We had our annual general meeting right before Thanksgiving, and every year I share some of what we covered. Last year, I shared this post about three potential states of the venture capital world, as I felt we were in the middle of significant shifts in the early-stage venture capital market. This year didn’t have a unifying theme, but I did discuss making sense of the returns potential for multi-billion-dollar VC funds, how AI companies are setting the bar for everyone when it comes to fundraising, and how outbound sourcing has changed over time.

The Math on How Multi-Billion Venture Funds Can Work - We Need $1T Companies

There has been a lot written about whether multi-billion-dollar VC funds can deliver meaningful returns for LPs. There is one loud camp that talks about the law of large numbers and how difficult it will be for those funds to generate attractive returns, whether measured by IRR or cash-on-cash multiple. In the back half of the year, I spent some time talking to LPs who are investors in these funds or looked closely at investing, as I was curious about how they thought the math could work - after all, they are the folks who are putting and keeping these multi-billion-dollar funds in business.

The LPs who made the strongest arguments for why these funds could work focused on rethinking the upper bound of terminal outcomes for VC-backed companies. We already have several private VC-backed companies valued at over $100B (including most of the Elonosphere companies, OpenAI, Stripe, Databricks, and Anthropic, to name a few). The math for a given large VC fund looks far less daunting if you believe we will see another step change increase in the upper bound of what’s achievable for private companies.

What if we are on the cusp of an era where we see $1 trillion private companies? While that seems a bit hard to imagine, it’s worth noting that 10-15 years ago, $1B private companies were rare enough that we called them unicorns.

I’ve lived through a variation of this myself. When I started working on Precursor in 2014, most of the portfolio modeling assumptions LPs used assumed that $1B valuations for private companies were rare and would remain rare. In such a world, it would be hard for funds like mine with larger portfolios and lower average ownership to deliver great returns. Thankfully, we are in a world with far more private companies valued at $1B or more than modeling predicted back then.

AI Companies Set the Terms for Everyone’s Fundraising

One of the most common conversations I had with founders this year was what it takes to raise money in this market. Particularly, those who are not in AI want to know why these AI companies are raising so much money so quickly while non-AI companies struggle to get attention. Most founders who ask this question underestimate the asymmetry of the VC business. Related to the point above about $1T companies, most venture capitalists are looking for companies they think will be outliers, not just good, solid companies.

Right now, rapid AI-driven traction is what gets VCs’ attention. Founders have figured this out, so they are doing whatever they can to generate steep revenue ramps over short periods of time to get venture capitalists’ attention. In some cases, what is being counted as “revenue” wouldn’t meet the traditional tests that we used for SaaS companies. But we are in a moment where deeply investigating the quality of a revenue number might be at odds with winning an allocation in a hot deal.

The asymmetry is a big part of what’s happening here. Put yourself in the shoes of a venture capitalist looking for massive home run outcomes. Is it easier to believe that a company that appears to be growing will continue to grow or that a company that isn’t growing as quickly will begin to grow quickly when it gets access to your capital? Betting on something that’s already working and continues to work could return your whole fund. If growth slows or it never was quite what you thought it was when you invested, the amount of money you’ll lose is likely something you can absorb. It is far more costly not to be in the winners than it is to make a few false positive investments if you’re managing a large pool of capital.

By far, the greatest challenge in fundraising today is rising above the noise floor, given the level of activity in the ecosystem and the number of companies competing for investors’ attention.

The Realities of Outbound Sourcing - Founders Understand the Game, Too

Once upon a time, firms that were effective at outbound messaging to founders had an edge, as many firms didn’t have the capabilities to execute it or any interest in making it a key part of their founder sourcing strategy.

Fast forward to today, many venture firms are very focused on outbound sourcing to find the most promising companies even earlier. The spectrum ranges from firms with entire software teams building custom tooling to those that use off-the-shelf software with custom queries. What was once an edge is not a standard tool in the VC sourcing arsenal.

Over the past few years, I have noticed a substantial increase in the number of inbound messages the portfolio founders we work with receive. Some of the messages are very well written, while others feel more like generic, trigger-based outbound templates. The reason I wanted to talk about this at our AGM is that I’ve seen more and more people who have learned how these trigger-based early-detection systems work. They’ve become quite strategic about when they send out the subtle signals they know will result in more attention and inbound.

If there was an era when simply having an early signal put you at the front of the line for meeting with a new founder, that era is long over. In addition to early signal, your outreach needs to come with more than just a recognition that you know or suspect the person might be starting a new company - the best founders are discerning and being first or early doesn’t mean what it used to in a world where everyone has their own system for surfacing people who are on the verge of starting new companies.

Many articles about AI’s cognitive effects focus on the negative potential consequences of overreliance on AI - brain obesity, reduced creativity, and intellectual laziness. Based on data from a report by OpenAI1, these concerns are likely valid given that many people are using ChatGPT and other LLMs to ask questions (as they would of a traditional search engine) or get advice, as opposed to taking action and doing things on their behalf. I’ve been using AI in a much different way; I’ve been building tools that help me in the many investment and management tasks that take up a significant chunk of my workday. Instead of treating AI like a search engine, I’ve enlisted it to be my thought partner and assistant. I’ve learned so much during this process, and it is very much at odds with the idea that using LLMs is an adult version of brain rot. Teaching my AI assistants how I think has forced me to clarify, and in some cases update, the way I think and work2.

Over the past few months, I’ve really gotten into building AI assistants and tools to help me with many of the investing and management parts of my job. Along the way, I’ve learned a bit of Python (my first time learning a programming language), learned how terminal commands work so I can build with Claude Code and Cursor, expanded my knowledge of Markdown, learned how to use MCP with Claude Desktop to get access to a lot of existing data I need for analytical projects, and built a rudimentary version control system in GitHub to help me manage the iterative nature of what I have been building. I haven’t been writing here as much because The Learning Corner, the podcast I co-host with my colleague, Mia, has been a good outlet for me to talk through many of my ideas on the current state of the VC market. The time I used to spend writing has (for now) been replaced by building things in AI and learning how to get better at getting more out of LLMs.

Developer Workflows Finally Make Sense to Me

On average, I’m now spending a few hours per day building and refining tools to help me in my daily work.

A big inspiration for this post was a LinkedIn post I read from Chang Xu at Basis Set Ventures. The post talks about how many non-developer knowledge workers are starting to build workflows that mirror what developers have been doing for a long time. I could have spent time learning Python without a coding assistant, learning how version control systems like GitHub work, the basics of Markdown, and some basic terminal commands, before modern AI tooling arrived. I never did it before because the barrier to learning those things felt really high, and the relevance to my daily work felt really low. The only reason to spend the time would have been for pure intellectual satisfaction, and I never got over the hump on that front. AI assistants changed that calculus entirely for me once I started pushing myself to use them more.

The Joy of Being a Novice is Real

Today, I’m learning to apply these concepts to my daily work, and they feel very relevant to what I’m doing. I’m now using quite a few tools every day that were not part of my daily workflow, including Cursor and Claude Code for small Python projects and agents, GitHub for version control of Markdown files, Claude Desktop with MCP and Skills for building assistants and projects, WisprFlow for text dictation, and Obsidian for markdown file creation and editing. When I reflect on why this has been so enjoyable and engrossing and taken up more and more of my time, a few things stand out:

• I started this journey as a novice with no programming background and limited knowledge of the tools I use today; I didn’t expect to be good on day one because much of this was new to me. That turned out to be a very liberating place from which to start when learning something new.

• The tooling is actually pretty good at nudging you toward improvement and helping you get more out of what it can do once you learn how to ask for help and structure things to get the best bang for the buck. You can really learn quickly here once you figure out how LLMs work.

• I feel like I’m part of a large community of people on the internet who are also trying to figure this out. Many have shared helpful tips and tricks, but there are far fewer true experts here, given how new the technology is.

• I have really enjoyed the iterative nature of learning. I started out doing very basic things, and have gotten more sophisticated as I’ve learned more about how the tech works and what it can and can’t do well. You don’t have to understand everything or use every tool on day one; you can work your way toward complexity.

To make this more concrete, here’s how I would summarize my journey of learning how to use Claude Desktop as my primary tool:

1. I began by writing simple prompts and then refined them significantly in chat. This felt really inefficient after a few weeks, so I learned more about writing better prompts and how to structure them more effectively.

2. Next, I wrote prompts that I regularly used and turned them into projects. The projects still required a lot of post-prompt chat to get to the answers I wanted, but it was still an improvement over where I started.

3. Then, I discovered project instructions, and that was a revelation. Projects became much easier to manage, but I was still writing very long project instructions to get projects to work.

4. Later, a friend showed me how to rethink project instructions. I started creating Markdown files I could swap out to handle more of the logic. This simplified the project instructions, resulting in huge gains, but it's still not perfect.

5. Recently, I discovered how to create more robust project logic files in Markdown, separate from project instructions, and how to refine them as I gained a better understanding of what I wanted the system to do and deliver.

6. Managing version control in Obsidian became too tedious, so I migrated the project logic's version control (and other core Markdown files) to GitHub to share them with others and track changes over time.

One thing I’ve learned is that many people who want to build stuff in AI are intimidated. They see fully polished, functional projects that others have created, and they assume the final product looks a lot like the original version they’ve been shown. In my experience, nothing could be farther from the case; these things emerge over time, and the starting point and end point often look very different.

Teaching AI to Think Like Me Made Me Better

So was all of this worth it? Yes, but not for the reasons I expected.

The most significant benefit hasn’t been time saved or better decision support. It’s been that teaching Claude to be a good and useful partner forced me to be explicit about how I make decisions about founders, markets, and companies. I’ve always had a model in my head for how I weigh information, but I’d never really articulated it to anyone - including myself.

To make this more concrete, I recently built a reference-checking skill in Claude to help me identify where to spend my time and energy when checking references. I built the skill to cover all the use cases where I want references: founders we’re evaluating, executives interviewing at our portfolio companies, and prospective lead investors for follow-on rounds where I don’t have a strong prior relationship. As I built out the branching logic, I realized something uncomfortable: my beliefs about where to find the most accurate references had evolved significantly over the years, but I was still following some old practices.

One particular area where my beliefs have changed is how to handle founders who have raised capital for a previous venture and are now raising money for their new company. Two things became clear as I started to articulate my reference logic to my AI companion:

1. First, the seed ecosystem has grown dramatically - there are many more founders who have raised VC money before, and I hadn’t digested what that meant for our reference-checking process.

2. In the last ten years, I’ve learned there is much more nuance about why investors and founders choose to partner again on their next venture; it’s dangerous to assume that you know why two people have chosen not to work together again without knowing more about the outcome and the relationship.

Building the tool forced me to confront this drift between what I used to believe and what I’ve learned over time. The branching logic wouldn’t let me be vague or rely on intuition - I had to articulate clear rules about which reference sources to prioritize, which questions mattered most, and where my time would be best spent. Seeing it all written out made the gaps obvious. I’ve made this change, and a bunch of others, as a result of having to explain how I think and reflect on how I feel about the words I see on the page.

I’ve had similar experiences with other tools I’ve built. The systems I’ve built sometimes produce results that surprise me, forcing me to re-evaluate how I weigh factors in the model I’ve built. Each surprise signals that my implicit model and explicit reasoning have diverged.

This is a different value proposition for AI than the efficiency-and-automation narrative that dominates most conversations. For me, the real insight isn’t that AI can do your job (I don’t believe it can, yet) - it’s that trying to teach AI how you think and make decisions forces you to make the implicit ideas you hold explicit. The act of making implicit thinking explicit doesn’t just help the AI understand you. It forces you to understand yourself.

I started this journey worried I might become intellectually lazy. I’ve found the opposite: building with AI has made me sharper, not by offloading thinking, but by forcing me to examine thinking I’d previously taken for granted.

Share Your Thoughts on LinkedIn

When I was raising our second fund in 2017, one of my LPs asked me a bunch of questions about how conflicts of interest work in venture capital. He came from the world of hedge funds and credit investing, and the notion that making an initial investment in one company would be a blocker to investing in a second company that appeared better positioned to win struck him as strange; that’s not how his business worked an he found it strange that I, and other VCs, would agree to tie our hands in this way. I told him the whole backstory on how venture works and why this norm existed. The TL;DR version of that conversation was that venture capital firms invest in private companies and often take board seats. Part of that arrangement grants the investor access to private information about the company's performance, product roadmap, and other sensitive details regarding the company’s plans. It’s not practical to have access to the same level of information about two competing companies while being a good fiduciary to each. I also believed the “no conflicts” norm forced firms to ask hard questions about whether they were willing to invest in a given company, knowing that it would foreclose the opportunity to invest in the known and unknown set of present and future competitors. I’m not sure that he found the answer satisfying, but it made sense to me at the time.

That being said, I think the norm of not investing in competitive companies has been a feature, not a bug, for our industry for a few important reasons:

1. It sends a signal to the entrepreneur that if you take a given firm’s money, they will not turn around and back your competitor. It’s a trust signal and a vote of confidence in the company, its vision, and its prospects relative to others in the same category.

2. It gives the investor an incentive to devote all their energy and effort to making the company successful, rather than exploring other investments in the category.

3. Starting at the seed stage, most investments come with the expectation of a board seat for the lead investor. Board seats come with duties and responsibilities that are much harder for an individual, but not a firm, to execute and fulfill if they are also involved with another company that sells a similar product to the same customer base.

4. This was not a really expensive policy to maintain when it wasn’t essential to be in every winning company. Being in the second or third best company in a category instead of the winner wasn’t material when funds were smaller, as you didn’t need to be invested in the far outliers to generate returns from smaller funds.

As venture fund sizes keep getting larger, I do think that tradition of venture firms having a norm (if not a stated policy) to not invest in competitive companies is likely to go away. This is simply a function of the fact that as venture funds have grown larger, it has become increasingly essential for those firms to be associated with the biggest and most important companies. The larger the fund, the more important it is to be an investor in the companies that are true outliers; there is no way to make the fund math work if you are not in those companies unless you are in other, similarly-situated companies. My sense is that there is more money chasing outliers at the moment than there are outlier companies to fund.

In many ways, this was a norm or policy that, although not entirely free to implement and support, was not prohibitively expensive for most firms to put in place, maintain, and operate as a general business principle. Given the current state of fund sizes, I wonder if that will continue to be true. The only way to return some of these really large pools of capital that are being raised in our business is to be in the biggest, most consequential companies that are created. If you missed them at Series A or Series B, can you afford to stand by your norms and principles and not secure a position in those companies? Here’s how I think most firms are dealing with this today:

1. Use a seed fund or scout strategy to meet as many promising, early-stage companies as you can. This gives good market coverage and access to what’s happening on the ground. There’s the risk of investing in too many companies that could block you from investing in the winner in the future, but this is an acceptable risk for most funds.

2. Focus on investing in Series A and Series B (instead of seed) rounds and pay up to get into the winners when it’s clear which companies are working. This helps reduce the odds of conflicting investments but pushes up the average entry price. This is a workable model so long as the terminal values are large relative to entry price.

At some point, this will break. There will be important companies where big funds can’t get primary ownership at the Series A or Series B, while also respecting conflict norms. I can see a few ways to finesse this issue and break the norm without creating unmanageable challenges:

1. Buy secondary positions in the companies that matter but that you missed - Buying passive secondary stakes in companies that you missed but want to own could work. Those passive, mostly common stock positions don’t come with board seats, information rights, or control provisions that could create conflicts. This solves for the financial impact of missing out on those companies.

2. Invest in competitors but have different investors take board seats and create firewalls to limit information spillover - While conflicts of interest are hard to manage at the individual investor level, there are more options at the firm level. There is a world in which firms tell founders that these conflicts can be managed by having different people work on each investment and having internal firewalls that prevent information leakage. We will see if founders believe these internal processes are effective.

One thing we haven’t discussed that’s particularly important is the founder's perspective! I don’t think founders want the investors who back them investing in competitors. This is an issue where the business model for funds is at odds with what most founders want. To be clear, most founders don’t want their existing investors to invest in directly competitive companies. Most founders lack significant “hard power” (i.e., the right to block an investment) in these negotiations; funds can and do invest in competitors if they choose to do so. However, there will always be a set of founders who possess soft power and will utilize it to encourage their investors not to engage in such behavior. The universe of founders with meaningful soft power to influence this is very small, but that universe of founders is very powerful.

I am very curious to see how this plays out in the coming years. I am happy to hear thoughts that others have on this topic.

Comment in LinkedIn

I’ve been in many rooms and events with other VCs and tech executives, and every conversation naturally gravitates toward discussing the present realities and future possibilities of AI. These conversations have been interesting, but most have left me uneasy about a handful of questions that have gone unanswered or been avoided altogether. These aren’t new questions, but they certainly feel more urgent and timely to me lately based on the tone and tenor of those conversations.

We Encouraged A Lot of People to Learn to Code - Now What?

Starting about 10 or 15 years ago, there was a concerted effort to get many people to learn to code to get jobs in the tech industry, whether those folks worked for a big tech company or a traditional big company trying to modernize its operations. The logic was that getting people to learn computer science at university or how to code outside of traditional universities was an excellent idea and would open pathways to good, in-demand jobs. There are many valid educational reasons to learn to code or to study STEM fields beyond the economic payoff. Still, much of the rhetoric focused on employment opportunities in the tech industry as the primary driver for this focus. I do not have a degree in CS and have never been a professional software developer, so take my views with a healthy grain of salt. Still, I wonder whether we oversold people the transformative nature of the opportunities here given how quickly the landscape for entry-level software developers has changed in the last 24 months.

AI-driven white collar job destruction is happening much faster than predicted even last year.

I’ve been reading a handful of these CEO letters about how they are pivoting their larger, established companies to be AI-first. Two that stand out are this one from the Shopify CEO and this one from the Duolingo CEO. While neither memo explicitly talks about replacing employees with AI, the discourse around these articles suggests that informed employees see the writing on the wall and know that AI is coming for many of their jobs even sooner than people thought a few years ago. Whatever comfort there might have been that AI would take a while to go for white collar jobs in sales, customer support, and HR has been dispelled. Whether the technology is entirely up to the task or not, CEOs are pushing hard to get the most out of AI-powered efficiency and cost improvements, and that trend is only accelerating. For every published memo we read, many more are being written or haven’t been shared.

Who is going to take responsibility for the labor market displacement?

This is the one theme that nobody seems to want to touch. Companies pursuing aggressive AI deployments often do so to get more efficient and compete with other companies offering their customers AI-powered solutions; it’s hard to compete if you aren’t pushing the envelope and staying on the cutting edge. I don’t expect profit-maximizing private and public companies to feel deep responsibility for what this change means for the segment of the workforce impacted by this transition; we don’t live in an era where that is a genuine concern for most companies and I believe most employees are fairly clear-eyed about the social contract between them and their employers and the nature of their employment (whether they think it’s the optimal relationship is another question for another post).

The other entity that could take responsibility is the government. I sense that the government at all levels understands AI's coming disruption, but it isn’t well situated to get ahead of its impact on the labor force. Government is deliberative and slow by design, which works well in stable times but can be a real challenge in moments of great upheaval like the one we are in now.

The impact on displaced white collar workers, which appears to be happening faster than anticipated, is a problem without an owner and kind of hot potato that nobody wants to grab or own, as it’s a tricky problem to solve and doesn’t have obvious, straightforward answers. I wonder who will step into the void to address this as it accelerates and becomes a more visible and urgent problem.

Senior and Junior People are Having Very Different Experiences

When I first wrote this section, I planned to frame it as an older versus younger worker difference, but that felt off. The real divide seems to be between executives and leadership (many of whom in the tech industry are young) and non-executive management and individual contributors. Admittedly, I spend most of my time talking to senior executives, CEOs, founders, and people generally in leadership positions. Much of what I am hearing comes from how they view AI's transformative power and what it will mean for the companies and organizations they run. They are uniquely positioned to capture the benefits of these savings, and they view their jobs and positions as generally safe from AI disruption for now.

When I talk to people who are not in management but are in white-collar occupations like software development, finance, consulting, or other high-prestige entry-level jobs, their perspective is very different from that of management; they are acutely aware that they are the eventual targets of these efficiency and cost-cutting plans. They seem to intuit that they studied, interned, and trained for a world that might not exist.

This Time Feels Different - More Work but Less Labor

At the risk of sounding like a Luddite, this time does feel different. Two of the most significant platform shifts I’ve seen in my career, the move to the cloud and then the move to mobile, created a lot more labor demand. With all of this cloud and mobile software, we needed people to manage data centers, focus on uptime, work in security centers to keep an eye on intruders, and perform various other essential tasks that arose as software penetrated new industries and parts of life. And, because there weren’t technologies that could do most of that work, the only solution was to hire and train people to do new(ish) jobs like site reliability, developer operations, and the like. From everything I’ve seen, the rise of AI is replacing labor with machines and agents because they can do most, but not yet all of the work. As is the case with any innovation, the optimist in me wants to believe that this new tech wave will create new job categories and titles that didn’t exist before, but this wave feels different because I think the companies that would otherwise create those jobs for people are going to push much harder to have software do more of the work.

I don’t have the answers to anything I have written about here - I wrote this to help me think this through. But I’ve been in many meetings lately where everyone else either doesn’t have the answer, has an answer but is unwilling to share it, or won’t say the quiet parts out loud.

Yesterday, we announced the successful close of our $66 million fifth fund at Precursor Ventures. So much has changed about the venture business in general and the seed stage venture market in particular since I first started working on pre-seed investing almost ten years ago. I feel very fortunate to have raised our fifth fund in this environment, but I also have many questions about our industry and where we are heading.

What it means to be a venture capitalist has dramatically changed over the last 10 years. Our business has become much more institutionalized than it was back then. We have many more venture capital funds than we did a decade ago. When measured by assets under management (AUM), the largest funds in our business are bigger in both absolute size and relative to the median-sized firm in our industry. What was once a cottage industry of small, boutique firms has become a more mature, concentrated industry that looks a lot more like other mature segments of the finance industry.

One consequence of this maturation is that our industry has become very prescriptive regarding what works. Limited partners have developed relatively fixed views of portfolio constructions and investment strategies that will generate returns and primarily back VC firms that conform to those models. Increasingly, venture capitalists focus on algorithmic approaches for founder selection, emphasizing folks with low employee badge numbers at well-known companies, people in their social or professional networks, or people who have otherwise achieved some level of fame or notoriety online. Maybe it’s just me, but it feels like our industry is trending in a direction where everything feels the same - firms feel the same, the founders we back feel the same, and certain parts of the business feel like they’ve devolved into a cookie-cutter approach to finding and funding businesses. This is not a good direction for an industry focused on finding and backing outlier businesses.

Some of these prescriptive processes worked well in the past. Some repeat founders do have a gift for starting multiple successful businesses. Some portfolio constructions and fund strategies have performed better than others over time. Somewhere along the line, though, these best practices have hardened into an orthodoxy that crowds out different, new ideas about how this business could work for fund managers and entrepreneurs. There are many days when it feels like there is an odd contradiction in the venture business; we are very focused on funding people who have disruptive ideas that challenge the status quo but the leaders who run VC funds and the limited partners who provide capital to funds are stuck in a dynamic that makes it very difficult to try new things when it comes to the mechanics of who we fund and how funds operate.

This feeling of unease is not new for me. I first felt this way when the venture capital business went from an industry with a relatively healthy balance between B2B and B2C companies to an industry almost exclusively invested in B2B software-as-a-service (SaaS) companies. We narrowed the focus on one sector and found a way to reduce most of these businesses to a series of core metrics that made them easy to evaluate on quantitative dimensions. We focused on customer acquisition cost (CAC), lifetime value (LTV), net revenue retention (NRR), and many other metrics that made it easy to compare these businesses in spreadsheets and databases and put them into percentiles and histogram bins to determine where they ranked relative to their peers. Somewhere along the way, these KPIs went from ways to measure these businesses to goals in and of themselves, in a weird recreation of Goodhart’s Law.

The dramatic increase in money flowing into and through the venture capital business has exacerbated this trend toward conformity. More global limited partner dollars are looking to invest in venture capital funds. Venture capital funds have expanded their appetites for LP capital and can now absorb and invest more money than ever. Ambitious startups are willing and able to raise more money than ever before. Our industry is finding ways to put more capital to work than ever before. While the venture industry has increased its ability to put more money to work, it comes at a cost.

Somewhere along the line, we traded metrics for taste. Metrics give you the comfort of absolute and relative comparisons. Venture firms can invest like algorithmic hedge funds when metrics are the only input that matters. But taste is where the magic is in this business. Taste does not scale. Taste requires a point of view about an industry, a founder, and his or her vision for what to build. Investment themes are not taste. Themes are the opposite of taste. Themes reduce taste and insight into formulas that skip the vital work of building your point of view of what will work and what good looks like. Themes are just a terrible shortcut to original thinking and taste.

Part of the inspiration for this post is the current moment. We are in a moment where we have what I call concentration squared. We have not only concentrated most of the money in this business in fewer hands; those firms that control the money have also gone all-in on a very narrow set of investment themes, including AI and defense tech. Even in an industry with many firms, I suspect most of the industry’s dollars are going to the same themes regardless of which firm writes the check. And, in my experience, most themes only produce a very small number of really meaningful winners. I worry about what this means for the venture ecosystem.

This job used to be about hunting for the most revolutionary, high-risk ideas. From the earliest internet days to the social media revolution, these ideas were so visionary that they sometimes sounded unbelievable. That promise of invention brought us the most formidable tech companies of this generation. I worry that we are losing the ability to invest in weird and wonderful new ideas that will be the great companies of the future. Determining who exactly can break those patterns is part of the magic of demonstrating taste. Taste is revealed by the decisions you make about who and what to fund and why.

Venture capital is at an inflection point. We have the opportunity to lean into new and interesting areas or continue forward toward what feels safe. I hope we choose to go where innovation flows.