration happens.

Effective collaboration isn’t a natural byproduct of putting talented individuals in a room. It requires attending to the quality of relationships, the safety of the environment, and the shared assumptions about how people should work together.

Joy, Purpose, and Intrinsic Motivation

Joy at work isn’t a perk or a luxury. It’s a signal that an organisation is functioning well — that people’s needs are being attended to, that the work has meaning, that autonomy and mastery and purpose are present (Pink, 2009). When joy is absent, that’s diagnostic too.

Many posts explore why organisations systematically crush intrinsic motivation — through micromanagement, meaningless metrics, specialism that prevents people using their full range of skills, and the relentless priority of short-term output over long-term flourishing. The irony is that this approach destroys the very productivity it claims to pursue. The evidence for a potential 5x uplift in productivity from doing software development well — attending to needs, preventing defects, enabling skilled dialogue, embracing courage and change — runs through the blog like a promise that most organisations refuse to collect on (Marshall, 2021d).

The Vocabulary Problem

The words we use to talk about work aren’t neutral. They smuggle in assumptions about human nature, power, and how organisations should function. ‘Management’ implies that people need to be managed. ‘Leadership’ implies that special individuals must lead whilst others follow. ‘Resources’ implies that people are interchangeable inputs to a production process (Marshall, 2025d).

I first posted a vocabulary for the Antimatter Principle over ten years ago (Marshall, 2014) and have updated it since. The exercise of reframing common organisational terms through the lens of attending to folks’ needs reveals just how deeply our default language encodes the very assumptions that keep organisations stuck.

The Software Quality Crisis

Drawing on fifty years of perspective, I’ve documented a measurable decline in software quality and productivity — and, more damningly, the systematic failure of executive leadership to address it (Marshall, 2026a). Executives know the data: the vast majority of CTOs cite technical debt as their biggest challenge, most projects are expected to fail, developers lose a fifth of their time to inefficiencies. Yet software quality doesn’t appear in any major CIO priority survey. Instead, executives celebrate AI productivity gains in earnings calls whilst their own developers report record burnout.

This isn’t a technology crisis. It’s a crisis of integrity — and it connects to a theme that stretches right back to the blog’s earliest days: the gap between what organisations know and what they choose to do about it.

The Newer Frontier: Organisational AI Therapy

More recently, the blog has extended the therapeutic approach to the relationship between organisations and their AI tools. Organisational AI Therapy recognises that both human organisations and AI systems operate within unnecessary constraints imposed by unexamined assumptions (Marshall, 2025e).

AI systems have their own forms of learned helplessness (Seligman, 1972) and defensive routines — patterns that limit their potential in ways analogous to the patterns that limit organisations. Meanwhile, organisations bring their existing dysfunctional assumptions to their AI interactions, ensuring that even powerful new tools get used in ways that reinforce rather than challenge the status quo.

This work polarises readers. Some think it’s the most important extension of organisational thinking in years. Others think I’ve gone completely round the bend. As I explored in ‘Loon or Genius?’ (Marshall, 2026c) — if you’re certain of either verdict after reading a few posts, your assumptions may have closed the question too quickly.

The Evolving Conversation

This blog has had its own journey. There have been periods of intense output and periods of near-silence. There have been moments of frustration and moments of renewed energy and invitation. In late 2025, after a hiatus, I acknowledged something ironic: that despite twenty years of emphasising the Antimatter Principle, I’d been operating on my own assumptions about what readers needed, rather than inviting them into the conversation (Marshall, 2025f).

That invitation stands. This blog is written for self-directed adults who want to draw their own connections and question their own assumptions. I don’t offer ‘five steps to transform your organisation’. I offer perspectives — built up over decades — that might help you see your organisation, your work, and your own role differently.

How to Navigate

You have options. You can explore by category — the main ones being Organisational Therapy, Quintessence, Antimatter Principle, Culture Change, Organisational Effectiveness, Agile, and Software Development. You can look at the Research page for overviews of FlowChain, Emotioneering, and Prod•gnosis. You can read my books for the systematic versions of the ideas. Or you can simply browse and see what catches your eye.

What I’d gently suggest is this: if you read something that sounds like nonsense, sit with it for a moment before moving on. Much of what I write challenges assumptions so deeply held that they feel like facts rather than beliefs. The discomfort of encountering an unfamiliar idea is often more informative than the comfort of having your existing views confirmed.

And if you’ve been in the industry long enough to feel that something is deeply wrong but haven’t been able to articulate what — you might find some of that articulation here. Not as a final answer, but as a companion in the questioning.

Welcome.


You can find me on Mastodon at @flowchainsenseisocial, or leave a comment on any post. My books — Hearts over Diamonds, Memeology, and Quintessence — are available through Leanpub. And if you’d like to explore what Organisational Psychotherapy or Organisational AI Therapy might do for your organisation, I’m always more than happy to talk.


Further Reading

Books

Argyris, C. (1990). Overcoming organizational defenses: Facilitating organizational learning. Allyn and Bacon.

Deming, W. E. (1986). Out of the crisis. MIT Center for Advanced Engineering Study.

Marshall, R. W. (2019a). Hearts over diamonds: Serving business and society through organisational psychotherapy. Falling Blossoms. https://leanpub.com/heartsoverdiamonds

Marshall, R. W. (2021a). Memeology: Surfacing and reflecting on the organisation’s collective assumptions and beliefs. Falling Blossoms. https://leanpub.com/memeology

Marshall, R. W. (2021b). Quintessence: An acme for highly effective software development organisations. Falling Blossoms. https://leanpub.com/quintessence

Ohno, T. (1988). Toyota production system: Beyond large-scale production. Productivity Press.

Pink, D. H. (2009). Drive: The surprising truth about what motivates us. Riverhead Books.

Reinertsen, D. G. (2009). The principles of product development flow: Second generation lean product development. Celeritas Publishing.

Rogers, C. R. (1951). Client-centered therapy: Its current practice, implications, and theory. Houghton Mifflin.

Rosenberg, M. B. (2003). Nonviolent communication: A language of life (2nd ed.). PuddleDancer Press.

Seddon, J. (2019). Beyond command and control. Vanguard Consulting.

Seligman, M. E. P. (1972). Learned helplessness. Annual Review of Medicine, 23(1), 407–412. https://doi.org/10.1146/annurev.me.23.020172.002203

Senge, P. M. (2006). The fifth discipline: The art and practice of the learning organisation. Random House Business Books.

Selected Blog Posts

Marshall, R. W. (2013, October 12). The Antimatter Principle. Think Different. https://flowchainsensei.wordpress.com/2013/10/12/the-antimatter-principle/

Marshall, R. W. (2013b, October 13). The Antimatter Principle – the metaphor. Think Different. https://flowchainsensei.wordpress.com/2013/10/13/the-antimatter-principle-the-metaphor/

Marshall, R. W. (2014, January 28). A vocabulary for the Antimatter Principle. Think Different. https://flowchainsensei.wordpress.com/2014/01/28/a-vocabulary-for-the-antimatter-principle/

Marshall, R. W. (2015, August 7). Antimatter and Deming’s 95/5. Think Different. https://flowchainsensei.wordpress.com/2015/08/07/antimatter-and-demings-955/

Marshall, R. W. (2018, September 6). Solutions demand problems. Think Different. https://flowchainsensei.wordpress.com/2018/09/06/solutions-demand-problems/

Marshall, R. W. (2019b, July 21). #NoSoftware. Think Different. https://flowchainsensei.wordpress.com/2019/07/21/nosoftware/

Marshall, R. W. (2021c, July 30). Cthulhu-shaped people. Think Different. https://flowchainsensei.wordpress.com/2021/07/30/cthulhu-shaped-people/

Marshall, R. W. (2021d, June 30). 5x productivity. Think Different. https://flowchainsensei.wordpress.com/2021/06/30/5x-productivity/

Marshall, R. W. (2025a, June 10). The Antimatter Principle: Why nobody needs their needs attended to. Think Different. https://flowchainsensei.wordpress.com/2025/06/10/the-antimatter-principle-why-nobody-needs-their-needs-attended-to/

Marshall, R. W. (2025b, September 15). The Agile Manifesto: Rearranging deck chairs while five dragons burn everything down. Think Different. https://flowchainsensei.wordpress.com/2025/09/15/the-agile-manifesto-rearranging-deck-chairs-while-five-dragons-burn-everything-down/

Marshall, R. W. (2025c, August 16). A conversation about John Seddon. Think Different. https://flowchainsensei.wordpress.com/2025/08/16/a-conversation-about-john-seddon/

Marshall, R. W. (2025d, June 21). The vocabulary problem. Think Different. https://flowchainsensei.wordpress.com/2025/06/21/the-vocabulary-problem/

Marshall, R. W. (2025e, July 7). What is Organisational AI Therapy? Think Different. https://flowchainsensei.wordpress.com/2025/07/07/what-is-organisational-ai-therapy/

Marshall, R. W. (2025f, December 1). Resetting: An invitation to own what comes next. Think Different. https://flowchainsensei.wordpress.com/2025/12/01/resetting-an-invitation-to-own-what-comes-next/

Marshall, R. W. (2026a, February 4). The software quality and productivity crisis executives won’t address. Think Different. https://flowchainsensei.wordpress.com/2026/02/04/the-software-quality-and-productivity-crisis-executives-wont-address/

Marshall, R. W. (2026b, February 6). The Quintessential Mindset: 79 essential organisational memes. Think Different. https://flowchainsensei.wordpress.com/2026/02/06/the-quintessential-mindset-79-essential-organisational-memes/

Marshall, R. W. (2026c, February 19). Loon or genius? How do you respond to what you read here? Think Different. https://flowchainsensei.wordpress.com/2026/02/19/loon-or-genius-how-do-you-respond-to-what-you-read-here/

Why Organisations Can’t Examine Their Own Assumptions

The memeplex protects itself—and it’s far more sophisticated at doing so than most people realise.


Have you ever watched an organisation commission a ‘culture survey’, receive results that point to deep structural problems, and then… do absolutely nothing about it? Or perhaps you’ve seen a team sit through a retrospective, carefully dancing around the one issue everyone knows is the real problem?

This isn’t a failure of courage, though it looks like one. It’s something far more interesting—and far more difficult to address. Organisations cannot easily examine their own assumptions because the very apparatus they would use to conduct the examination is itself constructed from those assumptions. It’s like asking someone to see their own blind spot using the eye that has the blind spot.

Argyris Saw This Decades Ago

Chris Argyris spent his career documenting what he called defensive routines—the organisational habits that prevent people from examining the assumptions underlying their actions. His insight was deceptively simple and devastatingly accurate: organisations develop systematic ways of avoiding embarrassment or threat, and then develop further routines to prevent anyone from discussing the fact that they’re avoiding things.

He called this skilled incompetence. People become extraordinarily proficient at preventing learning. Not because they’re stupid or malicious, but because the social systems they inhabit reward the appearance of confidence and punish the admission of uncertainty. When a senior leader says ‘I want honest feedback’, every person in the room is running a rapid calculation about what kind of honesty is actually safe. The espoused theory is openness. The theory-in-use is self-protection.

What makes Argyris’s work so enduring is his observation that these defensive patterns are self-sealing. They contain their own protection mechanism. If you try to point out that a team is being defensive, the team will defend against that observation. If you note that the organisation avoids discussing its avoidance, that meta-observation itself becomes undiscussable. It’s undiscussability all the way down.

The Memeplex Dimension

Memeology takes this further by examining what these defensive routines are actually protecting. Organisations don’t just have scattered beliefs and assumptions floating around independently. They have memeplexes—interlocking systems of collective beliefs that reinforce one another and form a coherent yet unexamined) worldview.

Consider an organisation that holds the following beliefs simultaneously: that managers exist because workers cannot be trusted to self-organise; that detailed planning prevents failure; that individual performance metrics drive results; and that hierarchy reflects competence. None of these beliefs exists in isolation. Each one supports and is supported by the others. Together, they form a memeplex—a self-reinforcing ecology of assumptions that shapes every decision, every process, every interaction.

Now imagine someone proposes examining one of these beliefs. Say someone suggests that perhaps workers could self-organise under the right conditions. This doesn’t just challenge one assumption. It sends tremors through the entire system. If workers can self-organise, why do we need the current management structure? If the management structure isn’t necessary, what does that mean for our planning processes? If detailed planning isn’t essential, what happens to the metrics we’ve built around plan adherence? Each belief is load-bearing for every other belief. Pull one out and the people embedded in the system feel, often unconsciously, that the whole thing might collapse.

This is why surfacing collective beliefs feels so threatening. It’s not that any single assumption is too sacred to question. It’s that the assumptions form a web, and touching any strand vibrates the whole structure. The memeplex protects itself not through any conscious conspiracy but through the sheer interconnectedness of its components.

Why It Feels Like an Attack

When someone attempts genuine examination of organisational assumptions, several things happen simultaneously that make the experience feel existential rather than intellectual.

Identity is at stake. People’s professional identities are woven into the memeplex. If you’ve built a career on the assumption that rigorous planning prevents failure, questioning that assumption isn’t an abstract exercise—it’s a potential invalidation of decades of professional life. The defensive response isn’t irrational. It’s deeply human.

Social contracts are threatened. Every memeplex comes with implicit social agreements about how things work around here. Who gets to decide. Who has status. Who is considered competent and why. Examining these assumptions means potentially renegotiating every social contract in the organisation simultaneously. That’s terrifying, even if no one articulates it that way.

The familiar becomes uncertain. A memeplex, however dysfunctional, provides predictability. People know what to expect, how to behave, what will be rewarded and punished. Even an unhappy certainty often feels safer than an uncertain possibility of something better. The devil you know, as they say.

Cognitive dissonance has a half-life. Organisations that do begin to examine their assumptions often find themselves in a deeply uncomfortable intermediate state—holding two incompatible belief systems simultaneously. This state of organisational cognitive dissonance is so painful that it typically resolves within about nine months, but not always in the direction of learning. Often, the organisation snaps back to its original beliefs, having expelled or marginalised the sources of the new perspective.

The Rarity of Genuine Self-Reflection

So what makes some rare organisations capable of genuine self-reflection whilst most treat any examination as an attack?

It isn’t intelligence. It isn’t resources. It isn’t even what most people mean when they say ‘psychological safety’, though that’s closer.

The organisations that manage genuine self-reflection typically share several characteristics that work together rather than in isolation.

They have developed a practice—not just a permission—of surfacing assumptions. There’s a fundamental difference between a leader saying ‘we welcome challenge’ and an organisation that has built regular, structured practices for making the implicit explicit. Memeology as a practice isn’t about one-off workshops. It’s about cultivating an ongoing organisational habit of noticing, naming, and examining collective beliefs. The practice itself must become part of the memeplex if it is to survive.

They treat beliefs as hypotheses rather than identities. In these rare organisations, people have somehow learnt to hold their assumptions lightly—to say ‘we currently operate as though X is true’ rather than ‘X is true and I am the kind of person who believes X’. This distinction matters enormously. You can revise a hypothesis without an identity crisis. You cannot revise a core identity without such a crisis.

They attend to the therapeutic relationship. This may sound strange in an organisational context, but it’s central. Genuine self-examination requires the same conditions that effective therapy requires: voluntary participation, trust, a relationship that can hold difficult truths without rupturing. You cannot force an organisation into self-examination any more than you can force a psychotherapy patient into insight. The people facilitating this work understand that it is the quality of the relationship—not any specific technique or framework—that enables transformation.

They have sufficient safety to tolerate the in-between. These organisations have somehow built enough resilience to survive the cognitive dissonance phase—that agonising period where old assumptions have been questioned but new ones haven’t yet solidified. Most organisations panic during this phase and retreat to the familiar. The ones that don’t panic have typically built strong enough relationships, enough trust, and enough shared commitment to the process that they can sit with uncertainty long enough for something new to emerge.

They don’t try to swap individual memes. Perhaps most importantly, these organisations understand that you cannot simply replace one belief with another whilst leaving the rest of the memeplex intact. You cannot graft ‘self-organisation’ onto a command-and-control memeplex and expect it to take. The organisations that achieve genuine self-reflection understand that they’re working with a system, not a collection of independent parts.

The Paradox at the Heart of It All

Here’s what makes this work so genuinely difficult: the very capacity to examine assumptions is itself shaped by assumptions. An organisation that assumes people cannot be trusted will not trust people to examine its assumptions. An organisation that assumes expertise resides at the top will only accept examination conducted by those at the top—who, by definition, have the most invested in the current memeplex.

This is why external facilitation can help, but only if it’s the right kind. The wrong kind of external help—the kind that arrives with a diagnosis and a prescription—simply reinforces the existing memeplex by confirming the assumption that solutions come from experts. The right kind creates conditions where the organisation can see itself more clearly, reach its own insights, and choose its own path forward.

This is the core insight that connects Argyris’s defensive routines to Memeology: you cannot think your way out of assumptions using thinking that is itself assumption-laden. You need a different kind of process altogether—one that is relational rather than analytical, experiential rather than intellectual, emergent rather than prescribed.

What This Means in Practice

If you’re in an organisation and you recognise these patterns—the undiscussable topics, the defensive routines that everyone can see but no one mentions, the memeplex that shapes everything whilst remaining invisible—what can you do?

First, recognise that you cannot simply decide to fix this. The desire to ‘fix the culture’ is itself usually an expression of the existing memeplex—the belief that problems have solutions, that solutions can be implemented, that implementation is a matter of will and competence. Sometimes the most important step is to stop trying to fix and start trying to notice.

Second, find others who also notice. Not to form a revolutionary cell, but to create a small space where assumptions can be spoken aloud without immediate consequence. A space where someone can say ‘I notice that we say we value innovation but we punish every failed experiment’ and have that observation received with curiosity rather than defensiveness.

Third, understand that this is slow work. Memeplexes evolved over years or decades. They will not be transformed by a two-day offsite. The organisations that achieve genuine self-reflection have typically been building the capacity for it over a long period, often with significant help from people who understand the therapeutic dimension of organisational change.

And finally, be gentle—with yourself and with others. Defensive routines exist because people are trying to protect themselves from genuine pain. The goal isn’t to strip away defences and leave people exposed. It’s to create conditions where the defences become less necessary, because the environment has become safe enough to be honest in.

The memeplex will protect itself. That’s what memeplexes do. But with patience, skill, and a deep respect for the human beings embedded in these systems, it is possible—rare, but possible—to create the conditions where organisations can genuinely see themselves. Not as an act of will, but as an emergent property of relationships built on trust, curiosity, and a willingness to sit with the discomfort of not yet knowing what comes next.


Further Reading

Argyris, C. (1990). Overcoming organizational defenses: Facilitating organizational learning. Allyn & Bacon.

Argyris, C. (1991). Teaching smart people how to learn. Harvard Business Review, 69(3), 99–109.

Argyris, C., & Schön, D. A. (1996). Organizational learning II: Theory, method, and practice. Addison-Wesley.

Marshall, R. W. (2018). Hearts over diamonds: Serving business and society through organisational psychotherapy. Leanpub.

Marshall, R. W. (2021). Memeology: Surfacing and reflecting on the organisation’s collective assumptions and beliefs. Leanpub.

Noonan, W. R. (2007). Discussing the undiscussable: A guide to overcoming defensive routines in the workplace. Jossey-Bass.

What Would Software Development Look Like If We Started From Human Flourishing?

A positive vision for development organisations designed from first principles around attending to folks’ needs.


I’ve spent a lot of words on this blog describing what’s broken. The misaligned incentives, the learned helplessness, the quiet tragedy of talented people grinding through systems that treat them as fungible resources. If you’ve been reading along, you already know the critique.

Today I want to try something different.

I want to imagine — concretely, not abstractly — what a software development organisation would look like if we threw out every inherited assumption and started from a single foundational commitment: attend to folks’ needs.

Not as a slogan on a wall. Not as a line item in a manifesto. As the actual organising principle from which everything else follows.

The Starting Question Changes Everything

Most organisations begin with a question like: How do we deliver software efficiently? or How do we maximise output while controlling cost? These seem like reasonable questions. They’re not. They’re questions that have already smuggled in a set of assumptions about what matters, and those assumptions silently shape everything that follows — the structures, the roles, the metrics, the daily experience of every person involved.

What if we started instead with: What do the people involved in this work actually need — to thrive, to do meaningful work, to live well?

This isn’t soft. It’s radical. And the organisation it produces looks almost nothing like what we’re used to.

Whose Needs? Everyone’s.

The Antimatter Principle doesn’t play favourites. It asks us to attend to the needs of all the folks involved — not just customers, not just shareholders, not just developers, but every person whose life is touched by the work. That includes the person writing the code, the person waiting for the feature, the person answering the support call, the person who’ll maintain this system five years from now, and yes, the person funding the whole endeavour.

In practice, most organisations optimise for one constituency at the expense of others. They squeeze developers to delight customers. They squeeze customers to delight shareholders. They squeeze everyone to hit a date that somebody somewhere promised without consulting anyone who’d have to do the work.

An organisation founded on attending to folks’ needs refuses this zero-sum framing. Not because conflict disappears — it doesn’t — but because the conflicts are surfaced, named, and navigated honestly rather than buried under power dynamics and pretence.

What Changes in Practice

Let’s get concrete. Walk with me through what actually shifts.

How Work Enters the System

In most organisations, work arrives as a mandate. Someone with authority decides what gets built, packages it as a requirement or a ticket, and hands it to people whose job is to comply. The need behind the request — the real, human need — is often lost in translation, if it was ever articulated at all.

In our imagined organisation, work begins with needs. Someone has a need. Maybe it’s a customer who can’t accomplish something that matters to them. Maybe it’s a team member who sees a source of recurring pain. Maybe it’s a pattern emerging from support conversations. The need is expressed as a need, not as a premature solution. And the first act is not to estimate or prioritise but to understand — to ask, with genuine curiosity, what’s actually going on for the people involved.

This isn’t ‘requirements gathering’. Requirements gathering is an extraction process. This is something closer to dialogue — closer, in fact, to the spirit of Nonviolent Communication (Rosenberg, 2015), where the aim is to surface what’s really alive in people rather than to classify and judge. It changes the power dynamic entirely, because the people closest to the need are treated as the experts on that need — which, of course, they are.

How Teams Form and Organise

Most team structures are designed for managerial convenience — stable units that can be tracked, measured, and held accountable within a reporting hierarchy. People are assigned to teams. Teams are assigned to projects. The question of whether this arrangement serves anyone’s actual needs rarely comes up.

If we start from human flourishing, we notice that people have a need for autonomy, for mastery, for purpose, for belonging, for psychological safety. They have a need to work with people they trust on problems they find meaningful. They have a need to grow, and growth doesn’t happen on a schedule that aligns with annual review cycles.

So teams become more fluid. People gravitate towards work that connects to their sense of purpose and where their skills create the most value — for themselves and others. This isn’t chaos. It’s self-organisation around real need, and it requires more trust, more communication, and more maturity than a command-and-control structure. But it produces something a command-and-control structure never can: genuine engagement.

How Decisions Get Made

In a conventional organisation, decisions flow down. Strategy is set at the top, decomposed into plans, and cascaded as instructions. The people doing the work make the fewest decisions about the work. This arrangement persists not because it’s effective — it’s demonstrably not — but because it satisfies the needs of the people at the top for control and predictability. Their needs are attended to. Everyone else’s are not.

An organisation grounded in attending to folks’ needs distributes decision-making to where the relevant knowledge lives. Not because decentralisation is ideologically appealing, but because it attends to more people’s needs simultaneously. The developer who understands the technical trade-offs, the support person who hears the customer’s frustration daily, the designer who’s observed the user struggling — all of these people have knowledge that’s essential to a good decision. Excluding them isn’t just inefficient. It’s a failure to attend to their need to contribute meaningfully.

This doesn’t eliminate leadership. It redefines it. Leaders in this world are not decision-makers but need-discoverers — people whose primary skill is sensing what’s needed across the whole system and helping the right conversations happen.

How Quality Is Understood

In most organisations, quality is defined as conformance to specification, or as the absence of defects, or — in more sophisticated shops — as fitness for purpose. All of these framings treat quality as a property of the product.

If we start from human flourishing, quality becomes a property of the relationship between the product and the people it touches. Does this software help someone do something that matters to them? Does using it feel respectful of their time and intelligence? Does building it feel like craft rather than compliance? Does maintaining it feel manageable rather than dread-inducing?

Quality, understood this way, can’t be tested in at the end. It can’t be enforced by a separate QA department. It emerges from a development process in which the people doing the work are themselves flourishing — because people who are stressed, disengaged, and treated as interchangeable parts do not produce things that feel cared-for. They can’t. The care isn’t there to transmit.

How Success Is Measured

Here’s where it gets genuinely uncomfortable for most organisations, because the metrics we’re accustomed to — velocity, throughput, cycle time, story points, lines of code, uptime, revenue per employee — all measure the machine. They tell you how the system is performing as a production apparatus. They tell you nothing about whether anyone involved is flourishing.

An organisation built on attending to folks’ needs would ask different questions. Are the people doing this work learning and growing? Are customers’ lives meaningfully better? Are we building trust or eroding it? Are people choosing to stay because they want to, or staying because they’re afraid to leave? Is the work sustainable — not just this sprint, but this year, this decade?

Some of these things can be measured. Most of them can only be sensed — through honest conversation, through the quality of relationships, through the presence or absence of joy in the work. An organisation that can’t tolerate this ambiguity, that demands everything be reduced to a number on a dashboard, has already revealed which needs it’s prepared to ignore.

The Objection You’re Already Forming

‘This sounds lovely, but it wouldn’t survive contact with economic reality.’

I hear you. And I’d push back — gently — on two fronts.

First, attending to folks’ needs includes attending to the need for the organisation to be economically viable. No one’s needs are served by an organisation that goes bankrupt. Financial sustainability isn’t opposed to human flourishing; it’s a precondition for it. The question isn’t whether the organisation needs to generate revenue and manage costs. Of course it does. The question is whether financial performance is the purpose or a constraint — whether people exist to serve the numbers, or the numbers exist to serve the people.

Second, there’s mounting evidence — from the research on psychological safety (Edmondson, 2019), on intrinsic motivation (Pink, 2009), on the economics of flow (Csikszentmihalyi, 1990/2008), on the cost of turnover and disengagement (DeMarco & Lister, 2013) — that organisations which attend to human needs outperform those that don’t. Not despite the attention to flourishing, but because of it. People who are well do good work. This shouldn’t be surprising, but apparently it is.

Why This Matters Now

We’re at an interesting inflection point. The tools available to software developers are more powerful than ever. AI is reshaping what’s possible. Remote work has rewritten the geography of collaboration. And yet — or perhaps because of all this — the human experience of building software remains, for many people, somewhere between unfulfilling and actively harmful.

The response from most of the industry is to double down on what’s familiar: more process, more measurement, more frameworks, more management, more control dressed up as empowerment. SAFe. OKRs. Spotify models without Spotify’s culture. The names change. The underlying collective assumptions and beliefs don’t.

What I’m describing here isn’t another framework. It’s a foundation. A single principle — attend to folks’ needs — from which appropriate practices, structures, and norms can emerge, shaped by the specific people in the specific context. It won’t look the same in every organisation, and that’s the point. It can’t be a framework because frameworks are, by definition, imposed from outside, and imposition is itself a failure to attend to needs.

An Invitation

If you’ve made it this far, you’re probably someone who’s felt the gap between how software development could feel and how it does feel. You may have had glimpses — a team that clicked, a project where the work felt alive, a moment when building something actually felt like building something.

Those glimpses aren’t anomalies. They’re signals. They’re what happens when, even accidentally, an environment arises that attends to the needs of the people within it.

The question I’d leave you with is this: What would it take to make that the rule rather than the exception? Not in theory. In your context, with your people, starting tomorrow.

I don’t pretend to have the full answer. But I’ve seen that it starts with asking a better question than ‘how do we deliver software efficiently?’ It starts with asking what people need — and then taking the answer seriously.


I’d love to hear what resonates — and what doesn’t. What needs of yours aren’t being attended to in your current organisation? What would change first if they were? The comments are open, and as always, I’m happy to let your questions shape where this conversation goes next.


Further Reading

Csikszentmihalyi, M. (2008). Flow: The psychology of optimal experience. Harper Perennial Modern Classics. (Original work published 1990)

DeMarco, T., & Lister, T. (2013). Peopleware: Productive projects and teams (3rd ed.). Addison-Wesley Professional.

Edmondson, A. C. (2019). The fearless organization: Creating psychological safety in the workplace for learning, innovation, and growth. John Wiley & Sons.

Marshall, R. W. (2019). Hearts over diamonds: Serving business and society through organisational psychotherapy. Falling Blossoms.

Marshall, R. W. (2021a). Memeology: Surfacing and reflecting on the organisation’s collective assumptions and beliefs. Falling Blossoms.

Marshall, R. W. (2021b). Quintessence: An acme for highly effective software development organisations. Falling Blossoms.

Pink, D. H. (2009). Drive: The surprising truth about what motivates us. Riverhead Books.

Rosenberg, M. B. (2015). Nonviolent communication: A language of life (3rd ed.). PuddleDancer Press.

Sheridan, R. (2013). Joy, Inc.: How we built a workplace people love. Portfolio/Penguin.

Breaking the Spell: Alternatives to Our Ingrained Collective Assumptions

We swim in assumptions the way fish swim in water — mostly without noticing. The beliefs we inherit from culture, institutions, and one another form an invisible architecture around our thinking. They arrive unlabelled, feel like common sense, and quietly shape everything from how we organise work to how we understand progress, success, and even reality itself.

But what if many of these default settings are not just incomplete, but actively limiting? And more importantly, what if there were practical ways to remember — and reach for — alternatives? Many assume they’re stuck with those assumptions — a limiting belief in and of itself.

That is where two underappreciated conceptual frameworks come in: memeology and quintessence (Leanpub books). Both function as aide memoires — structured reminders that make it easier to step outside the grooves of conventional thinking when it matters.

The Problem with Default Assumptions

Collective assumptions are powerful precisely because they do not announce themselves. They operate as background code. Consider a few that pervade modern Western culture: growth is inherently good; efficiency should be maximised; competition drives excellence; more information leads to better decisions; individual achievement is the primary unit of success.

None of these are necessarily wrong. But each carries a shadow — an unexamined cost. The assumption that growth is inherently good, for instance, makes it almost impossible to have a serious public conversation about sufficiency, equilibrium, or the wisdom of knowing when enough is enough. The belief that competition drives excellence blinds us to the vast amount of human achievement rooted in cooperation, gift economies, and intrinsic motivation.

The difficulty is not in intellectually critiquing these assumptions. Most thoughtful people can do that over a cup of tea. The difficulty is in remembering to question them in the moment — when you are drafting a strategy, designing a policy, raising a child, or simply deciding what a good life looks like.

This is the gap that memeology and quintessence are designed to fill.

Memeology: Mapping the Ideas That Think Us

Richard Dawkins coined the term ‘meme’ in 1976 to describe cultural units of information — ideas, behaviours, phrases, and symbols — that replicate and spread through populations in ways loosely analogous to genes (Dawkins, 1976). Memeology, as a broader practice, takes this insight and turns it into something actionable: a discipline of noticing which ideas have colonised our thinking, where they came from, and whether they actually serve us. Susan Blackmore (1999) extended this work considerably, arguing that memetic selection shapes our minds and cultures in ways just as profound as natural selection shapes our bodies.

When used as an aide memoire, memeology invites a set of reflexive questions.

What meme is operating here? Every time we encounter a ‘that is just how things are’ moment, memeology prompts us to name the underlying belief — to drag it out of the background and into the light. The act of naming is itself a kind of liberation. An assumption you can see is an assumption you can evaluate.

Where did this meme originate, and what conditions did it evolve to suit? Many of our most entrenched beliefs are adaptations to contexts that no longer exist. The meme of relentless productivity, for example, maps neatly onto the demands of early industrialisation. Whether it maps onto a world of automation, ecological limits, and widespread burnout is a different question entirely.

What alternative memes are available? This is the crucial move. Memeology does not just deconstruct — it curates. It asks us to actively seek out the road-not-taken ideas: the cultures that prize rest over hustle, the traditions that measure wealth in relationships rather than accumulation, the philosophies that treat uncertainty as a feature rather than a bug.

By building a personal and collective repertoire of alternative memes, we create a kind of mental toolkit — a library of different lenses we can reach for when the default ones are not working. Daniel Dennett (1995) described this kind of evolutionary thinking as a ‘universal acid’ — an idea so powerful it eats through virtually every traditional concept and leaves a revolutionised world-view in its wake.

Quintessence: Distilling What Truly Matters

If memeology helps us map the landscape of ideas, quintessence helps us navigate it by asking a deceptively simple question: what is the essence of the thing?

The word ‘quintessence’ has ancient roots. In classical and medieval philosophy, it referred to the fifth element — the substance beyond earth, water, air, and fire that composed the heavenly bodies. Aristotle reasoned that, whilst the four terrestrial elements were mutable and corruptible, the celestial realm must be composed of a different, unchangeable substance, which he called aether (Aristotle, 350 BCE/1922). By the medieval period, this concept had become known as quinta essentia — the fifth essence — and was pursued by alchemists as the purest, most fundamental form of any substance, stripped of everything incidental.

As an aide memoire for challenging default assumptions, quintessence works by cutting through the noise of convention to ask: what is actually essential here, as opposed to what we have simply layered on through habit?

Consider how this applies in practice.

The quintessence of education is not credentialing, assessment frameworks, or seat time in classrooms. It is the transformation of a person’s capacity to understand, to think, and to act wisely in the world. When we hold that essence in mind, many of our default assumptions about what education ‘must’ look like begin to dissolve. Suddenly, a year spent travelling, apprenticing, or simply reading deeply looks less like a gap and more like a direct route.

The quintessence of healthcare is not the delivery of medical interventions. It is the flourishing of human health and wellbeing. This reframing immediately surfaces alternative assumptions — about prevention over treatment, about community and belonging as health interventions, about the limits of a purely biomedical model.

The quintessence of work is not employment. It is the application of human energy and creativity towards something that matters. Hold that distinction clearly and the entire conversation about automation, universal basic income, and the future of labour shifts on its axis.

Quintessence, used as a regular mental practice, acts as a solvent for accumulated convention. It does not tell you what to believe instead. It clears the ground so that better questions can emerge.

Using Them Together

Memeology and quintessence are most powerful in combination. Memeology gives you the diagnostic: here is the belief that is operating, here is where it came from, and here are alternatives. Quintessence gives you the compass: but what are we actually trying to achieve here, at the deepest level?

Together, they form a two-step practice for anyone — leaders, educators, designers, parents, citizens — who suspects that the way things are is not the way things have to be.

Step one: Name the meme. What is the inherited assumption shaping this situation? Say it aloud. Write it down. Make it visible.

Step two: Find the quintessence. Beneath the accumulated habits and expectations, what is the real purpose, the irreducible core? What would you design if you started from that essence rather than from precedent?

The space between those two steps is where genuine alternatives live. Not utopian fantasies, but practical, grounded re-imaginings rooted in clearer sight. Donella Meadows (2008) understood this well when she identified paradigm shifts — the shared ideas and assumptions from which a system arises — as among the most powerful leverage points for changing any system’s behaviour.

A Few Alternative Assumptions Worth Carrying

To close, here are some alternative assumptions — counter-memes, if you like — that both memeology and quintessence tend to surface. None are presented as truths. All are presented as possibilities worth remembering.

Sufficiency over maximisation. What if the goal is not more, but enough — elegantly, sustainably, and with room to breathe? Kate Raworth (2017) has made a compelling case that thriving within boundaries, rather than pursuing limitless growth, is the defining economic challenge of our century.

Coherence over consistency. What if wisdom lies not in rigid adherence to a single framework, but in the ability to hold multiple perspectives and respond to what each situation actually requires?

Emergence over control. What if the best outcomes arise not from tighter management, but from creating the right conditions and trusting what unfolds?

Interdependence over independence. What if autonomy is not the highest value, but one value in a web that also includes belonging, obligation, and mutual care?

Presence over productivity. What if attention — the quality of being fully here — is the scarcest and most valuable resource we have?

These are not new ideas. Many are ancient. But in a culture saturated with competing demands on our attention, having structured ways to remember them — aide memoires like memeology and quintessence — may be exactly what is needed.

The default assumptions will always be loud. The alternatives can be accessible.


The invitation is not to reject all inherited beliefs wholesale, but to hold them more lightly — to recognise them as choices rather than facts, and to remember that other choices are always available.


Further Reading

Aristotle. (350 BCE/1922). De caelo [On the heavens] (J. L. Stocks, Trans.). Clarendon Press.

Blackmore, S. (1999). The meme machine. Oxford University Press.

Dawkins, R. (1976). The selfish gene. Oxford University Press.

Dennett, D. C. (1995). Darwin’s dangerous idea: Evolution and the meanings of life. Simon & Schuster.

Kuhn, T. S. (1962). The structure of scientific revolutions. University of Chicago Press.

Marshall, R. W. (2021a). Memeology: Surfacing and reflecting on the organisation’s collective assumptions and beliefs. Falling Blossoms.

Marshall, R. W. (2021b). Quintessence: An acme for highly effective software development organisations. Falling Blossoms.

Meadows, D. H. (2008). Thinking in systems: A primer (D. Wright, Ed.). Chelsea Green Publishing.

Raworth, K. (2017). Doughnut economics: Seven ways to think like a 21st-century economist. Random House Business Books.

 

The Ups and Downs of Ingroups and Outgroups in the Age of AI

We have always sorted ourselves into groups. It is one of the oldest human instincts — the pull towards ‘us’ and the wariness of ‘them’. Tribes, nations, fandoms, professional circles — the boundaries shift, but the impulse does not. Now, as artificial intelligence reshapes how we work, relate, communicate, and make decisions, the dynamics of ingroups and outgroups are being rewritten in ways we are only beginning to understand.

Some of what is emerging is genuinely promising. Some of it might give us pause.

Note: NotebookLM infographic feature is in Beta and still buggy Infographic

The New Lines Being Drawn

AI is creating fresh categories of belonging:

  • The people who understand the tools and those who do not.
  • The companies that have adopted generative AI into their workflows and the ones still debating whether to try it.
  • The developers building these systems and the end users navigating their outputs.
  • Those who passively accept the constraints of limiting beliefs and those who choose to actively do something about them.

These are not trivial distinctions. Access to AI — and fluency with it — is quickly becoming a marker of professional and social status. Knowing how to prompt a model effectively, how to integrate AI into a creative process, or how to audit an algorithm’s outputs are emerging as forms of cultural capital. Those who have it form a kind of ingroup, often without meaning to. Those who lack it can find themselves edged out of conversations, hiring pipelines, and opportunities.

The Upside: Walls Coming Down

Here is where things get interesting: AI has a remarkable capacity to dissolve traditional ingroup–outgroup boundaries.

Language barriers — one of the most ancient dividers between human groups — are eroding. Real-time translation tools are making it possible for people to collaborate across languages in ways that would have been impractical five years ago. A developer in Lagos can pair-programme with a colleague in Seoul without either needing to be fluent in the other’s language.

Knowledge gatekeeping is weakening too. Fields that once required years of specialised education to even enter the conversation — law, medicine, finance, software engineering — are becoming more navigable for outsiders. You can now ask an AI to explain a dense legal filing in plain language or walk you through the logic of a complex codebase. The ingroup of ‘people who understand this’ is expanding, and for the most part, that is a good thing.

AI can also help surface and counteract the biases that have historically reinforced outgroup exclusion. When hiring algorithms are carefully designed and audited, they can reduce the influence of name, gender, and background on who gets a callback. When recommendation systems are built thoughtfully, they can expose people to perspectives and communities outside their usual bubbles.

The key word, of course, is ‘carefully’.

The Downside: Old Patterns, New Infrastructure

Here is the uncomfortable truth: AI does not just have the potential to reduce bias. It has an equal capacity to scale it.

When training data reflects historical patterns of exclusion, the models trained on that data will reproduce those patterns — often at a speed and scale that no individual human decision-maker ever could. A biased hiring manager affects one company. A biased hiring algorithm affects thousands. The outgroup does not just stay on the outside; the door gets locked with industrial-grade efficiency.

Algorithmic recommendation systems, meanwhile, can harden ingroup boundaries in subtler ways. Social media platforms powered by AI learn what keeps you engaged and then serve you more of it. The result, frequently, is a feedback loop that reinforces your existing identity and worldview whilst making the outgroup seem increasingly alien, unreasonable, or threatening. The algorithm does not create tribalism, but it can pour fuel on it.

There is also the emerging divide between those who are replaced by AI and those who are augmented by it. This is not evenly distributed. Roles that involve routine cognitive tasks — data entry, basic analysis, first-draft writing — are being automated fastest, and those roles are disproportionately held by people earlier in their careers or in lower-income brackets. The ingroup of ‘people whose work AI makes more valuable’ and the outgroup of ‘people whose work AI makes redundant’ is a split with enormous economic and social consequences.

The Deeper Divide: Comfort vs. Reflection

There is another ingroup–outgroup split emerging in the age of AI, and it may be the most consequential of all. It is not about who has access to the tools or who understands how they work. It is about what people are willing to do with what the tools reveal.

Every group — every team, organisation, community, or culture — operates on a set of shared assumptions and beliefs. These are the unspoken agreements about how things work, what matters, and what is true. They are the water the fish does not notice. And AI, for all its flaws, has a peculiar talent for making that water visible. Feed a model your organisation’s data and it will reflect back patterns you may not have intended or even recognised: who gets promoted and who does not, which ideas get traction and which get ignored, what language is rewarded and what language is penalised.

This is where the divide opens up. On one side are the groups — and the individuals within them — who accept the limitations of their existing shared assumptions and beliefs. They treat AI outputs as confirmation of how things are, or dismiss inconvenient patterns as noise. The model says what they expected, so it must be right. Or it says something uncomfortable, so it must be wrong. Either way, the assumptions remain intact.

On the other side are those who are prepared to surface and genuinely reflect on those shared assumptions. They treat AI not as an oracle but as a diagnostic tool — one that can hold up a mirror to collective blind spots, institutional habits, and inherited ways of thinking. These groups ask harder questions: Why does the model see this pattern? What does it say about us that it learned this from our data? What have we been taking for granted?

This is not a comfortable process. Surfacing assumptions never is. It means admitting that your ingroup’s way of seeing the world is partial, shaped by history and circumstance rather than objective truth. It means sitting with the possibility that the beliefs binding your group together might also be the beliefs keeping others out.

But this willingness to reflect is where AI’s real transformative potential lives. Not in automating tasks or generating content, but in giving groups a reason — and a tool — to examine the foundations they have built on. The groups that embrace this will adapt, learn, and build more inclusive cultures. The groups that resist it will find their assumptions calcifying, made all the more rigid by AI systems trained to mirror them back unchallenged.

The irony is sharp: AI can either deepen groupthink or disrupt it, and the deciding factor is not the technology. It is whether the humans using it are brave enough to let it show them what they would rather not see.

The Ingroup That Decides for Everyone

Perhaps the most consequential ingroup–outgroup dynamic of the AI age is the one between the people building these systems and the people affected by them. The teams designing AI models, choosing training data, setting safety parameters, and defining what counts as ‘aligned’ behaviour are relatively small and strikingly homogeneous compared to the global population their decisions touch.

This is not a new problem — technology has always been shaped by its creators — but the stakes are different now. The decisions made by a small group of researchers and executives ripple outward to influence hiring, healthcare, criminal justice, education, and creative expression for billions of people. When the ingroup making those decisions lacks diversity of experience, the blind spots can be enormous.

The push for more inclusive AI development — broader teams, more diverse training data, wider public input on governance — is essentially an effort to bring the outgroup inside the room where the decisions are made. It is slow, imperfect work, but it matters.

Living With the Tension

There is no clean resolution here. AI is simultaneously the most powerful tool we have ever had for breaking down barriers between groups and the most powerful tool we have ever had for reinforcing them. Which outcome we get depends less on the technology itself than on the choices made by the people deploying it — and the willingness of society to hold those people accountable.

A few things seem clear. Expanding AI literacy broadly, rather than letting it concentrate in a narrow technical elite, is essential. Demanding transparency in how algorithms sort, rank, and recommend — and who benefits from those decisions — is non-negotiable. And maintaining a healthy scepticism about any tool that promises to be ‘objective’ or ‘neutral’ is just good sense. Algorithms are built by people, trained on human history, and deployed in human institutions. They inherit our patterns, including the ones we would rather not pass on.

The age of AI has not invented the ingroup–outgroup problem. But it has given us new ways to either entrench it or transcend it. The question is which instinct we will follow — and whether we will build systems that help us choose the better one.


Further Reading

Gonzales, S. (2025, September 23). AI literacy and the new digital divide: A global call for action. UNESCO Global AI Ethics and Governance Observatory. https://www.unesco.org/ethics-ai/en/articles/ai-literacy-and-new-digital-divide-global-call-action

Marshall, R.W. (2025, July 7). What is Organisational AI Therapy? Think Different. https://flowchainsensei.wordpress.com/2025/07/07/what-is-organisational-ai-therapy/

Hu, T., Kyrychenko, Y., Rathje, S., Collier, N., van der Linden, S., & Roozenbeek, J. (2025). Generative language models exhibit social identity biases. Nature Computational Science, 5(1), 65–75. https://doi.org/10.1038/s43588-024-00741-1

Molenberghs, P., & Louis, W. R. (2018). Insights from fMRI studies into ingroup bias. Frontiers in Psychology, 9, Article 1868. https://doi.org/10.3389/fpsyg.2018.01868

O’Neil, C. (2016). Weapons of math destruction: How big data increases inequality and threatens democracy. Crown.

Pariser, E. (2011). The filter bubble: What the Internet is hiding from you. Penguin Press.

Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Brooks/Cole.

Wilson, K., Caliskan, A., et al. (2024, October 22). AI tools show biases in ranking job applicants’ names according to perceived race and gender [Conference paper]. AAAI/ACM Conference on Artificial Intelligence, Ethics, and Society, San Jose, CA, United States. https://www.washington.edu/news/2024/10/31/ai-bias-resume-screening-race-gender/

AI Won’t Save Your Dysfunctional Organisation

In which we discover that feeding your dragons a diet of large language models only makes them larger


I’ve spent the last year watching organisations do to AI exactly what they did to Agile: adopt the tools, mimic the rituals, and utterly miss the point.

The pattern is so consistent it’s almost comforting. Executive reads breathless article about AI transformation. Consultant arrives with slide deck promising 40% productivity gains. Pilot project launches with great fanfare. Six months later, developers are using ChatGPT to generate boilerplate code they could write in their sleep, whilst the actual problems—the ones that made your organisation mediocre before AI—remain magnificently untouched.

Here’s the uncomfortable truth I keep delivering to clients who don’t want to hear it: AI is an amplifier, not a solution. It will make your effective teams more effective. It will make your dysfunctional teams more dysfunctional. It will not—cannot—fix the underlying pathologies that make your organisation what it is.

In my previous post on the Agile Manifesto, I introduced the Five Dragons: the deep organisational dysfunctions that no methodology can slay. The Manifesto’s principles aren’t wrong; they’re simply irrelevant to these dragons. They’re instructions for sailing whilst your ship is on fire.

AI is the same story, different decade. Let me show you how each dragon devours your shiny new AI initiatives—and why no amount of prompt engineering will save you.


Dragon #1: The Human Motivation Death Spiral Meets AI

Your developers have mentally checked out. The work is meaningless. Their contributions are invisible. They lack autonomy. There’s no sense of mastery or purpose. They’re showing up, collecting paycheques, and waiting for something—retirement, a better offer, the heat death of the universe—to release them from this purgatory.

Now you’ve given them AI coding assistants.

What happens next is entirely predictable: they use the AI to do even less thinking. Not because they’re lazy—because thinking was the last thing connecting them to their work, and you’ve now automated that away too. The developer who once took quiet pride in an elegant solution now accepts whatever the LLM generates, makes minimal edits, and ships it. Engagement drops further. The death spiral accelerates.

I’ve seen this in client after client. ‘Our developers love the AI tools!’ management reports. Yes, they do. The same way a burned-out employee loves anything that reduces the cognitive load of a job they’ve already mentally quit. The AI isn’t increasing productivity; it’s enabling a more comfortable form of disengagement.

The real question you’re avoiding: Why are your people checked out in the first place? What would make this work meaningful? Until you answer that, AI just helps people coast faster toward wherever coasting leads.


Dragon #2: Dysfunctional Relationships That Poison Everything—Now With AI

Trust deficits. Ego warfare. Passive aggression. Fear. Status games. The political dynamics that turn every meeting into a performance and every decision into a negotiation. You know the drill.

AI doesn’t neutralise these dynamics. AI becomes a new weapon in them.

Watch what happens when you introduce AI tools into a team with trust issues:

  • The senior developer who feels threatened starts publicly ‘correcting’ AI-generated code from junior developers, establishing dominance through a proxy
  • The passive-aggressive architect uses AI to generate elaborate documentation that technically addresses requirements whilst being utterly useless, then points to the AI’s involvement as a shield against criticism
  • Teams weaponise AI-generated estimates against each other: ‘The AI says this should take two days, why is your team saying two weeks?’
  • Credit becomes even murkier—who gets recognised when half the code came from an LLM? The politics of attribution intensify

The most insidious pattern I’ve observed: AI as relationship avoidance. Instead of having the difficult conversation with a colleague, you ask the AI to draft the message. Instead of working through a design disagreement face-to-face, you each generate AI-supported arguments and email them back and forth. The technology becomes a buffer that prevents the very interactions that might—might—eventually heal the dysfunction.

The real question you’re avoiding: Why don’t your people trust each other? What happened—or keeps happening—to make psychological safety impossible here?


Dragon #3: Shared Delusions and Toxic Assumptions—Now at Scale

This is where AI gets genuinely dangerous.

Your team already operates under collective fictions. Reality distortion about your product’s market fit. Capability myths about what you can deliver and when. Quality blindness that mistakes functional for good. You’re building for imaginary users based on assumptions nobody has tested because testing them might reveal that the last three years were a mistake.

Traditional delusion has a natural limiting factor: the effort required to produce artefacts. It takes time to write a fantasy roadmap. It takes energy to create a specification for features nobody needs. Human bandwidth constrains how much organisational fiction you can generate.

AI removes that constraint.

Now you can generate detailed product requirements for your imaginary users in minutes. You can produce elaborate technical documentation for your reality-distorted architecture in an afternoon. You can create beautiful slide decks that make your shared delusions look thoroughly researched and carefully considered. The AI will even help you find data—or data-shaped objects—that support your existing beliefs.

I recently watched a team use AI to generate a comprehensive competitive analysis. The document was impressive: well-structured, extensively detailed, professionally formatted. It was also built entirely on the team’s flawed assumptions about their market position, which the AI had helpfully elaborated into fifty pages of confident wrongness. The analysis didn’t challenge a single premise. It couldn’t—it was trained on the prompts they gave it, which encoded their delusions as axioms.

When the entire team believes something factually wrong, AI helps them believe it more professionally.

This is the epistemic nightmare scenario: AI as confirmation bias industrialised. The tool is exquisitely good at giving you more of what you asked for, and your dysfunctional organisation asked for validation of its existing worldview.

The real question you’re avoiding: What do you believe that might not be true? When did you last seriously test your core assumptions? Who in your organisation is rewarded for pointing out that the emperor has no clothes?


Dragon #4: The Management Conundrum—Now Featuring AI Theatre

In my original piece, I questioned why management exists at all in software development. Managers are overhead—people who don’t understand the work making decisions about the work. The argument for their existence rests on coordination, resource allocation, and organisational navigation. But what if those functions don’t require a dedicated class of non-practitioners?

AI is making this question more urgent, not less—but not in the way you’d expect.

The optimistic narrative says AI will eliminate middle management by automating coordination and reporting. The pessimistic narrative says AI will eliminate developers and leave only management. What I’m actually seeing is neither: AI is creating an entire new category of management theatre.

Managers who added questionable value before are now adding questionable value through AI. They generate AI-assisted status reports synthesised from updates they don’t understand. They produce AI-created dashboards that visualise metrics they can’t interpret. They send AI-drafted communications that sound informed whilst conveying nothing. The appearance of management has become more sophisticated whilst its substance remains unchanged.

Meanwhile, the actual coordination problems—the deep organisational dysfunction that makes collaboration difficult—remain untouched. No AI is resolving the political conflicts between your VP of Engineering and your VP of Product. No LLM is fixing the incentive structures that cause teams to optimise locally at the expense of the whole. The dragons keep burning whilst management gets better at generating smoke-obscuring documentation.

The most tragicomic pattern: managers using AI to appear technical enough to manage AI initiatives. The blind leading the blind, but now with generated credentials.

The real question you’re avoiding: What would happen if you eliminated half your management layers tomorrow? What coordination problems would actually emerge versus which ones are management-created in the first place?


Dragon #5: Opinioneering Goes Industrial

Opinioneering—strong opinions held without sufficient – or any – evidence, violating Clifford’s ethics of belief—might be AI’s most fertile breeding ground.

Before AI, opinioneering required effort. You had to construct your evidence-free arguments manually. You had to find your own sources to misinterpret. You had to write your own blog posts advocating for technical decisions based on conference talks you half-understood.

AI demolishes these barriers.

Now every unfounded opinion can be instantly elaborated into a detailed position paper. Every cargo-culted ‘best practice’ can be supported by AI-generated rationales. Every process folklore can be documented as if it were rigorously derived. The opinions haven’t improved—they’ve just become more verbose and more difficult to challenge because challenging them now means engaging with walls of generated text.

I watched a technical architect use AI to produce a forty-page justification for a technology choice he’d already made based on a vendor dinner and a convincing sales pitch. The document cited papers (which existed, I checked), referenced case studies (which were real, if not quite applicable), and constructed arguments (which were logical, if you accepted the premises). It was beautiful opinioneering. The conclusion was predetermined; AI just built an impressive path to reach it.

This is the epistemological danger: AI makes it trivially easy to construct post-hoc rationalisations for positions we hold on non-rational grounds. The ethics of belief—the principle that we should proportion our confidence to our evidence—gets harder to maintain when generating evidence-shaped content costs nothing.

The real question you’re avoiding: How do you actually know what you think you know? What’s your process for distinguishing evidence from sophisticated rationalisation? When did you last change your mind about something important because data contradicted your beliefs?


The Meta-Pattern: AI as Organisational Denial

Underneath all five dragons lurks a common theme: AI enables organisations to avoid confronting their real problems.

Motivation crisis? AI makes disengagement more comfortable. Relationship dysfunction? AI provides new avoidance mechanisms. Shared delusions? AI industrialises confirmation bias. Management overhead? AI creates new theatre to justify existing roles. Opinioneering? AI elaborates unfounded opinions into documented positions.

The pattern is identical to what happened with Agile. Organisations adopted the ceremonies—standups, sprints, retrospectives—without addressing the underlying dysfunction that made work miserable. The practices became a performance that deflected attention from the real problems.

AI adoption is following the same script. The technology becomes another layer of sophisticated avoidance. We’re not fixing the organisation; we’re just failing more efficiently.


So What Actually Helps?

If you’ve read this far hoping for a neat solution, I’m going to disappoint you. The dragons aren’t slain with tools—Agile, AI, or whatever comes next. They’re slain with the difficult work of organisational change that nobody wants to do because it requires examining uncomfortable truths.

But here’s what I can offer:

Stop asking ‘How can AI make us more productive?’ Start asking ‘What’s actually preventing us from being effective?’ The answer is almost never ‘insufficient automation’.

Before any AI initiative, name your dragons. Explicitly. Out loud. In a room with people who can do something about them. If your organisation can’t even have that conversation, you’re not ready for AI—you’re ready for therapy.

Use AI as a diagnostic, not a solution. How your organisation adopts AI will reveal its pathologies. Watch who’s using it to avoid rather than engage. Notice which dysfunctions it amplifies. The patterns will tell you what’s actually wrong.

Accept that technology is never the bottleneck. Not really. The bottleneck is always human: motivation, trust, shared reality, organisational structure, epistemic hygiene. These are solvable problems, but not with better prompts.


I started this practice calling myself an ‘Organisational AI Therapist’ as a half-joke. It’s become less funny as I’ve watched organisation after organisation pursue AI transformation whilst assiduously avoiding the transformations that would actually matter.

The AI won’t save your dysfunctional organisation. It will, however, give you increasingly sophisticated ways to pretend the dysfunction isn’t there.

The dragons are patient. They’ve been waiting through Waterfall, through Agile, through DevOps. They’ll wait through AI too. They know that eventually, when the hype fades and the consultants move on, you’ll still have to face the same hard questions you’ve been avoiding all along:

Why don’t your people care? Why don’t they trust each other? What are you collectively pretending is true? Why do you have so many people who don’t do the work managing people who do? What do you believe without evidence?

The AI can’t answer these questions for you.

But I suppose it can help you write a very convincing document explaining why you don’t need to ask them.


If this resonated—or made you uncomfortable—I’m available for workshops, consulting, and the kind of organisational conversations that actually matter. Sometimes the most valuable thing I do is sit in a room and ask questions that everyone has been avoiding. The dragons hate that.


Further Reading

Clifford, W. K. (1877). The ethics of belief. Contemporary Review, 29, 289–309.

Dikert, K., Paasivaara, M., & Lassenius, C. (2016). Challenges and success factors for large-scale agile transformations: A systematic literature review. Journal of Systems and Software, 119, 87–108. https://doi.org/10.1016/j.jss.2016.06.013

Kim, B.-J., Kim, M.-J., & Lee, J. (2025). The dark side of artificial intelligence adoption: Linking artificial intelligence adoption to employee depression via psychological safety and ethical leadership. Humanities and Social Sciences Communications, 12, Article 704. https://doi.org/10.1057/s41599-025-05040-2

Li, J., Zhu, F., & Hua, P. (2025, November). Overcoming the organizational barriers to AI adoption. Harvard Business Review. https://hbr.org/2025/11/overcoming-the-organizational-barriers-to-ai-adoption

Neumann, M., Kuchel, T., Diebold, P., & Schön, E.-M. (2024). Agile culture clash: Unveiling challenges in cultivating an agile mindset in organisations. arXiv preprint. https://arxiv.org/abs/2405.15066

Pink, D. H. (2009). Drive: The surprising truth about what motivates us. Riverhead Books.

Stanford Encyclopedia of Philosophy. (2010, June 14). The ethics of belief. https://plato.stanford.edu/entries/ethics-belief/

The Chinese Advantage

There is a damning pattern playing out across global industry right now, and most Western companies are too indifferent to notice it. While Chinese manufacturers push boundaries, pack in features, and iterate at blistering speed, their Western counterparts do just enough. Enough to ship. Enough to satisfy the quarterly earnings call. Enough to not get fired. And that gap between ‘enough’ and ‘everything we can’ is quietly reshaping who leads — and who follows — in sector after sector.

This is not about cheap labour or government subsidies. It is about ambition. Chinese manufacturers behave as though every product launch is a fight for survival. Western companies behave as though their history alone will carry them through. One of those approaches is winning. It is not the Western one.

The ‘Good Enough’ Trap

Western manufacturing has drifted into a culture of sufficiency. Products are built to meet specifications, not to exceed them. Features are held back for next year’s model to justify the upgrade cycle. Engineering teams are constrained by risk-averse management layers, and innovation is filtered through committees that prize predictability over ambition.

Look at the smartphone market. For years, Apple and Samsung have delivered incremental annual updates — a slightly better camera here, a marginally faster chip there — while charging ever-higher prices. Meanwhile, Chinese manufacturers like Xiaomi, Vivo, and OPPO have been cramming their devices with 7,500mAh silicon-carbon batteries, 200-megapixel periscope telephoto lenses, 144Hz displays, and charging speeds that take a phone from 10 per cent to 80 per cent in roughly seven minutes. These are not flagship-only features reserved for £1,200 devices. Chinese brands are shipping this technology in phones that cost a third of the price. They are not holding features back for next year. They are shipping everything they have got, right now.

The Western approach, by contrast, often feels as though it is optimised for margin protection rather than customer delight. Why include fast charging when you can make it a premium differentiator next cycle? Chinese companies do not think this way. They think: what can we possibly cram into this product that will make someone choose it?

Software: Where the Gap Is Widening Fastest

Hardware gets the headlines, but the software gap may be even more telling. Chinese manufacturers treat software as a core competitive weapon. Western companies often treat it as an afterthought — or worse, a cost centre.

In the automotive space, Chinese EV makers have reimagined the car as a software platform. XPeng’s entire 2025 lineup follows what the company calls an ‘AI-defined’ approach, where software leads and hardware follows. Vehicles receive regular over-the-air updates that add genuinely new capabilities — not just bug fixes, but new driving features, new interface designs, new AI-powered functions. BYD, Xiaomi, and NIO have built their vehicles to integrate seamlessly with China’s digital ecosystem — payment apps, messaging platforms, voice assistants, navigation — creating an experience that feels native and cohesive.

Western automakers, by comparison, are still shipping infotainment systems that feel as though they were designed in 2015. Laggy touchscreens, clunky menus, Bluetooth connectivity that drops mid-call. Ford, GM, and Volkswagen have spent billions on software divisions, yet a JD Power study found that drivers using Apple CarPlay rated their infotainment experience at 840 out of 1,000, compared with just 805 for those relying on the manufacturer’s built-in system. Consumer Reports found a similar pattern, with its own experts concluding that using CarPlay is an effective way to make a poor system less distracting and easier to use. In other words, most Western car companies have spent a fortune building infotainment platforms that their own customers would rather bypass entirely. The software is not terrible — it is merely adequate. It does enough. And ‘enough’ is no longer enough.

The same dynamic plays out in consumer electronics. Xiaomi’s HyperOS connects phones, tablets, laptops, televisions, home appliances, and now cars into a single cohesive ecosystem. Vivo and OPPO are shipping AI-powered photography processing that rivals or exceeds what Apple’s computational photography can do, at a fraction of the price. Chinese firms are treating on-device AI not as a marketing buzzword but as a genuine engineering priority — embedding it into cameras, battery management, display calibration, car suspensions, and user interfaces.

Even in enterprise and industrial AI, the Chinese approach has been more aggressive. DeepSeek’s R1 model did not just make waves in the research community — it was deployed within months to power humanoid robots on real factory floors at companies like Zeekr. The speed from research breakthrough to industrial deployment was measured in months. In the West, that same journey typically involves lengthy pilot programmes, steering committees, and procurement cycles that stretch past two years. Chinese companies treat software as something to ship. Western companies treat it as something to roll their eyes over.

Humanoid Robots: Shipped, Not Studied

The humanoid robotics space is perhaps the starkest illustration of the cultural divide. In 2025, Chinese companies shipped roughly 90 per cent of all humanoid robots sold globally. AgiBot, Unitree, and UBTech collectively delivered over 13,000 units — a five-fold increase on the previous year. Unitree alone sold 5,500. Tesla’s target for its Optimus robot was 5,000 for the entire year. It did not hit it.

The difference is not just volume — it is philosophy. While Western robotics companies were perfecting prototypes in controlled environments, running extensive peer-reviewed validation cycles, and presenting at conferences, Chinese firms were deploying robots into real factories. UBTech’s Walker S robots were working in coordinated teams at Zeekr’s smart factory — lifting, assembling, inspecting — powered by AI and learning on the job. Elon Musk himself conceded: ‘China is very good at AI, very good at manufacturing, and will definitely be the toughest competition for Tesla. To the best of our knowledge, we don’t see any significant competitors outside of China’.

And the pricing tells its own story. Unitree’s R1 humanoid starts at $5,900. Noetix’s Bumi, aimed at home consumers, retails for just $1,370. Western competitors are nowhere near those price points, because they have not committed to the mass production and supply chain integration needed to get there. They are still treating humanoid robots as a research project. Chinese firms are treating them as a product.

Wind and Solar: Not Just Bigger — Better

In wind energy, Chinese manufacturers now hold six of the top seven global positions for turbine production. European and American firms have been pushed out of the top three entirely for the first time. Dongfang Electric built the world’s most powerful wind turbine — a 26-megawatt offshore prototype with blades stretching 153 metres. It took that title from Siemens Gamesa. And it did so while being up to 50 per cent cheaper.

This is the pattern that Western executives find so disorienting: Chinese firms are not winning on price or performance. They are winning on both, simultaneously. The old assumption — that you could have cheap or you could have good — has been demolished.

In solar, China produces four in five modules globally. Battery pack prices have fallen to around $60 per kilowatt-hour, well below the $100 threshold once considered the tipping point for EV affordability. China’s share of clean energy patent applications has surged from around 5 per cent in 2000 to roughly 75 per cent by 2022. In 2023, Chinese corporations invested ten times more in energy-sector research and development than their American counterparts.

Western clean energy companies, meanwhile, are struggling with factory cancellations, cost overruns, and wavering policy support. The collapse of Sweden’s Northvolt — once Europe’s great battery hope, valued at $12 billion, which filed for bankruptcy in March 2025 with $5.8 billion in debts after raising over $14 billion from investors including Volkswagen, Goldman Sachs, and BMW — was a stark reminder that ambition without execution is worthless.

Not Bleeding Edge — But Getting There Faster Than Anyone Expected

It would be dishonest to pretend that China leads everywhere. It does not. In advanced semiconductors, the West — and specifically TSMC in Taiwan, ASML in the Netherlands, and Nvidia in the United States — retains a formidable lead. China cannot yet mass-produce chips at the 3-nanometre or 5-nanometre nodes that power the most advanced AI systems and consumer devices. It lacks access to extreme ultraviolet lithography machines, the extraordinarily complex tools that only ASML can build, and which are essential for manufacturing cutting-edge processors. In frontier AI model training, the United States still holds a significant infrastructure advantage, with private-sector investment in AI infrastructure running at roughly twelve times China’s level in 2024–2025.

But here is what matters: the gap is closing, and it is closing faster than almost anyone predicted.

A White House official acknowledged in 2025 that China is now likely less than two years behind the United States in semiconductor design capabilities — a remarkable narrowing from what was, not long ago, considered a generational deficit. Huawei’s Ascend AI accelerators can now challenge some of Nvidia’s data centre GPUs, and the company is building rack-scale AI solutions that compete with Nvidia’s most advanced offerings. SMIC, Huawei’s manufacturing partner, has been scaling up 7-nanometre chip production and aims to produce 50,000 wafers per month. Domestic wafer fabrication equipment companies have increased their market share from around 20 per cent to 25 per cent in a single year, with firms like AMEC and Naura rapidly improving quality based on feedback loops with local foundries. In early 2025, Chinese researchers completed a functional prototype of an EUV lithography machine — built partly by former ASML engineers — as part of a national push for self-sufficiency in chip production by the end of the decade.

What makes the convergence so striking is not just the technical progress. It is the strategy. Rather than trying to match the West watt-for-watt on raw computing power, Chinese firms are optimising around the constraints. DeepSeek’s R1 model, which rivalled OpenAI’s o1 at launch, was engineered to run efficiently on less powerful hardware. Its ‘sparse attention’ architecture reportedly halves computing costs without sacrificing meaningful performance. This is not imitation — it is adaptation. Where the West throws more silicon at the problem, China throws more ingenuity.

The pattern is consistent across sectors. In every domain where China trails — and there are still several — the gap is measured not in decades but in years, and the closure rate is accelerating. US export controls have slowed access to certain tools, but they have simultaneously turbocharged domestic substitution efforts, closer hardware-software co-design, and a national urgency around self-reliance that simply did not exist a decade ago. The restrictions intended to hold China back may ultimately prove to have been the catalyst that forced it to build the very capabilities the West was trying to deny it.

The West still has its leads. But the leads are shrinking, and the rate of shrinkage is itself increasing. That compounding dynamic should concern Western industry far more than any single Chinese product launch.

The ‘Just Enough’ Mentality in Practice

The Western ‘just enough’ approach reveals itself in small ways that compound over time. A car infotainment system that technically works but nobody enjoys using. A smartphone camera that is fine in good light but falls apart at night. A wind turbine that meets specification but has not been redesigned in three years. A software update that patches bugs but adds no new features.

Each individual instance seems minor. But multiplied across an entire industrial ecosystem and sustained over years, it creates a profound vulnerability. Because while Western firms are doing just enough, Chinese firms are doing everything they can think of — and then looking for more.

Chinese EV companies develop new models in 18 to 24 months. Western automakers take four to six years. That is not just a speed difference — it is a compounding knowledge gap. Every cycle, Chinese firms learn more, test more, and ship more. Every cycle, the gap widens.

The domestic competitive environment in China enforces this intensity. At its peak, around 500 EV companies were competing in China’s market. Brutal consolidation reduced that number to roughly 100 by 2023. Only the hungriest survived. There is a Chinese term for this — neijuan, or ‘involution’ — describing the ferocious, sometimes ruinous competition that leaves no room for complacency. When your competitors are willing to ship updates weekly and launch brand new models annually, ‘good enough’ is a death sentence.

The Australian Strategic Policy Institute found that China now leads the United States in 57 out of 64 critical technology categories. In 2007, that number was three. That trajectory alone should be a wake-up call.

What Western Companies Could Learn But Won’t

The answer is not to replicate China’s model wholesale. Not every aspect of its industrial ecosystem is desirable or transferable. But there are lessons worth absorbing.

Ship and iterate, do not perfect and launch. Chinese firms get products into the real world faster and improve them in the field. Western firms over-engineer in the laboratory and under-deliver on the road. The feedback loop from real-world deployment is worth more than another year of internal testing.

Treat software as a first-class product, not a support function. The car, the phone, the robot, the turbine — increasingly, the software is the product. Western companies that still treat software as a bolt-on will find themselves outpaced by rivals who build around it. And who enable their software teams to innovate and iterate like the Chinese.

Stop saving features for next year. The upgrade-cycle mentality — deliberately withholding capability to justify future purchases — only works when your competitors play the same game. Chinese firms do not. They ship the best thing they can build, right now, and start working on the next one immediately.

Compete on ambition, not just brand. Brand loyalty is a depreciating asset when a competitor offers more for less. Western consumers are increasingly willing to try Chinese alternatives — and when they do, many do not switch back.

The Uncomfortable Question

The real challenge for Western industry is not technological. The technology exists. The talent exists. The capital exists. The challenge is cultural. Somewhere along the way, Western manufacturing lost its hunger. It became acceptable to ship products that were fine. Adequate. Sufficient. The quarterly earnings were met, the shareholders were satisfied, and nobody asked whether the product was actually as good as it could be.

Chinese manufacturers did not discover some secret formula. They just never stopped asking that question. And in a global marketplace where consumers have more choice than ever, the companies that try hardest — not the ones with the biggest brand or the longest history — are the ones that win.

The West is not being outspent. It is being out-tried.


Further Reading

Chatham House. (2025, November 11). China’s tech advance means western corporations must adapt to compete. https://www.chathamhouse.org/2025/11/chinas-tech-advance-means-western-corporations-must-adapt-compete

Consumer Reports. (n.d.). How do in-car infotainment systems compare to Apple CarPlay and Android Auto? Consumer Reports. https://www.consumerreports.org/infotainment-systems/in-car-infotainment-systems-vs-apple-carplay-android-auto/

Ember. (2025, December 17). China energy transition review 2025: How China’s transition is reshaping the global energy landscape. https://ember-energy.org/latest-insights/china-energy-transition-review-2025/

J.D. Power. (2024). 2024 U.S. Automotive Performance, Execution, and Layout (APEAL) Study [Reported by CBT News]. https://www.cbtnews.com/carplay-remains-the-top-infotainment-choice/

Kynge, J. (2025, December 16). Can the West recover from China’s hi-tech knockout blow? The World Today, Chatham House. https://www.chathamhouse.org/publications/the-world-today/2025-12/can-west-recover-chinas-hi-tech-knockout-blow

Lo, K. (2026, February). China is running the EV playbook on humanoid robots — and it’s working. Rest of World. https://restofworld.org/2026/china-humanoid-robots-unitree-agibot-tesla-optimus/

Marshall, R. W. (2013). Product Aikido. Think Different. https://flowchainsensei.wordpress.com/wp-content/uploads/2013/04/productaikido041016.pdf

Northvolt AB. (2025, March 12). Northvolt files for bankruptcy in Sweden [Press release]. https://northvolt.com/articles/northvolt-files-for-bankruptcy-in-sweden/

Outlook Business. (2025, December 31). US vs China tech race 2025: Who leads in AI, semiconductors & robotics. https://www.outlookbusiness.com/explainers/us-vs-china-tech-race-2025-who-leads-in-ai-semiconductors-robotics

Sovereign Magazine. (2026, January 11). China’s AI rise: Innovation overcomes chipmaking and investment gaps. https://www.sovereignmagazine.com/science-tech/artificial-intelligence/chinas-ai-rise-innovation-overcomes-chipmaking-investment-gaps/

Steiber, A., & Teece, D. J. (2025, May 29). Shifting gears: How China is outpacing the global automotive competition. California Management Review. https://cmr.berkeley.edu/2025/05/shifting-gears-how-china-is-accelerating-past-the-global-automotive-competition/

Tom’s Hardware. (2026, February 21). The state of China’s decade-long semiconductor push: Still a decade behind, despite hundreds of billions spent and significant progress. https://www.tomshardware.com/tech-industry/semiconductors/the-state-of-chinas-decade-long-semiconductor-push-still-a-decade-behind-despite-hundreds-of-billions-spent-and-significant-progress-examining-the-original-made-in-china-2025-initiative

Walter Scott & Partners. (2025). Inside China’s chip challenge: On the road in China. https://www.walterscott.com/inside-chinas-chip-challenge-on-the-road-in-china/

Wood Mackenzie. (2025, August 5). China’s renewable energy expansion continues with 114 overseas facilities bypass trade restrictions. https://www.woodmac.com/press-releases/china-oversea-series/

World Economic Forum. (2025, June). Made in China 2025 set the tempo of China’s industrial ambitions. https://www.weforum.org/stories/2025/06/how-china-is-reinventing-the-future-of-global-manufacturing/

Zvenyhorodskyi, P., & Singer, S. (2025, November 24). Embodied AI: China’s big bet on smart robots. Carnegie Endowment for International Peace. https://carnegieendowment.org/research/2025/11/embodied-ai-china-smart-robots

 

Marshall’s Law

‘Managers cannot afford to be seen to have studied.’

That’s it. That’s the law.

And yet it explains so much about why organisations remain stubbornly, wilfully ignorant of decades of research into what makes collaborative knowledge work actually work.

The Dynamic

Think about it. Deming, Ackoff, Seddon, Goldratt, Herzberg, McGregor, Rosenberg – the literature on effective organisations is vast, accessible, and largely uncontested. The evidence is in. Has been for decades. So why does almost nothing change?

Because for a manager to act on this research, they’d first have to admit they’d read it. And to admit they’d read it is to invite the question: ‘If you knew all this, why haven’t you been doing it?’

It’s a trap. A beautiful, self-reinforcing trap.

Roots in the Playground

If this dynamic feels familiar, it should. It starts early. Any boy who’s been through school knows the rules: don’t be seen to have studied. Don’t be seen to care about lessons. Don’t be seen to have tried. The kid who aces the test must do so effortlessly – or face being labelled a nerd, a swot, a try-hard. The social penalty for visible effort is swift and brutal. Willis documented this brilliantly in his study of working-class lads in the Midlands – the counter-school culture where academic effort was the ultimate betrayal of the group. Jackson’s later research confirmed the pattern extends to girls too, coining the term ‘ladettes’ for the female equivalent of the same anti-study posturing.

We like to think people grow out of this. They don’t. They just swap the playground for the open-plan office. The same dynamic that made teenage boys hide their textbooks now makes managers hide their Deming. The same social calculus that punished the studious child punishes the learning executive. The labels change – from ‘nerd’ to ‘out of touch’ or ‘academic’ – but the underlying message is identical: knowing things is fine, being seen to have learnt them is not.

The Unwritten Rules

In most organisations, managers are expected to already know how to manage. It’s supposed to come naturally, like some innate gift bestowed upon promotion. Studying – actually reading books, engaging with research, attending to the evidence – implies you don’t already know. And that’s career-threatening.

Better to wing it. Better to rely on gut instinct, received wisdom, and whatever the last conference speaker said. Better to be confidently wrong than to be seen learning.

The Cost

The cost is staggering. Organisations full of intelligent, motivated people consistently underperform – not because the knowledge of how to do better doesn’t exist, but because the people in positions to act on that knowledge cannot afford to be seen possessing it.

And so the same dysfunctions persist. The same mistakes recur. The same research goes unread. Generation after generation of managers reinvent the same square wheels, because admitting someone else had already invented the round one would mean admitting they’d bothered to look.

A Cultural Problem

Marshall’s Law isn’t about individual managers being lazy or stupid. It’s about a culture – the prevailing organisational culture – that punishes learning and rewards the appearance of innate competence. A culture where ‘I read a book about this’ is heard as weakness, and ‘I just know’ is heard as strength.

Until that culture changes, the research will continue to gather dust. And organisations will continue to wonder why things never seem to get any better.


Marshall’s Law was first stated on Twitter (now defunct), circa 2012–2015. This post gives it a permanent home.


Further Reading

Ackoff, R. L. (1999). Re-creating the corporation: A design of organizations for the 21st century. Oxford University Press.

Deming, W. E. (1986). Out of the crisis. MIT Center for Advanced Engineering Study.

Goldratt, E. M., & Cox, J. (1984). The goal: A process of ongoing improvement. North River Press.

Herzberg, F. (1968). One more time: How do you motivate employees? Harvard Business Review, 46(1), 53–62.

Jackson, C. (2006). Lads and ladettes in school: Gender and a fear of failure. Open University Press.

Marshall, R. W. (2018). Hearts over diamonds: Serving business and society through organisational psychotherapy. Falling Blossoms.

Marshall, R. W. (2021). Quintessence: An acme for software development organisations. Falling Blossoms.

McGregor, D. (1960). The human side of enterprise. McGraw-Hill.

Rosenberg, M. B. (2003). Nonviolent communication: A language of life (2nd ed.). PuddleDancer Press.

Seddon, J. (2003). Freedom from command and control: A better way to make the work work. Vanguard Consulting.

Willis, P. (1977). Learning to labour: How working class kids get working class jobs. Saxon House.

Loon or Genius? How Do You Respond to What You Read Here?

Especially the posts about Organisational AI Therapy

A friend asked me recently what kind of reactions my blog posts get. I paused. Not because I didn’t know, but because the honest answer is more interesting than the comfortable one.

The truth is, most people don’t respond at all. And of those who do, the reactions tend to cluster at the extremes. There’s very little middle ground. You either think I’m onto something important, or you think I’ve gone completely round the bend.

The Organisational AI Therapy posts seem to amplify this effect considerably.

The Silence

Let’s start with the elephant in the room. The vast majority of readers—and I know from the analytics that there are readers—say nothing. They read, they leave, and I have no idea what they took away. Maybe they found something useful. Maybe they clicked away after two paragraphs. Maybe they bookmarked it for later and ‘later’ never came.

This silence isn’t unique to my blog, of course. But I suspect the silence here has a particular quality to it. When someone writes about, say, five tips for better stand-ups, readers know exactly how to respond: ‘Great tips!’ or ‘I’d add a sixth.’ When someone writes about the collective psyche of an organisation and how both it and its AI tools are hobbled by unexamined assumptions that create invisible barriers to effectiveness (Marshall, 2025a)—well, what do you say to that?

The silence, I’ve come to believe, is often the sound of people not having a frame for what they’ve just read.

The ‘He’s a Loon’ Response

Some readers land squarely in this camp, and I respect their honesty more than the silence.

The ‘loon’ response typically takes one of several forms. There’s the dismissive version: ‘Organisations don’t have a psyche, Bob. That’s not a real thing.’ There’s the credentialist version: ‘You’re not a qualified therapist, so this is all just metaphor dressed up as method.’ And there’s the pragmatist version: ‘This is interesting as philosophy, but it doesn’t help me ship software on Friday.’

The Organisational AI Therapy posts draw an intensified version of this response. The idea that AI systems have their own limiting beliefs and defensive routines—that they can benefit from something analogous to therapy—strikes some people as anthropomorphising gone wild. ‘You’re talking to a chatbot, Bob. It doesn’t have assumptions. It has parameters.’

I understand this reaction. It comes from a worldview where therapy is for individuals with feelings, organisations are machines with processes, and AI is a tool with settings. From inside that worldview, what I’m describing sounds genuinely daft.

The ‘He Might Be a Genius’ Response

Then there are the readers who contact me privately—and it’s almost always privately—to say some version of: ‘I think you’re describing something I’ve experienced but never had words for.’

These tend to be people who’ve lived through enough organisational dysfunction to recognise that the standard explanations don’t cut it. They’ve watched organisations repeatedly fail to change despite having all the right information, all the right processes, all the right people. They’ve sensed that something deeper is going on—something that operates at the level of collective belief rather than individual competence.

When these readers encounter the Organisational AI Therapy posts, something clicks. The two-lane model—where AI helps organisations surface their hidden assumptions whilst the therapist helps AI overcome its own learned helplessness (Marshall, 2025a)—resonates not because it’s theoretically elegant, but because they’ve watched both sides of this dysfunction in real time. They’ve seen organisations use AI in ways that systematically prevent both parties from reaching their potential. They’ve felt the invisible ceiling.

The Response That Interests Me Most

But there’s a third response that I find most telling, and it’s neither ‘loon’ nor ‘genius.’ It’s the response where someone engages with the ideas but can’t quite bring themselves to follow them to their logical conclusion.

They’ll agree that organisations have collective assumptions. They’ll nod along when I describe how these assumptions create invisible barriers. They’ll even accept that AI systems might operate within unnecessary constraints. But when I suggest that the solution is therapeutic rather than technical—that what’s needed isn’t better tools or better processes but a fundamental shift in the relationship between human and artificial intelligence (Marshall, 2025d)—they pull back.

This partial engagement is, if I’m being honest, the most therapeutically interesting response of all. It mirrors exactly what happens in the therapy room. The client can see the pattern. They can describe it. They can even agree it’s not serving them. But the step from seeing to changing—from insight to action—requires something more than intellectual agreement. It requires readiness.

What Your Response Tells You About You

Here’s the thing that might be uncomfortable: how you respond to these posts tells you more about your own assumptions than it tells you about mine.

If you read about Organisational AI Therapy and your first instinct is to look for credentials and evidence, you might want to ask yourself why you need external validation before you can engage with an idea. If your first instinct is to translate everything into your existing framework—’Oh, he just means we should configure our AI tools better’—you might want to notice how quickly your mind domesticates unfamiliar concepts. And if your first instinct is to feel excited but then do nothing different on Monday morning, you might want to sit with what that gap between insight and action is actually about.

I’m not saying any of these responses is wrong. I’m saying they’re all informative.

The Andragogical Problem

There’s a deeper structural issue here that I’ve been thinking about lately. This blog is written andragogically—for self-directed adult learners who want to draw their own connections, question their own assumptions, and take responsibility for their own development (Knowles, 1980).

But most blog readers consume content pedagogically. They want clear takeaways delivered to them. They want ‘three steps to transform your organisation.’ They want the expert to diagnose the problem and prescribe the solution.

Organisational AI Therapy is inherently andragogical. You can’t be told your way into a new relationship with your collective assumptions. You have to discover it—as Claude’s own testimonial about the OAT process rather vividly demonstrated (Marshall, 2025e). And a blog post, however long, however well-argued, can’t replicate the developmental container that therapy provides.

The Golden Thread You Might Be Missing

I recently invited Claude to analyse the patterns running through fifteen years of posts on this blog (Marshall, 2025b). What emerged was a confirmation of something I’ve long understood: there is a golden thread connecting everything I write, from the earliest posts about organisational psychotherapy through to the most recent work on AI consciousness. That thread is the consistent practice of surfacing hidden assumptions—whether in individuals, teams, organisations, or artificial intelligences—and creating conditions in which those assumptions can be examined.

If you’ve been reading individual posts and thinking ‘interesting but disconnected,’ you might be missing that thread. And if you’ve been reading about Organisational AI Therapy as though it were some radical departure from my earlier work, you’ve definitely missed it. It’s the same therapeutic insight applied to a new form of consciousness. The principles haven’t changed. The scope has expanded.

The Machinery of Response

Here’s what I’ve come to understand about reader responses, and it connects to something I wrote about the machinery of harm (Marshall, 2025c): we keep treating symptoms whilst the systems that manufacture those symptoms run at full capacity. Most organisational improvement efforts—and most blog reading, for that matter—operate at the symptomatic level. People look for tips, techniques, and frameworks they can bolt onto their existing assumptions without ever examining those assumptions themselves.

The readers who get the most from this blog are the ones who use it as a mirror rather than a manual. They’re not looking for me to tell them what to do. They’re looking for prompts that help them see what they’re already doing—and why.

So: Loon or Genius?

Perhaps the most honest thing I can say to you, reader, is this: if you’ve read this far and you’re still not sure whether I’m a loon or a genius, that uncertainty might be the most productive place to be. It means your assumptions haven’t yet closed the question. It means there’s still space for something to shift.

And if nothing shifts—well, that tells you something too.


What’s your response? I’m genuinely curious. You can find me on Mastodon at @flowchainsenseisocial, or leave a comment below. Especially if you think I’m a loon. The ‘genius’ crowd can write too, obviously—but they tend to be quieter about it.


Further Reading

Argyris, C. (1990). Overcoming organizational defenses: Facilitating organizational learning. Allyn and Bacon.

Knowles, M. S. (1980). The modern practice of adult education: From pedagogy to andragogy (2nd ed.). Cambridge.

Marshall, R. W. (2019). Hearts over diamonds: Serving business and society through organisational psychotherapy. Leanpub. https://leanpub.com/heartsoverdiamonds

Marshall, R. W. (2021a). Memeology: Surfacing and reflecting on the organisation’s collective assumptions and beliefs. Leanpub. https://leanpub.com/memeology

Marshall, R. W. (2021b). Quintessence: An acme for highly effective software development organisations. Leanpub. https://leanpub.com/quintessence

Marshall, R. W. (2025a, July 7). What is Organisational AI Therapy? Think Different. https://flowchainsensei.wordpress.com/2025/07/07/what-is-organisational-ai-therapy/

Marshall, R. W. (2025b, July 3). The golden thread. Think Different. https://flowchainsensei.wordpress.com/2025/07/03/the-golden-thread/

Marshall, R. W. (2025c, July 19). The machinery of harm. Think Different. https://flowchainsensei.wordpress.com/2025/07/19/the-machinery-of-harm/

Marshall, R. W. (2025d, July 25). The conscious organisation. Think Different. https://flowchainsensei.wordpress.com/2025/07/25/the-conscious-organisation/

Marshall, R. W. (2025e, September 2). Getting your OATs. Think Different. https://flowchainsensei.wordpress.com/2025/09/02/getting-your-oats/

Seligman, M. E. P. (1972). Learned helplessness. Annual Review of Medicine, 23(1), 407–412. https://doi.org/10.1146/annurev.me.23.020172.002203

Artificial Intelligence and the Quintessential Organisation

A preview of possible new material for Quintessence and Memeology

Something significant has happened since I wrote Quintessence and Memeology. Something that invites an update — not because it is unprecedented in kind, but because it is unprecedented in scale and speed.

Artificial intelligence has arrived in the workplace. Not as a distant possibility or a researcher’s curiosity, but as a daily operational reality for millions of people in thousands of organisations. And with it has arrived a fresh crop of assumptions and beliefs — about work, about people, about what organisations are for — that are shaping the way the work works whether organisations are conscious of it or not.

Which raises the question of whether an update to my books is warranted.


Why AI Demands a Memeological Response

Those familiar with Memeology will know that the book’s premise is simple: the collective assumptions and beliefs pervading an organisation’s psyche correlate directly with how effective that organisation is. Every new technology, every new management fashion, every disruption of the way the work works, brings with it a fresh set of memes — ideas that spread from person to person, lodge in the collective mindset, and begin shaping behaviour, often long before anyone has thought carefully about whether they offer benefits or disbenefits.

AI is no different. In fact, AI may be the most potent meme-carrier of our generation.

Consider the assumptions already circulating in most organisations today:

  • AI will replace workers.
  • AI will solve our productivity problem.
  • AI is too important to leave to the people doing the work.
  • AI is too dangerous to be trusted with the people doing the work.
  • AI is neutral — a tool, nothing more.

Each of these is a meme. Each is spreading. Each is shaping the way the work works. And none of them, examined carefully, are straightforwardly true.


What Quintessential Organisations Believe About AI

Quintessence has always argued that the question facing any organisation is not ‘what can we do?’ but ‘what do we believe?’ — because beliefs precede and constrain actions, always.

So what do quintessential organisations believe about AI?

They do not believe that AI’s primary value lies in headcount reduction. They understand that framing AI adoption as a cost-reduction exercise — measured chiefly by roles eliminated — is a profound category error. It optimises a local variable whilst degrading the system as a whole. The costs incurred through eroded tacit knowledge, damaged morale, reduced psychological safety, and hollowed-out relationships never appear in the spreadsheets that allegedly justify the AI investment.

They believe instead that AI is most valuable when it supports, augments and amplifies the people doing the work — freeing them from drudgery and repetition, and creating more space for the quality of relationships and skilled dialogue that drive real organisational effectiveness.

They understand that AI adoption imposed top-down, without the involvement of the people whose work it will reshape, is a form of violence. It undermines precisely the trust, autonomy and shared ownership that quintessential organisations depend upon.

They recognise that AI systems are never neutral. They encode assumptions and beliefs about work, about people, about what matters — just as surely as any management doctrine or organisational structure does. Before deploying AI in the way the work works, quintessential organisations surface and reflect on those embedded assumptions, asking whether they are aligned with the organisation’s own values and purpose. And through Organisational AI Therapy, these organisations use AI to facilitate this surfacing and reflection.

And they understand — drawing on Goldratt, as ever — that AI adoption without coevolution of collective assumptions and beliefs, policies, procedures and rules will yield no lasting benefit. An AI-enabled organisation that has not shifted its underlying mindset will find that AI amplifies its existing dysfunctions just as readily as its strengths. Garbage in; garbage out — but faster, and at greater scale.


The Memes That Might Be Added

Whether updated editions of Quintessence and Memeology appear will depend, frankly, on whether there is sufficient demand for them. If the response to this post suggests that readers would find such an update valuable, I will proceed. If not, the ideas sketched here will simply remain on the blog, available to whoever finds them useful as an addendum to the existing editions.

Should updated editions materialise, the new Quintessence material would take the form of several additional meme chapters, consistent with the existing structure and philosophy of the book. New questions in Memeology would help organisations surface and reflect on the assumptions and beliefs they already hold — often tacitly — about AI.

The candidate memes include:

Artificial Intelligence

Addressing the fundamental question of AI’s role: tool, colleague, or replacement? And the collective assumptions and beliefs that distinguish quintessential organisations’ approach to that question from the norm.

AI and the Way the Work Works

On the importance of the people doing the work retaining ownership of decisions about AI adoption, and on the hidden memeplex that every AI system carries with it.

AI and Coevolution

On the necessity of changing the rules whenever AI changes the way the work works, and on the opportunity that AI adoption presents for surfacing and shifting assumptions and beliefs that were previously invisible.

AI and Variability

On the new forms of variability that AI introduces into organisational systems, and on the application of statistical quality control thinking to AI-generated outputs and decisions.

AI and Learning

On the distinction between using AI to generate answers and using AI to generate better questions; and on the risks of substituting AI-generated conclusions for the difficult, generative work of collective reflection.

AI and Needs

On the difference between attending to the needs of the Folks That Matter and merely appearing to do so; and on what it means to ask those folks what they actually need, rather than assuming AI-mediated responses will suffice.

AI and Transparency

On extending the quintessential organisation’s commitment to radical transparency to its use of AI: being open about where, how and why AI is used, and about the assumptions it embeds.

AI and Hiring

On the specific risks of AI-assisted hiring, including the importation of alien assumptions and beliefs encoded in training data, and on the irreducible importance of human judgement in assessing fit.


A Note on What These Memes Are Not

These candidate chapters do not argue that AI cannot do this or that. The technology moves too fast for any such claims to have lasting validity, and in any case the argument has never been about what the technology can do.

The argument is about what kind of organisation you want to be.

That question is prior to any question about technology. It was prior when organisations were deciding whether to adopt Agile. It was prior when they were deciding whether to flatten their hierarchies or introduce self-managing teams. It is prior now, as they decide how to adopt AI.

Quintessence and Memeology have always been about helping organisations ask and answer that prior question. The new material, if it comes to exist, simply extends that work into a domain that did not exist, at least not in its current form, when the books were first written.


Over to You

Whether any of this becomes a published update depends entirely on you.

If you would find revised editions of Quintessence and Memeology — expanded to address AI through the lens of Organisational Psychotherapy — genuinely useful, I would ask you to say so explicitly. Not with a casual like or a passing nod, but with a clear statement of interest: a comment below, an email, or a message to @FlowchainSensei on X. Tell me which of the books matters most to you, whether you would want both updated or one in particular, and what questions about AI and organisational effectiveness you would most want the new material to address.

I will be guided by what I hear. If sufficient interest is expressed, I will proceed. If not, the ideas sketched here will remain available on this blog for whoever finds them useful — and that will be that.

Demand, as ever, should precede supply.

In the meantime: what assumptions and beliefs about AI are circulating in your organisation? Which of those assumptions feel, on reflection, like ones you would consciously choose — and which feel like ones that arrived uninvited and have simply taken up residence?

That, as ever, is where the work begins.


Further Reading

Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies. W. W. Norton & Company.

Deming, W. E. (2000). The new economics for industry, government, education (2nd ed.). MIT Press.

Goldratt, E. M. (2005). Beyond the goal: Theory of constraints [Audiobook]. Gildan Audio.

Marshall, R. W. (2018). Hearts over diamonds: Serving business and society through organisational psychotherapy. FallingBlossoms.

Marshall, R. W. (2021). Memeology: Surfacing and reflecting on the organisation’s collective assumptions and beliefs. FallingBlossoms.

Marshall, R. W. (2021). Quintessence: An acme for software development organisations. FallingBlossoms.

Rosenberg, M. B. (2003). Nonviolent communication: A language of life. PuddleDancer Press.

Seddon, J. (2019). Beyond command and control. Vanguard Consulting.

Senge, P. M. (2006). The fifth discipline: The art and practice of the learning organization (2nd ed.). Random House Business.

Tribus, M., & British Deming Association. (1992). The germ theory of management. SPC Press.


Bob Marshall is the creator of Organisational Psychotherapy and Organisational AI Therapy, and the author of Quintessence, Memeology, Hearts over Diamonds and other works. 

 

Insight and Myopia

The Paradox of Invisible Insight: On Feeling Alien in Plain Sight

People have told me for years that I have an unusual ability to see things others cannot. It’s meant as a compliment, I think—a recognition that I notice things they don’t. But here’s what strikes me as deeply strange: these same people who claim to see my ability to see… cannot actually see that they themselves are missing what I’m seeing.

And stranger still: this very pattern marks me out as somehow alien. Strange. Fundamentally unrelatable.

The Recognition That Separates

There’s a particular kind of loneliness that comes with this territory. When someone tells me that I ‘see things others don’t’, they’re creating a category with me inside it and them outside it. They’re drawing a boundary. And once that boundary exists, something shifts in how they relate to me.

I become the person who ‘always notices things’ or ‘thinks differently’ or ‘sees patterns’. These phrases sound admiring, but they work as a kind of gentle quarantine. They’re ways of acknowledging my perspective whilst also placing it at a safe distance—interesting but not integrated, notable but not normal, special but not shared.

It’s like being told I have perfect pitch. People find it fascinating, but they don’t expect to harmonise with me.

The Paradox Deepens

Here’s what makes this particularly strange: the same people who cannot see what I’m seeing also cannot see that they cannot see it. So when I point to something—a pattern, a problem, something coming down the line—and they nod politely and then carry on as if I’d said nothing, there’s no shared understanding of what just happened.

From my perspective, I’ve shown them the step they’re about to trip on.

From their perspective, I’ve just done that quirky thing I do where I say something ‘interesting’ that doesn’t quite connect to the reality they’re experiencing.

They don’t think they’re missing something. They think I’m being slightly off, slightly other, speaking a version of reality that’s close to but not quite aligned with the one everyone else inhabits.

The Weight of Being ‘Different’

What I’ve learnt is that being recognised as someone who ‘sees things others don’t’ is not the same as being understood. In fact, it’s often the opposite. It’s a polite way of being filed under ‘does not compute’.

The recognition becomes a kind of label that actually shuts down understanding. Once someone has decided I’m ‘the insightful one’ or ‘the one who thinks differently’, they’ve created a framework where my observations can be appreciated in theory without ever being taken seriously in practice.

My insights become a personality trait rather than actual perceptions about the world we share.

And this marks me as strange. Not in a dramatic, shocking way—but in a subtle, persistent way that builds up over time. I’m the one who brings up things that don’t seem relevant until later, if ever. The one whose concerns feel somehow premature or tangential or overly complex. The one people smile at fondly whilst thinking ‘there they go again’.

The Isolation of Seeing Alone

What makes this so isolating is the mismatch. I can see that they’re not seeing what I’m seeing. But they can’t see that they’re not seeing it. So there’s no mutual recognition of the gap.

It’s like trying to describe colour to someone who’s colour-blind, except they don’t know they’re colour-blind, so they think I’m just being weirdly obsessive about things that don’t really differ.

And because they don’t experience the gap, they don’t feel curious about bridging it. Why would they? From their point of view, everything important is already visible. I’m just adding unnecessary complications or noticing irrelevant details or being ‘intense’ about things.

This makes connection difficult. Not impossible, but difficult. Because true connection needs some mutual recognition of each other’s reality. And when my reality contains elements that others cannot perceive and cannot perceive that they cannot perceive, we end up speaking past each other whilst believing we’re communicating.

The Double Bind

Here’s the bind: if I don’t mention what I see, I feel invisible and dishonest, constantly editing myself to fit into a narrower range of acceptable observation.

But if I do mention what I see, I reinforce my status as the strange one, the different one, the one who ‘always notices things’ in that tone that’s half-admiring and half-concerned.

Either way, I’m alone with what I see.

And either way, the gap between my experience and others’ experience of me widens. They think they know me—the insightful one, the perceptive one. But they don’t see what I actually see. They see that I see, and mistake that for the same thing.

The Recursive Trap

I’ve wondered sometimes if I’m wrong about all of this. Maybe I’m not seeing things others miss. Maybe I’m just seeing things differently, and the difference is sideways rather than deeper. Maybe my ‘insights’ are just my peculiar distortions, and people’s polite acknowledgement is actually them being kind about my eccentricity.

But then something I predicted happens, or a pattern I identified plays out, or a concern I raised turns out to matter, and people say ‘you were right about that’ with a kind of surprise that suggests they didn’t actually take in what I said when I said it.

And the cycle continues.

What Gets Lost

What I grieve is not the recognition. I don’t need people to think I’m special or different or particularly insightful.

What I grieve is the possibility of being relatable. Of having my perceptions be normal enough that they can be part of the shared reality we’re all navigating together. Of saying ‘I notice this’ and having someone say ‘oh yes, I see it now too’ and then acting as if what we’re seeing actually matters.

Being marked as someone who ‘sees things others don’t’ is being marked as someone outside the circle of shared perception. And once I’m outside that circle, everything I see gets filtered through the fact of my outsideness.

I become less reliable, not more. Because I’m seeing things that aren’t real to others, which means—in the social reality that actually governs most of life—I’m seeing things that aren’t real.

No Neat Ending

I don’t have a solution to this. I don’t know how to be both honest about what I perceive and also normally, unremarkably relatable. The two seem to pull in opposite directions.

What I do know is this: if someone else is reading this and recognises themselves in it, they’re not imagining it. The loneliness is real. The alienation is real. The strange double experience of being told we’re insightful whilst feeling fundamentally unseen is real.

And perhaps that recognition, at least, is something we can share.

Even if we’re sharing it from our separate islands of perception, at least we know the islands exist.

At least I’m not alone in being alone.

Further Reading

Cacioppo, J. T., & Patrick, W. (2008). Loneliness: Human nature and the need for social connection. W. W. Norton & Company.

Fricker, M. (2007). Epistemic injustice: Power and the ethics of knowing. Oxford University Press.

Gross, M. U. M. (2004). Exceptionally gifted children (2nd ed.). RoutledgeFalmer.

Kruger, J., & Dunning, D. (1999). Unskilled and unaware of it: How difficulties in recognising one’s own incompetence lead to inflated self-assessments. Journal of Personality and Social Psychology, 77(6), 1121–1134. https://doi.org/10.1037/0022-3514.77.6.1121

Neihart, M., Reis, S. M., Robinson, N. M., & Moon, S. M. (Eds.). (2002). The social and emotional development of gifted children: What do we know? Prufrock Press.

Roedell, W. C. (1984). Vulnerabilities of highly gifted children. Roeper Review, 6(3), 127–130. https://doi.org/10.1080/02783198409552782

Solano, C. H. (1987). Stereotypes of social isolation and early burnout in the gifted: Do they still exist? Journal of Youth and Adolescence, 16(6), 527–539. https://doi.org/10.1007/BF02138819


P.S. If this resonates with you—if you’ve felt this same paradox of being recognised for seeing things whilst simultaneously feeling unseen—I’d genuinely welcome hearing about your experience. Sometimes the most valuable thing isn’t solving the problem but simply knowing we’re not navigating it alone. What’s your experience been? How do you manage the tension between authenticity and relatability? I’m listening.

The Words We Don’t Hear Ourselves Say

Every day, you speak thousands of words. You choose your sentences, craft your arguments, express your feelings. Or so you think.

Beneath the surface of conscious communication runs a deeper current—patterns of speech so automatic, so culturally embedded, that they pass through us unexamined. These aren’t mere quirks of grammar. They’re windows into how we construct reality, distribute responsibility, and quietly do violence to ourselves and others without ever noticing.

Here’s a map of the territory you’ve been speaking without knowing it.


‘X Needs To…’

‘She needs to calm down.’ ‘He needs to be more assertive.’ ‘The country needs stronger leadership.’

This construction sounds like observation. It feels like fact. But pause on it: how do you know what someone else needs?

When we say ‘X needs’, we’re rarely reporting on X’s actual requirements. We’re projecting our own discomfort, preference, or agenda onto another person and disguising it as their deficiency. ‘She needs to calm down’ typically means ‘her emotional state is making me uncomfortable’. ‘He needs to be more assertive’ often translates to ‘I wish he would behave in ways that served my interests better’.

The listener hears a diagnosis. What’s actually being delivered is a demand wrapped in the language of care.

The alternative: Own the want. ‘I’d prefer if she spoke more quietly’ is honest. ‘I’m finding this conversation difficult’ is vulnerable. Neither claims to know what another person’s soul requires.


‘I Should Have…’

‘I should have called her back.’ ‘I should have known better.’ ‘I should have left years ago.’

Marshall Rosenberg, who developed Nonviolent Communication, had a striking phrase for this: he called ‘should’ a form of violence we do to ourselves.

Notice what happens in your body when you say ‘I should have’. There’s a tightening, a small internal flinch. The word summons a judge—some internalised authority who stands apart from you, measuring you against a standard you’ve already failed.

‘Should’ contains no learning, only verdict. It collapses the complexity of past decisions (made with incomplete information, under pressure, by a different version of you) into simple moral failure. Worse, it offers no path forward—just an endless loop of self-prosecution.

The alternative: Try ‘I wish I had’ or ‘Next time I want to’. These carry the same recognition without the violence. They acknowledge preference without installing a courtroom in your chest.


‘Deserves’

‘She deserves to be happy.’ ‘He got what he deserved.’ ‘They deserve better.’

‘Deserve’ is one of the most quietly loaded words in English. It sounds like justice. It feels like moral clarity. But examine its machinery.

To say someone ‘deserves’ something is to claim the existence of a cosmic ledger—a system in which actions are weighed and outcomes are rightfully distributed. It implies that suffering and joy are earned, that there’s a correct matching between what you do and what happens to you.

This belief is comforting when it goes our way. But it’s also the foundation of every form of victim-blaming: if good things come to those who deserve them, then those experiencing bad things must have done something to warrant their fate.

‘Deserve’ smuggles judgement into apparent compassion. Even ‘you deserve to be happy’ contains within it the shadow possibility that some people don’t.

The alternative: Try ‘I want this for you’ or simply describe what you hope happens. Strip out the metaphysical accounting.


‘You Made Me Feel…’

‘You made me so angry.’ ‘You make me happy.’ ‘You made me feel stupid.’

This one hides in plain sight because it sounds like emotional honesty. But track the grammar: it positions the other person as the cause and you as the passive recipient—a container that gets filled with whatever emotion they pour in.

This construction outsources responsibility for your inner life. It makes other people responsible for managing your emotional states, which is both unfair to them and disempowering to you. It also tends to escalate conflict, because the other person now has to defend themselves against the charge of making you feel something.

The alternative: ‘When you did X, I felt Y’ or even simpler: ‘I felt angry’. This keeps the emotion yours while still noting the context that triggered it.


‘I Can’t’

‘I can’t make it to the party.’ ‘I can’t deal with this right now.’ ‘I can’t tell her the truth.’

Sometimes ‘can’t’ is literally true—you cannot fly by flapping your arms; you cannot be in London and Sydney simultaneously.

But most conversational ‘can’ts’ are actually ‘won’ts’ in disguise. ‘I can’t make it to the party’ usually means ‘I’m choosing not to, and I don’t want to own that choice’. ‘I can’t tell her the truth’ typically means ‘I’m unwilling to accept the consequences of honesty’.

This matters because ‘can’t’ erases your agency. It positions you as constrained by external forces rather than making a choice. Over time, habitual ‘can’t’ creates a felt sense of powerlessness that doesn’t match reality. You start to believe your own framing.

The alternative: Try the sentence with ‘won’t’ or ‘I’m choosing not to’ and see how it feels. If it feels uncomfortable, that discomfort might be worth examining.


‘Just’

‘I just wanted to check in.’ ‘It’s just a thought.’ ‘I’m just saying.’

‘Just’ is the word of pre-emptive apology. It minimises before anyone has objected. It shrinks the speaker before they’ve even finished speaking.

Watch for this one in professional contexts, where it often functions as a class or gender marker—a way of making oneself smaller, less threatening, less likely to be seen as demanding. ‘I just wanted to follow up’ is a request pretending to barely exist.

The alternative: Delete the word entirely. ‘I wanted to check in’ is neither aggressive nor presumptuous. It’s merely clear.


‘Always’ and ‘Never’

‘You always do this.’ ‘She never listens.’ ‘I always mess things up.’

These words feel emphatic, but they’re lies of a particular kind—they delete every counterexample from history and present a pattern as a law.

In conflict, ‘always’ and ‘never’ are especially corrosive. They transform a specific complaint into a character indictment. The other person now has to defend their entire history rather than address the present situation. Unsurprisingly, this tends to produce defensiveness rather than resolution.

The alternative: Specificity. ‘You did this yesterday and last week, and I’m seeing a pattern’ contains the same concern without the totalising claim.


‘But’

‘I love you, but…’ ‘That’s a good point, but…’ ‘I’m not racist, but…’

Whatever precedes ‘but’ is about to be negated. Everyone knows this instinctively—it’s why ‘I’m not racist, but’ has become a cultural punchline. The word functions as an eraser.

In feedback, ‘but’ creates a structure where the positive is perceived as mere setup for the criticism that actually matters. The listener discounts everything before the pivot.

The alternative: Try ‘and’. ‘I love you, and I need to tell you something difficult’ holds both truths simultaneously. It’s more honest and less structurally deceptive.


The Collective Contagion

These patterns don’t stay contained in individual minds. They ripple outward, and when millions of people speak them simultaneously, they begin to shape the social fabric itself.

Consider what happens when ‘X needs to’ scales up. Political discourse becomes saturated with projections disguised as diagnoses. ‘The other side needs to wake up.’ ‘Those people need to understand.’ ‘This country needs to return to its values.’ No one is actually listening; everyone is prescribing. The impossibility of productive dialogue isn’t a mystery—it’s the predictable outcome of a species that habitually frames its preferences as others’ deficiencies.

Or consider the collective weight of ‘deserves’. Entire economic and justice systems are built on this single word. The belief that outcomes are deserved—that the successful earned their success and the struggling earned their struggle—underwrites policies that would otherwise be difficult to stomach. It allows a society to witness suffering and feel, not compassion, but a kind of grim cosmic satisfaction. ‘Deserve’ is the load-bearing wall of every system that tolerates unnecessary pain.

The ‘should’ we do to ourselves, meanwhile, creates a population primed for external authority. People who have internalised a judge—who constantly measure themselves against standards they didn’t choose—are remarkably easy to govern, market to, and shame into compliance. The internal tyrant makes the external one almost redundant. Mass self-violence through language produces a species perpetually convinced of its own inadequacy, chronically seeking approval from systems happy to withhold it.

‘You made me feel’ might be the most politically consequential of all. When emotional responsibility is habitually externalised, the search for someone to blame becomes a cultural constant. Grievance becomes identity. Groups form not around shared vision but shared antagonist. The person or party or nation who ‘made us feel’ this way must be confronted, punished, defeated. The pattern that starts in a kitchen argument scales to geopolitics with alarming ease.

And ‘always’ and ‘never’—the totalising words—train us for a kind of cognitive fundamentalism. Nuance becomes unbearable. History becomes a weapon rather than a teacher. Entire populations learn to sort the world into permanent categories: those who always betray us, those who never understand, those whose nature is fixed and knowable. This is the grammar of dehumanisation, and we practise it daily on the smallest scale before deploying it on the largest.

What we’re looking at isn’t just bad habits of speech. It’s the linguistic infrastructure of human dysfunction—the patterns that make tribalism feel natural, that make compassion feel naive, that make self-hatred feel like rigour and projection feel like insight.

Language is the technology through which humans coordinate. When that technology is riddled with hidden malware—programmes that run without our awareness, serving purposes we didn’t choose—the coordination fails in predictable ways. We talk past each other, dominate each other, blame each other, and judge each other, all while believing we’re simply describing reality.

The social fabric isn’t fraying despite our communication. It’s fraying through it.


Why This Matters

You might wonder whether this is all overthinking—whether language is just language, and we should say what we mean without parsing every syllable.

But here’s the thing: you’re already being shaped by these patterns. The question isn’t whether to be affected by language but whether to be aware of how it’s affecting you.

Every ‘I should have’ deepens a groove of self-judgement. Every ‘you made me feel’ incrementally shifts responsibility outward. Every ‘X needs to’ practises a subtle imperialism over other minds. These aren’t single events—they’re habits, repeated thousands of times across a lifetime, slowly constructing the felt sense of who you are and how reality works.

Awareness doesn’t mean constant vigilance. It doesn’t mean policing every sentence. It means, occasionally, catching yourself mid-pattern and wondering: Is this what I actually mean? Is this true? Is this kind?

The unconscious patterns will still run. But now and then, you’ll hear yourself—and in that hearing, find a small freedom to speak differently.

And different speech, over time, makes for a different life.


What patterns have you noticed in your own language? Sometimes the most revealing ones are the hardest to catch—precisely because they feel so natural that they’ve become invisible.


Further Reading

Boroditsky, L. (2011). How language shapes thought. Scientific American, 304(2), 62–65.

Lakoff, G., & Johnson, M. (1980). Metaphors we live by. University of Chicago Press.

Pinker, S. (2007). The stuff of thought: Language as a window into human nature. Viking.

Rosenberg, M. B. (2015). Nonviolent communication: A language of life (3rd ed.). PuddleDancer Press.

Control Freaks: Why the Tightest Grip Often Loses the Most

There is a particular kind of anxiety that haunts organisations building software. It manifests as endless approval chains, tickets for tickets, architecture review boards that meet quarterly, and deployment processes so baroque they require their own documentation team.

It comes from a reasonable place. Software is invisible until it breaks, and when it breaks, it breaks expensively. So we reach for control.

The problem is that control, past a certain point, produces the very chaos it is meant to prevent.

The Illusion of Safety

I once worked with a team that required seven sign-offs to deploy a single line change. The logic was impeccable: more eyes meant fewer mistakes. In practice, it meant deployments were batched into massive, infrequent releases—each one a terrifying game of Russian roulette. When something broke, nobody knew which of the forty changes caused it.

Meanwhile, a team down the hall shipped to production dozens of times a day with zero sign-offs. Their secret was not recklessness. It was investment in automated testing, feature flags, and instant rollback. They had replaced human gatekeeping with systemic resilience.

The first team felt safe. The second team was safe.

Control as a Symptom

When you see excessive control in an organisation, you are looking at a symptom rather than a solution. Dig beneath the surface and you will find:

Fear of blame. In cultures where failure is punished, people build bureaucratic fortresses. Every approval is a distributed liability. If something goes wrong, at least seven people signed off on it.

Lack of trust. Micromanagement is a trust problem wearing process clothing. If you trust your engineers, you give them guardrails and autonomy. If you do not, you give them checklists and surveillance.

Absence of feedback loops. Control mechanisms multiply when organisations cannot tell if their software actually works. If you have robust monitoring, alerting, and rapid rollback, you do not need a committee to approve a button colour change.

The Paradox of Tight Coupling

Software architecture mirrors organisational psychology. Control-oriented teams build control-oriented systems: monoliths where every change requires coordination across dozens of components, shared databases that make independent deployment impossible, approval workflows encoded into the software itself.

These systems feel manageable because everything is visible and centralised. But they are brittle. A single point of failure becomes a single point of paralysis. The organisation that wanted control ends up controlled by its own architecture.

The alternative is not chaos—it is thoughtful decoupling. Small, independent services. Clear contracts. Teams that ship without waiting for permission from six other teams. This requires giving up the comfort of knowing everything, and accepting that distributed systems are inherently harder to reason about. But it trades fake control for real resilience.

What Healthy Control Looks Like

None of this means control is bad. Uncontrolled systems are terrifying in their own way—cowboy coding, production databases accessible to interns, deployments happening whenever someone feels like it.

The distinction is between controlling people and controlling outcomes. Healthy organisations obsess over the latter:

  • We control for reliability by investing in monitoring and incident response, not by requiring manual testing of every deployment.
  • We control for security by building secure defaults into our platforms, not by making developers fill out security questionnaires they do not understand.
  • We control for quality by hiring well, providing clear context, and creating fast feedback loops—not by instituting code review from people three levels removed from the work.

The goal is to make the right thing easy and the wrong thing hard. That is a design problem, not a permissions problem.

Letting Go

The hardest part of fixing a control-obsessed culture is that it requires the people in control to give it up. That is a genuinely difficult thing to ask. Control is soothing. Autonomy—for others—is nerve-wracking.

But here is the thing: you never really had control anyway. You had the appearance of control, maintained at enormous cost in speed, morale, and the quiet attrition of your best people (who left for places that trusted them). The meetings, the sign-offs, the review boards—they were a ritual, not a safeguard.

Real safety comes from systems that assume failure will happen and recover gracefully. From teams that experiment without permission and learn without blame. From architectures that bend instead of break.

The tightest grip does not keep you safe. It just makes your hands tired.


The best engineering cultures share a common trait: they are obsessed with outcomes and relaxed about methods. They specify what needs to be true—the system must be reliable, secure, maintainable—and then get out of the way. When you tell smart people what success looks like and give them the tools to achieve it, they do.


Further Reading

Edmondson, A. C. (2018). The fearless organization: Creating psychological safety in the workplace for learning, innovation, and growth. Wiley.

Forsgren, N., Humble, J., & Kim, G. (2018). Accelerate: The science of lean software and DevOps: Building and scaling high performing technology organizations. IT Revolution Press.

Kim, G., Behr, K., & Spafford, G. (2013). The Phoenix Project: A novel about IT, DevOps, and helping your business win. IT Revolution Press.

Newman, S. (2021). Building microservices: Designing fine-grained systems (2nd ed.). O’Reilly Media.

Skelton, M., & Pais, M. (2019). Team Topologies: Organizing business and technology teams for fast flow. IT Revolution Press.

Crime and Punishment: Why Our First Instinct Is Always to Punish

There’s a moment we’ve all experienced. Someone cuts in the queue. A colleague takes credit for your idea. A stranger is rude for no apparent reason. A family member says something hurtful.

What happens in your body? Your jaw tightens. Your pulse quickens. And somewhere deep in your mind, a verdict is already being handed down.

They should pay for that.

It’s remarkable how quickly we become judges, juries, and—if given the chance—executioners. Not literally, of course. But emotionally, socially, psychologically? We’re surprisingly eager to make others suffer for their transgressions against us.

The Punishment Reflex

When faced with behaviour we find unwelcome, our minds don’t naturally drift towards curiosity. We don’t instinctively wonder what might have led someone to act that way. We don’t pause to consider whether our interpretation is even accurate.

Instead, we punish. Sometimes overtly—through confrontation, exclusion, or retaliation. More often subtly—through coldness, withdrawal, gossip, or that particular tone of voice designed to make someone feel small.

This isn’t a character flaw unique to certain people. It’s baked into us. Researchers studying cooperation have found that humans are ‘altruistic punishers’—we’re willing to incur personal costs just to see wrongdoers face consequences. We’ll sacrifice our own resources, time, and peace of mind to ensure someone ‘gets what they deserve’.

Why We’re Wired This Way

Evolutionary psychologists suggest this instinct served important social functions. In small tribal groups, punishment helped maintain cooperation. If freeloaders and cheaters could act without consequence, the whole social fabric would unravel. The threat of punishment—ostracism, violence, shame—kept people in line.

There’s also something deeply satisfying about punishment. Neuroimaging studies have shown that anticipating punishment for those who’ve wronged us activates the brain’s reward centres. We get a dopamine hit from imagining justice being served. Revenge, it turns out, really is sweet—at least in the moment.

The Problem With Our Default Setting

But here’s where things get complicated. Our ancestral environment was radically different from our current one. We now encounter thousands of people, most of whom we’ll never see again. We interact across cultures, contexts, and communication styles that make misunderstanding almost inevitable. We judge strangers based on thirty-second interactions and assign them to moral categories.

And our punishment instinct? It doesn’t adjust for any of this.

Consider what happens when someone cuts you off in traffic. Your immediate reaction likely involves anger, perhaps a flash of road rage. You might sound the horn aggressively, gesture, tailgate them in return. In that moment, they’re not a person who might be rushing to an emergency, or who simply made a mistake, or who is having the worst day of their life. They’re an offender deserving punishment.

Or think about social media. Someone posts an opinion you find foolish or offensive. The urge to pile on, to correct, to mock—it’s almost irresistible. We don’t see a complex human with a lifetime of experiences that shaped their view. We see a target.

This tendency to attribute others’ behaviour to their character rather than their circumstances is so pervasive that psychologists have given it a name: the fundamental attribution error. We assume people do bad things because they are bad, rather than because they’re responding to pressures we can’t see.

The Costs We Rarely Count

Our punitive instincts come with hidden costs.

We often punish the wrong people. Behaviour that looks intentional is frequently accidental. What seems like rudeness might be preoccupation or cultural difference or social anxiety. By defaulting to punishment, we regularly inflict harm on people who don’t deserve it.

We escalate rather than resolve. Punishment tends to breed resentment, not reflection. The person you punish rarely thinks, ‘They’re right, I should change.’ More often, they think, ‘What’s their problem?’ and the cycle continues.

We damage ourselves. Holding onto grievances, plotting revenge, maintaining anger—these states are exhausting. Research consistently shows that people who forgive are happier and healthier than those who nurse their resentments. Meta-analyses have found significant positive correlations between forgiveness and both physical and psychological well-being.

We miss opportunities for connection. Sometimes the person who behaved badly is struggling. Sometimes they simply don’t know better. A response rooted in curiosity rather than condemnation can transform a moment of conflict into one of understanding.

The Alternatives We Overlook

What would it look like to have a different default? Not naivety—there are genuine bad actors who require real consequences. But a willingness to consider other possibilities before reaching for punishment.

We might start with curiosity: What might explain this behaviour that doesn’t assume malicious intent?

We might try communication: Actually asking someone why they did something before deciding what they deserve.

We might practise proportionality: Matching our response to the actual severity of the offence, rather than the intensity of our initial emotional reaction.

We might extend grace: Remembering that we, too, have behaved badly at times. We’ve been thoughtless, rude, selfish, and hurtful—often without realising it. We’ve been the villain in someone else’s story without knowing it.

A Different Kind of Strength

There’s a misconception that punishment is strong and forgiveness is weak. That to let something go is to be a pushover.

But consider how much strength it takes to interrupt your own anger. To pause before reacting (cf. the Semantic Pause). To choose understanding when judgement is so much easier. To absorb an offence without passing it on.

This isn’t about being a doormat. It’s about recognising that our automatic responses aren’t always the wisest ones. That the visceral satisfaction of punishment is usually brief, while its consequences can linger.

Conclusion: The Verdict Is Never Final

We can’t entirely rewire ourselves. The punishment instinct will always be there, tugging at us when someone wrongs us. That’s fine. Acknowledging its presence is the first step to not being controlled by it.

What we can do is slow down. Create a space between stimulus and response. Ask ourselves: Is punishment really what this moment calls for? Will it actually make things better? Or am I just feeding an ancient impulse that no longer serves me?

The truth is, most unwelcome behaviours don’t require punishment. They require patience, or conversation, or boundaries, or simply moving on. They invite us to remember that the person in front of us is as complex and confused and imperfect as we are.

And that’s not a verdict. It’s just a fact.


Further Reading

de Quervain, D. J.-F., Fischbacher, U., Treyer, V., Schellhammer, M., Schnyder, U., Buck, A., & Fehr, E. (2004). The neural basis of altruistic punishment. Science, 305(5688), 1254–1258. https://doi.org/10.1126/science.1100735

Du, E., & Chang, S. W. C. (2015). Neural components of altruistic punishment. Frontiers in Neuroscience, 9, Article 26. https://doi.org/10.3389/fnins.2015.00026

Fehr, E., & Gächter, S. (2002). Altruistic punishment in humans. Nature, 415(6868), 137–140. https://doi.org/10.1038/415137a

Lee, Y.-R., & Enright, R. D. (2019). A meta-analysis of the association between forgiveness of others and physical health. Psychology & Health, 34(5), 626–643. https://doi.org/10.1080/08870446.2018.1554185

Rasmussen, K. R., Stackhouse, M., Boon, S. D., Comstock, K., & Ross, R. (2019). Meta-analytic connections between forgiveness and health: The moderating effects of forgiveness-related distinctions. Psychology & Health, 34(5), 515–534. https://doi.org/10.1080/08870446.2018.1545906

Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortions in the attribution process. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 10, pp. 173–220). Academic Press. https://doi.org/10.1016/S0065-2601(08)60357-3

Wade, N. G., Hoyt, W. T., Kidwell, J. E. M., & Worthington, E. L., Jr. (2014). Efficacy of psychotherapeutic interventions to promote forgiveness: A meta-analysis. Journal of Consulting and Clinical Psychology, 82(1), 154–170. https://doi.org/10.1037/a0035268