About
I like building and making things. There is no prosperity without freedom. I write about…
Articles by Paul
Activity
5K followers
-
Paul Ford posted thisWe introduce ClawBench, a benchmark that evaluates AI browser agents on 153 real-world everyday tasks across 144 live websites. Unlike synthetic benchmarks, ClawBench tests agents on actual production platforms. Key findings: - The best model (Claude Sonnet 4.6) achieves only 33.3% success rate - GLM-5 (Zhipu AI) comes second at 24.2% — surprisingly strong for a text-only model - Finance and Academic tasks are easier (50% for the best model); Travel and Dev tasks are much harder - No model exceeds 50% in any category — there's a long way to go https://lnkd.in/euza3dqn
-
Paul Ford reposted thisPaul Ford reposted thisEach of these events started with an email or a DM or a chat. Dozens of people have put in huge amounts of work to make them happen, finding spaces, budgets, permissions and people to attend. I’m incredibly grateful to all of them, and to the people who come along - it’s such a strange and wonderful privilege to see you make amazing music with this weird little box.
-
Paul Ford reposted thisPaul Ford reposted thisSome news! This weekend I launched openjournalism.news — a site and newsletter featuring biweekly roundups of new public GitHub repositories from news organizations. Every two weeks, I highlight what newsrooms have shared: new tools, data notebooks, AI experiments, helpful libraries, and so on. The raw material comes from the Open Journalism Bot, a BlueSky bot that Claude and I built that monitors ~350 news org GitHub accounts and posts whenever one creates a new public repo. The bot catches everything. The website and newsletter are where I highlight the projects I find particularly interesting. Why bother? Last year, Ben Welsh and I published research showing that open-source culture in newsrooms has collapsed — new public repos dropped more than 80% from their peak. But some newsrooms are still working in the open, and they deserve to be celebrated. Follow along at https://lnkd.in/ekS6muYt, or follow @openjournalism.bsky.social on BlueSky for real-time updates.
-
Paul Ford shared thisI thought this was just the swellest conversation because Hilary Mason has always thought things through much farther than I have but also remains focused on the actual doing of things. This is an unusual combination! And hers is a rare non-toxic vision of a future where the bots are there to help us have fun—and creative people get rewarded, too.Paul Ford shared this"The original sin of AI user experience design was ChatGPT," says Hilary Mason. "I would submit that chat is a relatively poor product experience." On the latest Aboard Podcast, the CEO of Hidden Door digs into product—the good and the bad—in the age of AI. Check out the whole conversation: https://lnkd.in/ex9C-A27
-
Paul Ford reposted thisI got to join Paul Ford and Rich Ziade on their Aboard podcast! And they let me have 🌶️ takes on AI product design!Paul Ford reposted this"The original sin of AI user experience design was ChatGPT," says Hilary Mason. "I would submit that chat is a relatively poor product experience." On the latest Aboard Podcast, the CEO of Hidden Door digs into product—the good and the bad—in the age of AI. Check out the whole conversation: https://lnkd.in/ex9C-A27
-
Paul Ford reposted thisPaul Ford reposted this"Change is coming, for sure, but if you take a step back, you can see that the moats are still there, and the castles aren’t quite collapsing, no matter how much some people would like to see them fall." In this week's newsletter, Paul Ford assesses competitive moats in the current AI landscape: https://lnkd.in/eX8ScsK8
-
Paul Ford shared thisSequoia: “For every dollar spent on software, six are spent on services.”
-
Paul Ford shared thisDeloitte: “HR FOR AGENTS. As agents mature within job functions, organizations will need equally mature approaches to managing them. This will likely require an entirely new framework for managing agents that not only leans on traditional human resource management concepts for areas where agents share similarities with human workers, but also diverges to account for their unique characteristics. Some areas of focus for HR, such as workplace culture, employee loyalty, and worker motivation, won't be applicable to agents but will remain key pillars of how organizations manage their human staff. Other features of worker management can be extended to apply to agents, even if they look slightly different.”The agentic reality check: Preparing for a silicon-based workforceThe agentic reality check: Preparing for a silicon-based workforce
-
Paul Ford reposted thisPaul Ford reposted thisI wanted to share a weekend project I just shipped: an NYC pothole reporter button embedded in my car's dashboard! It's powered by an ESP32 microprocessor which sends GPS coordinates to a cloud server via a long-range LoRa radio on button-push. From there, an agentic browser does the heavy lifting of actually filling out the DOT's pothole reporting form. (And then, y'know, actual DOT employees come out and fill the potholes, ideally.) This was my first experiment combining agentic, task-based web APIs with embedded computing, but it's easy to imagine a zillion more practical use cases. Even a straightforward and discrete task like reporting a pothole is still complex: the NYC DOT uses an older web form with multiple pages, finicky geocoding and lots of error states. It's annoying to fill out manually and surprisingly challenging to reverse-engineer and fill out programmatically. So it's super neat to boil all of this complexity down to quite literally the push of a button with the help of an LLM. And since the NYC DOT actually does fill potholes swiftly once you report them (thanks, Mamdani!) -- it's super satisfying to quite literally see the impact of this project on the streetscape. If you're curious to learn more, I blogged about it here! https://lnkd.in/e2Nz7MWT
-
Paul Ford liked thisMy dad died last week after a long war and then a short battle with cancer. Last year I stepped away for 6 months to help care for a relative in hospice. Starting the summer before that my husband lost, in succession, his dad, then his brother, then his mom. It's been a hard 24 months. My friend Ben asked yesterday what I've learned about running a startup through this. Hadn't thought about it, but he's right: how to manage hospice and death while running a startup is rarely talked about. Consider this a hard earned checklist. Maybe even save it. The starting point: ask yourself which decisions truly require you, and what would break if you stepped away. Everything else should be delegated, delayed, or dropped. Your job is to become a control tower, not a pilot. You don’t need to make a big announcement. Your team and one or two more people you trust should know enough to understand capacity is temporarily reduced. The message is simple: you're dealing with something at home, you may be slower on some things, and here is what you are prioritizing. You need an operating system with your team: a regular check-in, a weekly priorities document, and distributed decision ownership. Without that structure, things will break when you are least able to absorb it. The goal is to reduce dependence on your attention. Shorten your time horizon. Instead of focusing on the next 12 to 18 months, focus on the next two to four weeks. Protect cash runway, customer commitments, and team stability. As an advisor once told me “Your job at this stage is to keep this thing alive.” If you can't do all three things, then breathe through it and do the best you can. Protect your cognitive bandwidth. In a startup you're already constant context switching; personal stress makes that harder to manage. Batch communication instead of responding all day. Decline or delegate nonessential meetings. Make good-enough decisions on lower-stakes issues instead of draining yourself trying to optimize everything. The goal is to prevent decision fatigue from becoming the thing that takes you down. Your performance may dip for a while. The mistake is pretending otherwise. And resist the instinct to handle it alone. Isolation makes hard periods harder. Your people can help you keep perspective and make steadier decisions. In moments like this that matters a lot. And use it: a crisis often exposes the places where decision-making is too centralized or where systems are fragile. It's a useful stress test of where you have failure points and need better structures. It shows who is poised for growth. It's important to remember is that this is a phase, not a permanent weakness. The goal is not perfect execution. The goal is to get through it with the company still viable and growing, however slightly. I'm so grateful to my team, and to our supporters, and to my husband. And to Benjamin Richman for asking a thoughtful question. Looking forward to seeing what he's building.
-
Paul Ford liked thisPaul Ford liked thisThe latest Ad Age Agency A-List dropped a few days ago. What caught my eye? Of course, the continued rise of independents, but also the fact that big agency networks with dozens or hundreds of offices are fully out — not just unfashionable but totally off the top-ten list for the second year in a row. Seems like a development worth pondering. A little research found that the last time a holding company-owned global network sniffed the top spots was in 2024, when Weber Shandwick rode the ritual sacrifice of a Pop-Tart mascot to ad trade glory. As recently as 2020, seven of the top 10 spots belonged to owned agencies, including five global networks. By 2026, it's down to four owned agencies, and not one of them is a global network. Instead, a cadre of mainstay independents dominated, including micronetwork Wieden & Kennedy. Meanwhile, the holding companies are represented by the smaller footprints of Goodby Silverstein & Partners and BBH. Worth noting: this is the first A-List since the Omnicom-IPG merger eliminated thousands of jobs and retired global networks FCB, DDB, and MullenLowe as independent brands. The A-List matters (and it's highly sought after) because it's the closest thing we have to an all-purpose barometer in an industry that's translucent at best, with almost no public financial data aside from earnings reports. It measures agencies at the intersection of real creative work (as opposed to awards bait), client results, and revenue growth — imperfect, self-reported, but the best read we have. This year, the detail I found to be most interesting were the demonstrations of independence in action. Rethink, agency of the year, built a trust structure that makes the agency legally unsellable. Tombras is third-generation family-owned with an 80-year time horizon — practically a geological timescale in agency terms. Mischief carries no debt, so no one has leverage over their decisions. Highdive took private-equity investment and still turned down 100+ pitches last year while growing 24%. Compare that to the fate of big networks and, increasingly, it feels like the industry has split into two distinct businesses that happen to share a name. One optimizes for scale, media, and data infrastructure, and investor returns — and it controls the majority of client spend across media, CRM, and the infrastructure layer that's genuinely hard to unbundle. The other optimizes for creative output — and it's won the industry's affections. Whether this share of hearts and minds translates to share of wallet is a question that will define the next decade of this business. And of course there's the flip side of the coin: whether the big global players can once again couple creativity with scale. (Disclosures: As a former editor at Ad Age, I used to work on the A-List. And as a former English major, I can't count, so check my math.)
-
Paul Ford liked thisPaul Ford liked thisThis morning, Spencer Glendon spoke at ClimateTech Connect, sharing climate maps of our probable futures and exploring what these changes could mean for communities and decision-makers. Thank you to ClimateTech Connect for the opportunity to share climate data, along with insights into risk and adaptation, with a broad audience.
Experience & Education
-
Aboard
View Paul’s full experience
See their title, tenure and more.
or
Publications
-
Then Again, Maybe I Will
Slate
See publicationA review of "Why Have Children?" by Christine Overall for Slate's relaunched books section.
-
The Age of Mechanical Reproduction
The Morning News (later published, translated, in Internazionale [Italy])
See publicationNamed as one of the best pieces of 2011 by the staff of the New York Times Magazine, others.
-
Just Like Heaven
The Morning News (later republished in Harper's Magazine)
See publication"Experts answer what they know. The Non-Expert answers anything. This week we answer that eternal question: What happens after we die?"
-
Best Music Writing, 2009 ed. Daphne Carr and Greil Marcus
Da Capo
See publicationA piece I wrote for the Morning News (http://www.themorningnews.org/article/six-word-reviews-of-763-sxsw-mp3s) was selected for this anthology.
-
Miscellaneous spoken word pieces
NPR
See publicationFrom 2003-2006 I was a commentator on All Things Considered. Pieces included:
- Internet as Archive, If You Have the Time
- Losing My Music, Or When Good Hard Drives Go Bad
- Privacy in Cyberspace: Is It Possible?
- Stealing Identities, and Ideas
- Distracted No More: Going Back to Basics
- Memory and the Virtual World
- Commentary: Be Proud of Guilty Pleasures
- Commentary: Obsolescence, the Killer App
- Commentary: Tilting for iTunes
- Commentary: Age of…From 2003-2006 I was a commentator on All Things Considered. Pieces included:
- Internet as Archive, If You Have the Time
- Losing My Music, Or When Good Hard Drives Go Bad
- Privacy in Cyberspace: Is It Possible?
- Stealing Identities, and Ideas
- Distracted No More: Going Back to Basics
- Memory and the Virtual World
- Commentary: Be Proud of Guilty Pleasures
- Commentary: Obsolescence, the Killer App
- Commentary: Tilting for iTunes
- Commentary: Age of 'X'
- Commentary: Early Periodicals Go Online
- Commentary: New York Stories
- Commentary: Decoding the Money Puzzle
- Commentary: Web Standards
- Commentary: A Religious Conversion
- The Truth About JT Leroy
- Commentary: The Milt
- In the Legal Wars Over Internet Downloads, Could TV Be Next ...
- Commentary: All About Eve, and Other Mothers
- Commentary: A Theory of Time
- Commentary: Software
-
Best Software Writing, 2005, ed. Joel Spolsky
http://www.amazon.com/Best-Software-Writing-Selected-Introduced/dp/1590595009
See publicationContains a piece on the language Processing.
-
Facebook and the Epiphanator: An End to Endings?
New York
See publicationSelected as one of the best pieces of 2011 by Alexander Chee (novelist) and Matt Pearce (Los Angeles Times); named among others as " Best Media Writing" in Advertising Age. Nominated for (but did not win, alas) an S.I. Newhouse School Mirror Award. Cheerfully discussed on On The Media (http://www.onthemedia.org/2012/oct/26/facebook-versus-epiphanator/transcript/)
-
The Web is a Customer Service Medium
Self-published
See publicationNamed as one of the best pieces of 2011 by the Awl, by Ben Williams of New York Magazine, and in the Best Tech Writing of 2011 by the Verge, and as a favorite read in 2011 by Venturebeat.
Honors & Awards
-
Winner, Best Single Topic Issue
American Society of Magazine Editors (ASME)
I wrote an entire issue of Businessweek about code. 40+ people worked to make it something amazing. The issue won the "ellie" award for best single-topic issue.
http://www.magazine.org/asme/ellies-2016-finalists-announced -
Included in anthology
Best Music Writing
-
Included in anthology
Best Software Writing
Organizations
-
Association for Computing Machinery
Member
Recommendations received
6 people have recommended Paul
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content