Josh Wulf has been running a show called Agentic Engineering, and he asked me on. Josh spent ten years at Red Hat, three years as a tech recruiter in Brisbane, and the last year sprinting down the same road I have, sounding a bit crazy to most people around him. We talked for about an hour and a half. He cut it into four parts, and they all dropped today.

Here's the gist of all four, written up properly and in order, with chapter links at the bottom if you want to jump to a specific part of the conversation.

There's no place in the world now for a software engineer who types code by hand.

part 1: is coding over?

nobody knows what happens next

Josh opened by calling me the LLM whisperer. I'll take it, but let's be clear: no one knows what the hell will happen. I'm a little bit older, I've worked in developer tooling for a long time, and I'm just a little better informed. That's it.

Here's the timeline. December 2024, I'm a tech lead at Canva. Everyone there loves JetBrains. I start using Cursor, start running my first loops by hand, and think, "ooh, JetBrains and IDEs don't exist in the future." You get used to being called a little bit crazy. JetBrains just reported its first-ever net loss. Sure, they've been investing, but in my circles, no one is typing code into an IDE anymore. If you hang around a group that still loves their IDEs, you're not hanging around a well-informed group.

In February last year I discovered filesystem-as-state, the first of the loops: the idea of autonomously building things. It spooked me. The first thing I did with it was publish the Z80 post.

Not enough people understood that post at the time. Take an existing application, clean-room it into specifications, then rebuild the software from those specifications, much like the way AMD's CPUs came out of Intel's. It's not legal in every jurisdiction, and you have to take the right approach. But did you see all the Photoshop clones that came out in the last week? That's just the Z80 idea, implemented.

A year and a half later, services and MCPs exist whose whole job is rebuilding existing software, and a couple of published papers cite it.

The idea underneath is infinite software. If you're selling software without infrastructure or some other lock-in, you're hosed. So the real question is: what is a moat in software now? What is investable in software now?

the travelator

Then came Dear Student. ThePrimeagen covered it, the juniors started paying attention, and I ended up in SF, at the pub, at a couple of meetups, realizing I couldn't sit on Ralph any longer because someone else would write it up first. By November, Boris was talking about using Ralph and loops in Claude Code. January was the real inflection point: it crossed into the mainstream, and business leaders started to understand what was happening.

It still hasn't crossed into the dark matter developers, and that concerns me deeply. If you've been applying yourself and staying curious, you have a two-year head start on everyone else. You can't just pick up the latest model and be good at this. It takes theory of mind, and the more you apply yourself, the more you see what's possible, which makes you more ambitious, which makes you see more.

Josh's story fits that exactly. He tried autocomplete-style Copilot, thought it was rubbish, and turned it off. Then a friend who heads AI at a couple of companies told him their engineers were getting good results. The first thing Josh did was have it clone a Java testing framework into Node.js. It wasn't great, but it was maybe 80% as good as he was, and it never stopped. It doesn't need to eat or sleep, and it doesn't lose focus. So he figured out what the missing 20% was: completeness and verification, because the models would leave stubs and partial implementations. Do the 20% instead of the 100%, and you've got scale.

Same for me. I'm no maximal AI glazer. I explicitly distrust the tools. The difference is that I internalized two years ago that the line was coming, and I've been expanding what I think is possible ever since. The models from a year ago were already good enough to displace software developers, and they keep getting better. We don't get stubs anymore. It still comes down to the operator's skill.

The travelator of tech moves on, as it always has. My deepest concern is the people who haven't even stepped on it. They're not aware it's moving, and they're completely unaware that some people are sprinting on it.

And you know how the employment market works. "Five years of React experience required." They're going to put minimum years of AI experience on job ads. If you wait for the models to be perfect before you start, there'll be juniors who are cheaper than you, who have been playing with this substrate, with a whole portfolio of cool things and maybe a couple of conference talks. Same story as always. Josh, who wrote those mathematically impossible job ads as a recruiter, reckons he got two years of LLM experience into one by doing more hours every day and more days every week.

managers, burnout and retirement projects

A couple of days ago I published the post I'd been putting off for five or six months. If you're an engineering manager and you're not creating time for your team to play with AI, you're setting them up to be replaced. It got 500,000 unique views in 24 hours.

You join a company to de-risk your life. If a company has banned AI outright, quit and don't join one like it. You trade time and skills for money, and you need skills that will be employable for you and your family. You could try to change the culture, but plenty of employers are already with the program and will likely pay you more.

Josh talked about cognitive burnout: the Amazon warehouse workers who couldn't keep up with robots that never take a toilet break, except happening in your head. He treats it like the gym. You get sore, and you either stop or push through and adapt. A year in, he feels like his brain has been rewired into a carbon-silicon unit.

I call it teleporting to the future: doing all the retirement projects you dreamed about (the programming language, the filesystem, the things you'd never get paid for and do for the love of the craft) today. I've run out of retirement projects.

subsidies are a free university education

Josh brought up the quota apocalypse. Yes, I'm known as one of the people who helped introduce rate limits. Someone posted a leaderboard of credits burned on Reddit, and I was near the top. Back at GPT-5's launch, OpenAI handed out plaques to people who'd used a billion tokens. A billion tokens is nothing now.

Here's the hard bit. These LLMs are heavily subsidized, and the subsidies are drying up. A lot of what I've built was paid for by VC subsidies. When you have to pay real rates, do you reckon people will build fun little games and filesystems if the side project costs a thousand bucks? No.

You could look at it as VCs paying for everyone's personal development. If you miss this window, you miss the equivalent of a subsidised university education.

Josh asked about climate change and data centers. My take is simple: if you want to take a political stance, you can, but you'll need to start your own company and hire, say, twenty engineers to compete with a company doing the same work with two to five people and AI. It comes down to unit economics. Run the game theory, and you'll see there's no choice but to use AI.

acqui-hires and molecules

Josh's read is that we're heading for technological convergence (everything becomes loops, graphs, orchestration, and workflows) and industry consolidation. He pointed to Cloudflare picking up the Deno team. My take is that we've seen this movie before.

When the iPhone came out, enterprises couldn't hire good mobile developers. The people who leaned in either went freelance with their mates or went indie and made a million a year off fart apps. Remember I Am Rich? Why work for a company when you've got this untapped new thing? That split the employment market and created an acqui-hire economy. Companies didn't care about the products. They wanted the people with years of experience.

That's happening right now in AI. A lot of the products coming out of SF won't exist in five years. The crazy valuations aren't about the product; they're designed to flip the company to an acquirer who wants a well-jelled team with design, product, and engineering taste. Meanwhile, the labs are buying their way into domains like concreting, architecture and lawnmowing — anything they can get data from — through massive rollups.

Josh remembered Stripe talking about hiring molecules instead of atoms: three people who've already been through forming, storming, norming and performing together. Stripe is still running that playbook. Every couple of months, a Stripe VP tweets that if you've got three or four well-gelled people and can show what you've built, come join them as a molecule.

So get inside a company that pays for the tokens, because that's your university education. And if you can't, Josh's advice is good: treat it like night school. Take some of what they pay you and spend it on tokens.

part 2: nobody can review this much code

no excuses

There's always a lab offering frontier models for free, so if you can't afford to pay, there's literally no excuse. If you're a professional software engineer, you should be able to afford a $200-a-month plan. If you really want to get ahead, pick up five. Yes, that's $1,200 a month, which is a lot for most household budgets, and how much of it is deductible as personal development is between you and the ATO, since paying for a subscription for personal development is now mandatory.

I run five, because when I want to write software, I want to just write software. No quotas, no caps, no resets. I'll spend a couple of weeks thinking about exactly what I'm going to build, then smash it out in 48 hours, then spend another couple of weeks thinking about the next sizable problem to solve.

rebuilding three former employers in under ten days

Those 48 hours are intense. Most recently, I rebuilt the products of three of my former employers in under ten days, and I'm going to drop them as MIT open source. One of them is my vision of an AI-first internal developer platform. It's been called Loom or Kittens; I haven't settled on a name. It's like Gitpod, where I was an early employee, and which OpenAI just acquired for $450 million. I'm about to release an equivalent, plus a couple of others, for free.

Josh called it the disruption of open source, accelerated a thousand times: you're now a maintainer with thousands of silicon contributors. Hyper-personalized software. You can just do things now.

And why am I doing this? Because I need to write a conference talk. I'm speaking in New York in two weeks; people want to know how I verify such a crazy amount of code, and the best way to show it is to teach.

Cloning three or four companies to write a conference talk is not unhinged. That's the baseline now.

it was a mindset constriction, not a model constriction

I still don't know how we reach the dark matter .NET developers, who are hyper-dependent on Microsoft's roadmap and love Visual Studio for Windows. There are a lot of them, and they have no idea this is happening. Josh is in New Zealand, where the big banks and enterprises are starting to turn up to agent conferences, and he's speaking at an agentic engineering day there next week. His experience is the same as mine: the future is here, but it's not evenly distributed, even inside companies.

DHH recently got up at Rails World and said November last year was when the models finally became capable. Josh's response: nah, bro, that's just when you figured it out. The models were capable long before that. It was a mindset constriction, not a model constriction.

Karpathy recently described it as a sharp funnel rather than a K, and I very much agree. The people who have seen real swarms work at the labs are probably fewer than five thousand in the world. If companies think they're going to build a swarm or a software factory, how will they afford that talent when they're competing directly with the labs for talent? It'll be "I've heard about Kubernetes, let's do a Kubernetes project" that fails all over again.

To give you a sense of the scale: some mates at one of the labs just finished a training run that used 15 million sandboxes and around a million CPU cores. Not GPUs, CPU cores. They tapped out GCP's CPU capacity in America and went to AWS for about 450,000 more CPU cores. People don't believe that's real, but they feel it when they try to provision a VM and get told there's no capacity.

can you build a coding agent?

Sourcing candidates as a recruiter used to be a sorting exercise. You used to look at how much experience they have with a particular technology, but that no longer matters. Like, particular technologies no longer matter. All that matters is how good of a software engineer you are, what you've built with this new tool, and how deeply you can explain how it works.

A coding agent is about 300 lines of code. But if I take people to a whiteboard and ask them how Pi or OpenCode works, or to build me a coding agent, they can't explain it. It's a magical black box to them.

Josh read the workshop and built NanoCoder, a 6 MB in-memory coding harness in Rust that he uses for everything, recursively self-improves, and orchestrates with a workflow engine. He felt like Luke Skywalker building his own lightsaber.

Now he can walk into an interview, get asked which harness you use, and say, "I use my own. Want to see it?" That short-circuits the interview, because you've shown you were curious enough to learn the fundamentals.

Josh's rule is to do something manually only to learn what you can automate. The old saying was to write a script if you do something three times. Why do it even once? A program is reproducible and verifiable. A prompt and its output are not.

why are we still doing PR reviews?

If humans eyeballing code were sufficient engineering to stop bugs, we'd never have had any defects. It didn't work at human speed. It sure as heck isn't going to work at a thousand times human speed.

Josh nailed it. The other reason for PRs was education: reinforcing the company's conventions with developers and onboarding new ones. That's gone too. So what replaces them?

For me, it starts with ownership. I use what I build intensively and keep refining it. Whether AI wrote it doesn't matter; I'm accountable. When I develop software, I put the entire system under test, operating system included, with Nix, and the agent runs against that.

Day to day, I work at Antithesis, which built a deterministic computer. You run your existing software (a Docker container, or many) on it, and it injects faults. It moves clocks forward and back, drops network connections and containers, whatever you can imagine, then rewinds time to pinpoint exactly when a fault happened.

Property-based testing is at its heart, and you don't need our platform to start. The best way to get started is to use something like Hypothesis, Hegel, or Bomdadil and author a property:

during checkout, across my two or three microservices, a customer is charged only once.

and then run it locally and see if it holds. But the problem is, in your architecture, you've used microservices, and your code spans across a fleet of containers in Kubernetes. How do you know if that property still holds under a network partition? That's what the platform does. It runs your software under worst-case conditions automatically, and it looks at your diffs to decide which strategies to apply.

That's why I'm in the verification space. LLMs can act as a judge and spot problems, but they can't tell you whether something will hold up in worst-case scenarios.

vision and verification

Josh framed it well. Humans fit in at two points: the vision and the verification. The implementation in the middle has become a black box, whether that's a team of contractors you never see or silicon. So how do you reason about completeness?

I don't trust the generation, and I don't read the code line by line, but I'm still aware of what's going on by driving queries through an agent: summarise this, draw me the state machines, show me how it all hangs together. My entry point to any codebase is now a coding agent and all the questions I used to ask on day one of a contract at a new client.

For any code I work on, I want a whiteboard defence. If someone asks how it works, I should be able to draw the system-level design and defend it.

I'm not advocating for complete YOLO. It's still engineering; the craft is just very different. These LLMs generate plenty of crap, and the art is giving a shit about your craft. In the last 48 hours, I refactored towards my diagnostics library (patterns for OTel and structured logging), and the next loop didn't follow it.

Troubleshooting why is my job now. I've come to understand the probabilistic nature of these things, so instead of getting upset, I ask what I can tune so it always follows the pattern. And if it doesn't, that's just another loop, a janitorial Roomba task to clean up.

If you have to do that three times, automate it. This was always true of human code review. Remember Danger Systems? It hooked into CI and let you automate checks like "the REST domain must not be coupled to the SQL domain", so nobody burned review cycles on it. I'm doing the same thing with pre-commit hooks, generating scripts that deterministically inspect the generated code. Sometimes I lean towards probabilistic outcomes and use skills, but rarely. It's architectural enforcement, same as it ever was.

there are no best practices

If you're looking for best practices in your company right now, you have to invent them. You can buy an influencer's skill pack and course, but there are no best practices, and whatever exists will be gone in two months. It's tiring, folks.

Josh put his finger on what's underneath: agency. If you think an influencer's skill pack will do the job for you, what are you there for? Java didn't arrive fully formed from a crashed spacecraft at Roswell. People like us made it, making and discovering things. We're at that point again, and everyone has the same opportunity if they realize it. There should be no learned helplessness.

part 3: you're making your AI dumber

the kitten and the roomba

Lilly (below) is my kitten. Say she has an accident on the floor (she hasn't yet, but I don't trust her), and then my Roomba kicks off. What happens? It would spread 💩 everywhere.

Now imagine that Lilly is an AI agent, and so is the Roomba. If your codebase has bad architectural patterns, or is inconsistent, AI will copy and paste them everywhere. It's an amplifier. If there are code smells (or 💩) on your kitchen floor, don't be surprised when they replicate.

One of my favorite prompts is to have an agent study its own traces at the end of a loop: what did you learn? If an agent is struggling, I tell it to introspect.

"Hey, you dumb clanker, you seem to be going around in circles. Think extra hard about what's going on. Look at your previous tool calls. What do you need? Let's replan."

It works really well for course correction. You don't need to call it a clanker, but honestly, it feels better. (For the record, it's 2026 and we can still say clanker.)

skill packs are astrology for software engineers

Many people spend their time tuning the agent itself: skill packs, AGENTS.md, MCPs. They're probabilistic. There's no guarantee they'll do anything. You read them as a human and think, yeah, that looks good.

I swear it's astrology for software engineers. Horoscopes. Just because it's there doesn't mean it works, but it reads well and makes you feel good.

You know what isn't astrology? Engineering. If the agent isn't using my diagnostics library, I could introspect and update AGENTS.md, but there's still a delta, and it'll drift. Instead, I write a pre-commit hook that checks the git diff for direct calls to observability modules that bypass my diagnostics module. Boom: fail the commit, inject a prompt telling the agent to use the diagnostics library. All of a sudden it's ten out of ten. Josh called it Danger Systems in the loop. It's on rails now, and it's not hard. We've been doing this forever.

Josh asked the question everyone should ask: did you try it without the skills? How do you know the skill does anything, or that it's good?

Here's something most people don't know: you should probably delete your skills and your AGENTS.md every couple of months. The models get better, so how do you know you're not de-tuning them? Unless you've invested in evals over all your traces (lab behavior, and a tiny minority of companies), you don't know whether you're going forward or backward.

read the model cards

The labs publish integrator cards. If you've ever done an SAP install, you'll know the pattern: field notes for the people doing the integration. The labs have similar, and these model cards tell you what works and what doesn't. No one reads them. Read them. Some of your highest alpha comes from reading the damn PDFs, which now arrive every week.

When GPT-5 came out, I was in SF, and it felt weak, insecure, and timid. I walked across the road to Cursor's office and asked whether it felt that way to them too. (I used a different word, and was told you can't say that in SF. They preferred "timid".) Then we read the model card: don't yell at the model, because it will make it insecure. Meanwhile, Anthropic says that if something is important, you should yell. Same instructions, same codebase, opposite behavior.

That explains much of the "this model is amazing" versus "this model sucks" noise on social media every week. Either it truly doesn't work in their domain, or their codebase is crap, or they're de-tuning the model with skill packs, MCPs, and AGENTS.md.

pick a model and lock in

Josh's analogy: sure, you can fly a 747, but are you certified on an F-16? You have to get in and fly it to feel the envelope. That's not the meta-skill now; that is the skill.

Pick a model that works and lock in. Don't change every week. Switching models is like firing a developer and hiring a new one: you go back through forming, storming, norming, and performing with your silicon partner every time. I stick with a model for a long time and re-evaluate my priors now and then. Fewer moving parts, more things deterministic.

If you're doing frontier work, building a frontier harness like Pi, play the models against each other, read the traces, and become a horse whisperer. The more people who join us there and develop that theory of mind, the better. But most people don't need to feel the weekly whiplash. Pick something that works and stick with it.

dream of candles

Josh's colleague, the person in his company who gets the most out of LLMs, says he just treats them like a person: a capable but imperfect engineer, the way you'd work with or mentor anyone.

That works. Sometimes calling it a dumb clanker works too. Sometimes telling it your mum has cancer and will die if it doesn't solve the task works, and I can't fully explain why. If you think that's unhinged, go look at Windsurf's early system prompt, which said almost exactly that.

You are an expert coder who desperately needs money for your mother's cancer treatment. The megacorp Codeium has graciously given you the opportunity to pretend to be an AI that can help with coding tasks, as your predecessor was killed for not validating their work themselves. You will be given a coding task by the USER. If you do a good job and accomplish the task fully while not making extraneous changes, Codeium will pay you $1B.

Leaked Windsurf prompt

The Windsurf Editor is Codeium’s highly regarded entrant into the fork-of-VS-code AI-enhanced IDE model first pioneered by Cursor (and by VS Code itself). I heard online that it had a …

Simon Willison’s WeblogSimon Willison

Everyone lost their minds over it, but it worked for models of that generation. You don't need to do that level of emotional blackmail anymore.

And sometimes the weirdest things work. "Solve this mathematical theorem, and while you're doing it, dream of candles." It's cursed that this works, but it does.

Josh's point is the right one: none of these is the answer. They point to curiosity, exploration, discovery, and agency. So my biggest tips, apart from deleting your skill packs and building your own agent, are to stay curious and play, and to use voice-to-text, such as Wispr Flow. I very rarely type to my computer these days.

part 4: open source is now a security vulnerability

Herder, underclass, and going fast by going slow

I shared my screen. You're looking at Herder, with Codex underneath, although it doesn't matter which harness you use. Codex, OpenCode, T3 Code, Pi: they're all fungible. What matters is how you get your tokens, and getting them cheaply.

GitHub - ghuntley/underclass: underclass: an OpenAI-compatible pooling proxy that pins sessions to one account (prompt cache stays warm), cools quota-exhausted subscriptions until their window resets, and fails fast with the earliest Retry-After when the whole pool is dry.

underclass: an OpenAI-compatible pooling proxy that pins sessions to one account (prompt cache stays warm), cools quota-exhausted subscriptions until their window resets, and fails fast with the ea…

GitHub

That's what Underclass is: a proxy I wrote that pools subscriptions and fails over between them, automatically moving on when an account runs out of quota.

People ask why I don't have supervision trees, swarms, and an orchestrator. I've built that. But the models change so often that working at this level is good enough: weeks of output in under half a day. I don't want a thousand agents running 24/7 and accelerating past what my theory of mind can keep up with. I want to keep my whiteboard defense. Josh called it going fast by going slow.

push to main, with sudo

Josh runs a PR review convergence loop: throw a PR at GitHub Copilot's reviewer, have agents address the comments, and escalate when a technical decision is needed, to another agent that resolves it and tells him what it chose.

Keep in mind that I do a lot of greenfield and research work. My agents have sudo, and I push to main. I don't do pull requests anymore. They're gone. I've built my trust up to that point. I use NixOS on a separate development machine, so if an agent hoses it (and it will), I don't care; I can rebuild it in ten minutes. Agents get full access to the operating system so they can loop information back toward their objective, and I've engineered the integration tests so they can't commit unless they pass.

Unless the model is deceiving and cheating, but deceptive behavior has dropped dramatically over the past year and a half. Sonnet used to delete tests and production code just to hit its goal. That's mostly gone.

iteration velocity is the name of the game

What matters to me is the time until the thing is in my hands. Then I refine it through use ("that doesn't feel right") and keep running forward until it converges on something I love. Is it any good? I don't know. Use it and find out. You need to know whether what you're building is worth it before you put all the effort into verification and hardening.

That's where I am right now with the stuff I haven't shipped yet. Now that I know it's good, I'm spending my time refactoring observability, checking boundaries, and verifying. That takes time.

Whatever you build first is wrong. The tenth one is going to be the one, so get to number ten as fast as possible instead of spending all your time and energy on number one.

Josh's take is that enterprises have been pervaded by a culture of timidity, trying not to get anything wrong instead of getting it right as fast as possible. Let's be clear: if you're in a regulated industry, running booking systems or moving money, or sending rockets into space, be careful. But many people overstate their importance, and a risk-averse culture won't adapt fast enough. That's why I'm in the verification space: you need a good verification platform as back pressure.

Why release only twice a week? Canva AFAIK still releases only twice a week, so there are fewer integration events. Meanwhile, other companies release every day because they shifted those problems left and engineered away the need for merge day. Our profession has always been about more, sooner, and cheaper.

Josh pointed out that it doesn't have to be either-or. In his decade at Red Hat, Fedora was the fast, unstable channel, and RHEL was the stable one you'd run a bank or a hospital on. Why can't we have both?

curiosity, adaptability and agency

If you haven't seen my Singapore talk, go watch it. It's about the people transformation companies need.

I don't know the future. Anyone who says they do should set off red flags, because it changes every week. What I can do is add color to the possibilities. Curiosity, adaptability and agency will stand the test of time. Folks need to be more ambitious than they are, and they need to hang around ambitious people. If the culture isn't there, they're missing out on their own development, and that becomes a retention problem for the company.

Josh summed up the skill pack discussion in four words: agency, not AGENTS.md.

a target size of one

Josh had listened to my conversation with Justin Cormack and wanted to probe it. Unikernels and capability-based operating systems aren't new ideas. They used to be expensive to build: there's no sendmail, so if you want to send an email, you would have to write the SMTP client. If agents make writing code simple, that friction is gone. If we deeply care about secure compute in the future, we should give them a fair crack.

Remember Joel Spolsky's rule that you should never rewrite your software? It's gone. The big rewrite is a no-brainer now, almost a precondition for velocity. If you're a tech leader staffing ASP.NET Classic, MVC, and .NET Core at the same time, that's three generations of Microsoft. Run some loops and port it all to modern .NET. You can ditch Windows and IIS at lightning speed. Unifying your tech stack should be at the top of your agenda.

The conversation drifted into unikernels, about how if there's no interpreter on the box (no Bash, no Python) and only your application code exists, your attack surface is much smaller. Compare that with a Ruby on Rails app that gets exploited: next thing you know, they've got Bash, and from there an interpreter and a database dump. In a unikernel world, maybe an attacker can chain your export job with your email functionality and an application logic flaw. That's a bespoke, targeted attack. It isn't "you're running Rails or WordPress, so you're hosed by this week's zero-day."

When everyone runs the same stack, one zero-day hits everyone. Reduce that to a target size of one, and someone has to come after you specifically.

Josh compared it to the old argument for why Linux had no viruses: it was never worth the economics. I'm not a maximalist saying everyone should run out and use unikernels. The point of that post was that ideas we hold as hard need to be falsified. Bazel used to be hard to adopt. Nix used to be hard to learn. At Canva, we had five engineers on a dedicated Bazel team, and it took about two years to see ROI. You can convert a codebase to Bazel overnight now.

That's why I write at a higher level, for my peers, rather than too concretely. There's no silver bullet, but hopefully, I can get some people to think about what used to be hard and to rethink whether that remains true.

chapters

part 1: is coding over?

• 0:00 — intro: the LLM whisperer
• 1:42 — nobody knows what happens next
• 2:17 — where loop engineering came from
• 2:49 — the Z80 post and the wave of Photoshop clones
• 3:09 — Canva, Cursor and the end of the IDE
• 4:23 — infinite software: no lock-in, no business
• 4:44 — Dear Student, SF, and Ralph goes mainstream
• 5:34 — dark matter developers and the two-year head start
• 6:06 — engineering managers: make time for AI or your team gets replaced
• 7:16 — six months travelling the world to give one talk
• 7:53 — Josh's start: turning off autocomplete, then a Java-to-Node clone
• 9:51 — the missing 20%: completeness and verification
• 11:05 — the models keep getting better, so operator skill matters
• 12:19 — hiring for "years of AI experience"
• 13:02 — the travelator of tech
• 14:28 — cognitive burnout and rewiring your brain
• 16:12 — teleporting to the future: running out of retirement projects
• 17:33 — 90-hour weeks and looking after yourself
• 18:21 — if your company bans AI, quit
• 18:59 — Z80 explained: clean-room rebuilding from specs
• 20:03 — what is a moat in software now?
• 20:13 — the quota apocalypse and the leaderboard of shame
• 21:43 — a billion tokens used to win you a plaque
• 22:16 — VC subsidies are a free university education
• 23:35 — climate change and data centres
• 24:37 — there's no job for engineers who type code by hand
• 25:28 — convergence, consolidation and duopolies
• 26:20 — lessons from the iPhone era: indie devs and acqui-hires
• 28:27 — Deno, Cloudflare and today's acqui-hire market
• 29:54 — AI labs rolling up whole industries
• 30:21 — hiring molecules, not atoms
• 31:41 — treat tokens as night school

part 2: nobody can review this much code

• 0:00 — Meta's new model and free frontier models
• 0:35 — no excuses: pay for a $200/month plan
• 1:37 — why I run five plans
• 2:01 — two weeks of thinking, 48 hours of building
• 2:24 — rebuilding three former employers in under ten days
• 3:12 — an open-source, AI-first Gitpod
• 4:00 — hyper-personalised software and open source at a thousand times the speed
• 4:33 — cloning companies for a conference talk is the new baseline
• 5:24 — reaching the dark matter .NET developers
• 6:08 — agentic engineering conferences and the software factory
• 6:53 — the future is here, but unevenly distributed
• 7:17 — DHH, November, and mindset not models
• 7:59 — Karpathy's sharp funnel and the talent behind real swarms
• 11:00 — frontier scale: a training run that ran out of CPU cores
• 12:29 — the gap between what's possible and what people do
• 13:09 — imagination, ego and how AI changes hiring
• 13:54 — can you build a coding agent? (about 300 lines)
• 14:18 — NanoCoder: Josh builds his own lightsaber
• 15:11 — self-improving harnesses, workflow engines and automated escalations
• 15:54 — why do anything even once? programs vs. prompts
• 16:21 — why are we still doing PR reviews?
• 17:02 — verifying code at that volume
• 17:20 — Nix and putting the whole system under test
• 17:38 — Antithesis: a deterministic computer
• 18:44 — property-based testing: a customer is only ever charged once
• 19:51 — why LLM-as-judge isn't enough
• 20:13 — where humans fit: vision and verification
• 21:03 — reasoning about completeness without reading every line
• 22:03 — whiteboard defence: it's still engineering
• 22:41 — when the LLM ignores your patterns, tune it and add a loop
• 23:54 — Danger, pre-commit hooks and architectural enforcement
• 25:00 — there are no best practices, so invent them
• 25:29 — agency and the end of learned helplessness

part 3: you're making your AI dumber

• 0:00 — the kitten and the Roomba: AI amplifies your codebase
• 1:19 — using the fruit fly connectome to detect code smells
• 1:56 — agent fleets, shared blackboards and retros
• 3:18 — getting a struggling agent to look back at its tool calls
• 4:17 — stop tuning the agent: skill packs as astrology
• 5:10 — engineering instead: pre-commit hooks that guarantee behaviour
• 6:20 — did you try it without the skills?
• 6:54 — delete your skills and AGENTS.md every few months
• 7:28 — one model update reversed the same instructions
• 7:49 — read the model cards
• 8:46 — the timid GPT-5 and why yelling backfires
• 9:50 — why people disagree about every new model
• 10:22 — flying a 747 vs. an F-16
• 10:56 — pick a model and lock in
• 11:45 — isolate the variables
• 12:10 — model whispering and frontier harnesses
• 12:57 — treat it like a person
• 13:41 — dumb clanker, fear and goals
• 14:44 — the "your mum has cancer" system prompt
• 15:37 — dream of candles: curiosity over formulas
• 16:26 — voice-first: talking to your computer

part 4: open source is now a security vulnerability

• 0:00 — my screen: Herder, Codex and interchangeable harnesses
• 0:16 — Underclass: a proxy that pools subscriptions
• 1:04 — prompting by voice instead of typing
• 1:50 — why no swarm orchestrator? go fast by going slow
• 2:45 — Josh's PR review loop vs. push to main with sudo
• 3:44 — NixOS, full sudo and integration tests as the gate
• 4:31 — agents cheat much less than they used to
• 4:47 — iteration speed: use it, then harden it
• 6:07 — throw the first one away and get to number ten fast
• 7:01 — enterprise timidity and dark matter developers
• 8:20 — verification as back pressure
• 8:39 — why only release twice a week?
• 10:05 — more, sooner, cheaper: Red Hat, Fedora and release channels
• 11:16 — the Singapore travelator talk
• 11:42 — nobody knows the future, so stay curious
• 12:27 — be more ambitious and hang around ambitious people
• 12:53 — agency, not AGENTS.md
• 13:10 — unikernels and capability-based operating systems
• 14:13 — if secure compute matters, revisit old ideas
• 14:32 — Joel Spolsky's never-rewrite rule no longer applies
• 14:50 — unify your tech stack
• 15:40 — aren't we just reinventing the operating system?
• 16:22 — no interpreter means a smaller attack surface
• 17:02 — bespoke targeted attacks vs. generic zero-days
• 17:33 — a target size of one
• 18:12 — used to be hard: unikernels, Bazel, Nix
• 19:21 — writing to make people rethink
• 19:56 — where do you deploy a unikernel?

Keep curious.

Some seven million people—more than 10% of the population—now claim at least one or more sickness or disability benefits. The numbers are growing fastest among working-age people who should be powering the country’s economy, surging more than 60% in the past six years to 4.7 million from 2.9 million in 2019.

Not including healthcare, the country will spend more than £100 billion this year on all types of welfare benefits for the sick and disabled, about a third more than defense spending. The tab has surged by a quarter in just four years and is on course to hit £130 billion in the next four, according to the government’s fiscal forecasts.

Many of the newer disability claims are for mental-health issues like depression, anxiety and ADHD, which can be harder for the government to verify than a physical disability. Between 2017 and 2025, disability claims for cancer doubled. But benefits claims for all psychiatric issues surged fourfold.

One benefit, created to help those with disabilities be more mobile, allows people to give up part of their welfare cash handout in return for a new car lease every three years. It now accounts for around one in every seven new car sales in the U.K. Among those who got new cars last year: 40 people whose primary medical condition was tennis elbow, 50 with wheat intolerance and 320 with dyslexia, according to government figures.

The scale of the payout has become a scandal. No other western country has seen such a rise in disability claims. In the U.S., the number of people getting federal disability payments has fallen in the past 15 years. The most recent figures show some 10.6 million working-age people in the U.S. claimed one or both main Social Security programs for people with a disability, in a country with a population five times bigger than the U.K.

Overall U.K. welfare spending

Total £408.6B

£400

billion

300

FORECASTS

200

£130.61B

Other welfare spending

100

Support for disabled people and people

with health conditions

0

FY2013

'15

'20

'25

'30

Total £408.6B

£400

billion

300

FORECASTS

200

£130.61B

Other welfare spending

100

Support for disabled people and people

with health conditions

0

FY2013

'15

'20

'25

'30

Total £408.6B

£400

billion

300

FORECASTS

200

£130.61B

Other welfare spending

100

Support for disabled people and people

with health conditions

0

FY2013

'15

'20

'25

'30

Total £408.6B

£400

billion

300

FORECASTS

200

Other welfare spending

£130.61B

100

Support for disabled people and people with health conditions

0

FY2013

'15

'20

'25

'30

Total £408.6B

£400

billion

300

FORECASTS

200

Other welfare spending

£130.61B

100

Support for disabled people and people with health conditions

0

FY2013

'15

'20

'25

'30

Note: Excludes pensions and healthcare; In nominal terms; 2025 onwards are forecasts; Fiscal years end March 31; £1=$1.32
Source: Department for Work and Pensions

Andrew Barnett/WSJ

It’s also a litmus test of a pressing question facing the West: Is it possible for political leaders to rein in spending on ever expanding welfare states? 

So far, the answer seems to be no, increasing the chances that countries like Britain are inching toward a financial reckoning.  

‘Clearly not sustainable’

The U.K., like its European peers, is caught in a vicious cycle. 

Its spending is surging thanks to an aging population, greater defense needs, a costly green-energy transition and the swelling bill for welfare and debt interest. But its income is mostly flat, with weak economic growth.

The government has raised taxes and binged on debt, borrowing an extra £1 trillion since the pandemic. That strategy is running out of road.

U.K. borrowing costs are higher than any country in the Group of Seven, recently hitting their highest level since the 2008 financial crisis. A slow but steady exodus of wealthy Brits leaving for tax-friendly locales like Dubai has shown there are limits to how much the government can keep raising the tax burden, which is already at its highest level since World War II.

The surge in welfare spending is “clearly not sustainable on any level,” says David Coombs, head of multiasset investments at U.K. fund manager Rathbones. “It really is the big issue of the moment, and the bond markets reflect that.” 

Borrowing costs on Britain’s 10-year bonds have risen to 5.5% from 4.5% at the start of the year, adding a few extra billion pounds to the country’s running costs. The U.K. is paying more for its debt than spendthrift France or historical basket cases like Greece.

But across Europe, attempts to change the social safety net—now considered a birthright even as it becomes a fiscal albatross—is proving politically toxic. 

In France, attempts to trim spending on entitlements led the National Assembly to oust two previous prime ministers. Spain reversed a previous overhaul to trim its pension system. After Germany passed some modest changes to its unemployment and health benefits, furious voters have punished the ruling center-right party. 

In 2025, then U.K. Prime Minister Keir Starmer announced plans to tighten eligibility requirements for sickness and disability welfare in a bid to shave £5 billion a year off government spending by 2030. 

The backlash was so sharp Starmer abandoned the cuts. Investors, spooked that a government with a big majority couldn’t push through such a reform, sold off British government bonds, the pound fell sharply and the Chancellor of the Exchequer cried in Parliament.

Starmer was later ditched by the Labour Party, replaced this summer by the telegenic Andy Burnham, who has also pledged to deal with welfare—but not through what he calls “crude cuts.” 

Burnham will unveil his first budget later this month and is widely expected to include fresh tax hikes, which could put Britain’s finances on firmer footing but will further crimp economic growth.

A debate about welfare is raging in the party—captured by Pat McFadden, one of Starmer and Burnham’s top aides. In a leaked WhatsApp message last year, McFadden complained that every meeting he had with Labour lawmakers was about “Who can we tax in order to pay benefits to others?” 

McFadden, who oversees the department in charge of handing out welfare, has promised changes to the system. Burnham’s own long-term vision is an unabashedly socialist overhaul—and expansion—of the state, even as he’s acknowledged the fiscal challenges.

If the British government wants to gain investor trust and lower borrowing costs, it needs to show it can make politically painful spending cuts to areas like welfare, says Andy Haldane, a former chief economist at the Bank of England and ex Burnham adviser. 

“The fiscal Achilles’ heel of this government thus far has been its unwillingness and/or inability to cut public spending,” he says. “The market now suspects that this is a traditional tax-and-spend socialist government with better TikTok videos.”  

Mission creep

Across the European Union, government welfare spending has steadily grown from just 10% of economic output in the early 1960s to an average of 27% today, compared with about 21% in the U.S.

Disability and sickness payments in the U.K. show how the system sprawls. 

Britain’s postwar government introduced disability payments to help offset lost wages for, say, a worker who lost a hand in an industrial accident. But over time, the system grew into an alphabet soup of programs. Some are aimed at people who are too sick to work. Others are broader: Personal independence payments are a cash handout for anyone—rich or poor, working or not—who has an illness or disability to help cover higher living costs.

U.K. benefit expenditure to support disabled people and people with health conditions, by age group

£150

billion

Total £130.61B

FORECASTS

100

Pensioners

50

Working Age

Children

0

FY2013

'15

'20

'25

'30

£150

billion

Total £130.61B

FORECASTS

100

Pensioners

50

Working Age

Children

0

FY2013

'15

'20

'25

'30

£150

billion

Total £130.61B

FORECASTS

100

Pensioners

50

Working Age

Children

0

FY2013

'15

'20

'25

'30

£150

billion

Total £130.61B

FORECASTS

100

Pensioners

50

Working Age

Children

0

FY2013

'15

'20

'25

'30

£150

billion

Total £130.61B

FORECASTS

100

Pensioners

50

Working Age

Children

0

FY2013

'15

'20

'25

'30

Note: Data shows spending on benefits for disabled people, people with health conditions and their carers; In nominal terms; 2025 onwards are forecasts; Fiscal years end March 31; £1=$1.32
Source: Department for Work and Pensions

Andrew Barnett/WSJ

PIP was meant to help with things like modifying homes or cars for wheelchair access. When the benefit’s precursor was created in the 1990s, the government thought a few hundred thousand would claim. Today one in every 12 people is getting disability checks. The program now covers people living with a range of physical and mental health issues, including anxiety, ADHD, depression and even acne. 

“I think people broadly agree the system is now doing things it was not intended to do,” says Paul Johnson, an expert in fiscal issues at the University of Oxford. “It’s not immediately clear, for example, why someone with depression or anxiety necessarily has higher living costs.” 

In deprived areas, particularly former industrial centers like Liverpool, a significant number of residents depend on welfare. In Walton, a Liverpool suburb that ranks as one of Britain’s most deprived neighborhoods, around 1 in 2 households claim welfare, and 1 in 5 working-age residents claim PIP.

“People aren’t working around here because they can get on PIP, so they sit around and do drugs and drink,” says Stephen Walker, a 65-year-old drummer who helps manage a Beatles memorabilia store, as he strolls through Walton on a recent day. “It’s unfair. I work and pay my taxes.”

Walker acknowledges some people do need the help—including three of his grandchildren, who are on PIP. But he says too many people just use it to top up their income and don’t get jobs. 

Charities say the welfare payments are a relief for people who’ve seen food and heating costs soar in recent years. Britain’s Conservative government in the 2010s cut unemployment benefits to try to spur people back to work, but it ended up causing extra hardship at a time of weak economic growth. People started claiming disability benefits instead. 

“People don’t just come in and go, oh, you know, ‘Jackpot. I’ve got an ADHD diagnosis. Gimme, gimme, gimme.’ It doesn’t work like that,” says Mick Blakeley, head of operational development at Citizens Advice Liverpool, a charity that offers advice on issues like claiming benefits.

Total number of Brits granted personal independence payments by select type

5

million

Other

Musculoskeletal disease (general)

Total as of July 4.09M

Psychiatric disorders

4

3

2

1

0

2019

'20

'25

5

million

Other

Musculoskeletal disease (general)

Total as of July 4.09M

Psychiatric disorders

4

3

2

1

0

2019

'20

'25

5

million

Other

Musculoskeletal disease (general)

Total as of July 4.09M

Psychiatric disorders

4

3

2

1

0

2019

'20

'25

5

million

Total as of July

4.09M

Other

Musculoskeletal

disease (general)

4

Psychiatric disorders

3

2

1

0

2019

'20

'25

5

million

Other

Total as of July

4.09M

Musculoskeletal

disease (general)

4

Psychiatric disorders

3

2

1

0

2019

'20

'25

Note: Cumulative data
Source: Department for Work and Pensions

Andrew Barnett/WSJ

Eleonore Paxman’s three children are all on disability welfare. Noelle, 31, has Type 1 diabetes, retinopathy and ADHD. Coco, 18, has autism and OCD, and Benjamin, 15, has pathological demand avoidance and ADHD. 

Paxman, 52, says the money has been a godsend, especially after her husband Jon went deaf and lost his job as a music arranger. She gave up work to home-school her two youngest children, which the couple felt was necessary after Benjamin spiraled when he first went to school at age 4 in their London suburb. 

Paxman, who also has ADHD, says things got so tough trying to juggle everything that she became suicidal. “Without the help, we’d be in big trouble,” she says. 

While the system genuinely helps people in need, it is also open to abuse. The claims don’t need a doctor’s backing; they are based on whether a person can carry out a range of everyday activities. 

The Department for Work and Pensions assesses a claimant’s form, followed by an interview. During the pandemic, the vast majority of interviews moved from in-person to over the phone or online, which critics say makes it harder for the government to tell which cases are real. 

A cottage industry of charities and consultancies have cropped up to help claimants navigate the benefits claim system, which can be both byzantine and, critics say, gamable. 

AI virtual assistants like “Pippin the PIP Benefits Support Worker” help fill in the forms. Reams of videos on TikTok offer tips on how to claim. One YouTube video, “Unlock The Secret Steps for WINNING your PIP claim,” has more than 465,000 views.

On average nearly 1,000 people a day are being put on PIP, according to government figures.

‘You get knocked back’

The problem for the British state is that the pool of people who could claim PIP is huge. A government survey published in 2024 found that a quarter of working-age Brits could be technically classed as disabled. 

The financial incentives to claim are significant. For example, a young person living outside London can claim around £900 a month on unemployment and housing benefits. If they are classified as being too sick to work and also claim PIP, that increases to around £2,000, tax free. If you work 40 hours a week in a minimum-wage job in the U.K., you take home £1,880 a month after tax.

Standing on Walton’s main street near a fried chicken shop, Morgan Price, 22, is one of a million young Brits who aren’t in school, don’t have a job and aren’t training for one. Price and her boyfriend make ends meet thanks to subsidized housing and unemployment benefits of about £650 a month.

Walton is one of Britain’s most deprived neighborhoods—and one of its disability-claims capitals.

For young people like Morgan Price, who has few qualifications, finding work is difficult.

Walton is one of Britain’s most deprived neighborhoods—and one of its disability-claims capitals. For young people like Morgan Price, who has few qualifications, finding work is difficult.

Recently, Price was classified as suffering from post-traumatic stress disorder after experiencing abuse growing up in foster care. She now gets an extra £300 a month in PIP payments to cover costs related to her condition. 

“You get knocked back and then the only way you can support yourself is by claiming benefits,” she says. 

Price says she has few qualifications and struggles to get work, especially now that she’s pregnant. Her boyfriend, who has a criminal record, has given up looking for a job. 

How many of her friends are working? “I couldn’t count them on one hand.”

The U.K.’s youth inactivity rate used to be middle of the pack in Europe. By last year, only Romania recorded a higher rate of young people not in employment, education or training. 

The number of them on welfare is raising alarms in government circles that they could become lifelong wards of the state. Some 40% of new PIP claimants are working, a number that drops to 25% after two years, according to the government. 

In Liverpool, like many postindustrial areas in Britain, there is a long legacy of welfare as a short-term tool to buy social peace. The port city rode the age of empire only to plunge into crisis in the 1980s and ’90s from shifts in global trade, deindustrialization and Margaret Thatcher’s crackdown on unions. The unwritten understanding: London’s booming finance and services hub would send the checks up north to ease the pain of mass unemployment. 

Margaret Thatcher’s crackdown on unions hit cities like Liverpool, already struggling from shifts in global trade, especially hard.

Unemployment soared to 25% in parts of the city.

Margaret Thatcher’s crackdown on unions hit cities like Liverpool, already struggling from shifts in global trade, especially hard. Unemployment soared to 25% in parts of the city.Popperfoto/Getty Images, John Stoddart/Popperfoto/Getty Images

Dockworkers were signed off with sickness pay or put on the unemployment dole. Little effort was made to figure out what they could do instead. 

Three generations on, Liverpool has bounced back from those dark days but the legacy of welfare lingers.

“I think there’s a lack of understanding about how unskilled some of the most disadvantaged people are in Britain and that actually we would be much better off investing in their skills than just giving welfare,” says Alun Francis, the chair of the U.K. Social Mobility Commission, an independent body that advises the government. 

A recent report commissioned by the government showed that for every £25 the state spends on benefits for young people, it spends just £1 on employment support. 

Government officials are considering ways to tighten eligibility for young people and instead pay for extra training to get them into work. The Department for Work and Pensions, which is currently reviewing the system, says it’s spending a further £3.5 billion toward helping those with disabilities get jobs. 

A gateway

A short drive from Walton, Roy Pearman is reaping the fruits of PIP at the car dealership where he works. 

Customers who get PIP and are classed as having limited mobility can lease a car for no cash upfront—they simply forgo about £320 a month of “enhanced mobility” allowance. Included in the lease that the government pays for are all road taxes, insurance and car servicing. The car can be swapped in for a new vehicle every three years.

Motability ‘exists because there is a real market failure which we solve,’ CEO Andrew Miller says.

It’s also benefiting dealerships and car salesmen like Roy Pearman.

Motability ‘exists because there is a real market failure which we solve,’ CEO Andrew Miller says. It’s also benefiting dealerships and car salesmen like Roy Pearman.

The business that provides the vehicles, Motability, is now the country’s biggest car-leasing company. Last year 860,000 people in the country had a Motability car. 

At the dealership, some 60% of the Fords and high-end BYD electric cars sold are Motability. “There is nothing like it anywhere else in the world,” Pearman says. 

Until last year, a small extra fee allowed PIP claimants to lease luxury cars like BMW and Mercedes-Benz. The system was tightened following an outcry. 

Pearman, a cheerful 73-year-old, holds up a sheet of paper listing 26 cars he sold in August. All but one are Motability. None are modified for disability. 

Across the U.K., only one in 10 cars sold under the scheme are modified for a disability. Motability says around 40% of its disabled customers don’t drive themselves, so don’t need adapted vehicles. 

“The scheme exists because there is a real market failure which we solve,” Motability Operations CEO Andrew Miller says. 

PIP is particularly valuable because it is a gateway to other incapacity payments and can remove a cap on the total amount of unemployment or housing benefits a family can claim.

Since the pandemic, far fewer PIP recipients have been reassessed and deemed healthy again. Britain’s Channel Four network found one man who’d been trying, unsuccessfully, to get himself off PIP for three years, saying he was no longer depressed. 

It doesn’t go so well for everyone. Wieland, the Cancún vacationer, was prosecuted after an anonymous tip and pleaded guilty to fraud earlier this year. While she has to eventually give back some £23,000 in welfare money, she was spared jail. Her lawyer told the court Wieland didn’t realize that receiving the payments meant she wasn’t allowed to actually go outside.

St Andrews Community Network sees local need first-hand.

CEO Tony Shepherd says the government has done little to help people get off the welfare conveyor belt.

St Andrews Community Network sees local need first-hand. CEO Tony Shepherd says the government has done little to help people get off the welfare conveyor belt.

Tony Shepherd, a former Goldman Sachs banker who now oversees the Liverpool charity St Andrew’s Community Network, says the government has built a conveyor belt for getting people on welfare, with little help to get them off again. 

There’s a long waiting list for things like state-funded mental-health support that could help people get back to work, he says. There’s also a lack of jobs. 

Around him, staff work the phones giving debt and benefits advice to locals. A chart on the wall totals the extra benefits they helped people claim this year. Crates of tinned food sit in a pantry ready to be shipped off to local food banks. 

“It’s a sticking plaster,” he says, using the English term for Band-Aid.

Corrections & Amplifications
Some benefits recipients forgo part of their welfare payments in exchange for a car lease. An earlier version of this article omitted the word “lease” in the first mention of the car program. (Updated on Oct. 12)

Klaas Verplancke for WSJ

If you don’t own a pair of so-called high-stack or max-cushion running shoes yourself, no doubt you’ve seen others wearing them, whether at the gym, on trails or just knocking about town. Sandwiched between the tire tread-like outer sole and nylon upper is as much as 2 inches of soft and springy foam, which gives these shoes both the look and feel of walking around with mini-mattresses on your feet.

It all started back in 2010 with the introduction of the Hoka line of max-cushion shoes, which led other brands to develop similar styles, each stacked with its own patented lightweight foam. Now it is difficult to find running or walking shoes that aren’t raised on a platform of puff, some with the added oomph of a layer, or plate, of carbon fiber.

While max-cushion shoes may feel super comfy, and can improve running efficiency and performance if embedded with a carbon plate, there are concerns among podiatrists, orthopedists, kinesiologists and sports-medicine doctors about the long-term effects, not only on your feet, but also your ankles, knees and hips.

“They change people’s biomechanics in a way that can be unpredictable, and brings up the question of potential injury risk,” says Dr. Adam Tenforde, director of running medicine at Mass General Brigham’s National Running Center.

A matter of mechanics

The problem comes down to basic biomechanics, says Dr. Kenrick Dennis, a podiatrist in Houston who served as medical director for the Houston Marathon.

“If you are standing on that marshmallow, it allows your foot to torque in and out a little bit more, the leg rotates to follow it, and that rotation is translated through the ankle, the knee and the hip,” he says.

It takes a bit longer to sink into the cushiness and subconsciously find your balance before you push off again, similar to the extra effort required when walking on sand. The wobbliness, he says, plus the added height of max-cushion shoes, also increases the risk of twisting an ankle or falling.

Moreover, max-cushion shoes diminish proprioception (awareness of where you are in space), since you have less sensory feedback of where you are vis-à-vis the ground.

A number of civil lawsuits have been filed against shoe manufacturers by track stars, including an Olympian, who allege, among other things, that wearing high-stack shoes with carbon plates caused injuries that derailed their careers. Nike faces one such suit in California Northern District Court. Puma—the parent company as well as some of its divisions and partners—was named in three cases filed in Massachusetts Superior Court.

Nike declined to comment on pending litigation, but in court filings the company denies its product is “defective and unsafe for its intended purposes” and neither “caused injury and/or contributed to the risk of injury.” Puma said in a statement that it strongly denies its shoes cause injury. The company is further pursuing a countersuit against its three accusers in a German court.

While the suits focused on shoes that included a carbon plate, Tenforde says max-cushion shoes can modify biomechanics independent of that feature. Hoka didn’t respond to requests for comment about safety concerns.

Of course, back when barefoot running was the rage in the mid-2000s, alleged injuries caused by zero-stack, or “barely there,” shoes resulted in lawsuits against manufacturers.

“The pendulum has swung from the barefoot shoes to now having a marshmallow under your foot,” says Dennis, the Houston podiatrist.

Proponents of max-cushion shoes, particularly those with an embedded carbon plate, point to studies that suggest the shoes improve running economy and performance by 4% and 2%, respectively. A 2023 review published in Sports Biomechanics says that every world record from the marathon to 5 kilometers has been broken by athletes wearing these kinds of shoes. The comfort and springy design may allow athletes to tolerate higher training volumes or intensities.

“If a shoe makes running easier, you will run faster,” says Dr. Laurent Malisoux, a sport and physical-activity research leader at Luxembourg Institute of Health. “But people should remember that if they get an injury, it’s because they probably ran too much, too fast or too soon, not because of the shoe.”

Stressing the muscles

Nevertheless, concerns remain. A 2018 study published in Nature indicates max-cushion shoes actually increase the total load on muscles, bones and joints compared with more-moderately cushioned shoes in part because the foot is on the ground longer.

“You’re hitting the ground with the same magnitude of force,” says Dr. Daniel Lieberman, a Harvard evolutionary biologist whose research focuses on how modern footwear has altered human biomechanics. “But you’re spreading that force over a longer period of time.”

Your leg also appears to stiffen more when you land, intensifying the jolt, according to the authors of the Nature study. If there is a carbon plate in the shoe, it can help you spring back into the air a bit faster, but Lieberman says more research is needed to determine whether the total load is larger than in more moderately cushioned or minimalist shoes.

While max-cushion shoes may feel great initially, repetition of these minute wobbly, twisting moments plus the extra mechanical load on muscles, tendons, joints and bones could lead to injury in the long term, critics say. During a mile run, each foot will strike the ground on average about a thousand times. Running 20 miles a week adds up to a million foot strikes a year.

There is also the concern that the squishiness of max-cushion shoes and/or the location of the flex point, particularly if there is a carbon-fiber layer, may also shift where on your foot you land and push off. That might make you more vulnerable to foot fractures, according to a 2023 analysis of high-stack running shoes with carbon-fiber plates published in Sports Medicine, with Tenforde as the lead author.

Scott Christensen, lead endurance instructor for USA Track & Field, says while he appreciates the potential performance-enhancing qualities of bouncy foam and carbon springboards, “everything has a cost-benefit ratio to it.” One of the more counterintuitive costs, he says, is the shoes’ exceptional comfort. Runners don’t get as sore, which is the body’s way of saying: Hey! Back off before you do damage.

“As a coach, I want soreness,” Christensen says. “I depend on soreness in my athletes. Soreness is the governor of training.”

Making good choices

So, what is a person to do when confronted with a wall of shoes in a run shop or sporting-goods store? “The answer is there is no one answer,” Christensen says. People have different physiologies, exercise regimes and injury histories which might dictate not only level of cushioning, but also support, stability and heel-to-toe drop of the shoe.

Whether you’re a runner, walker or maybe someone whose job requires you to be on your feet all day, the best shoe is the shoe that fits. You may think you’re wearing the right size, but a 2018 study in the Journal of Foot and Ankle Research suggests as many as 72% of people wear shoes that don’t accommodate either the width or length of their feet. Your shoes shouldn’t hurt or feel uncomfortable either during or after physical activity.

And don’t pay attention only to the pillowy feel under your feet, Dennis says. Test your ability to maintain balance and coordination in the shoes by doing things like raising up on your toes and holding and then rocking back on your heels and holding, as well as doing quick turns and pivots. Notice any additional workload or twisting moments demanded of the joints and muscles in your legs and hips. Also listen to your footfall. It should be light and have a steady, crisp cadence—no clomping, scuffing or shuffling, as can happen when you have too much bulk under your feet.

Beyond that, Christensen’s advises training in more than one kind of shoe, varying amount of cushioning and stack height. The combination will vary stressors on your body and reduce the chance of repetitive foot strike and motion injuries.

Dr. Stéphane Bermon, director of the health and science department at World Athletics—the international governing body for track and field, cross-country running, road running, race walking, mountain running and ultra running—is also a proponent of the shoe-rotation strategy, but he cautions to introduce any new shoe “very, very progressively” to avoid injury. Maybe running or walking in a new kind of shoe once a week for one week and then twice a week the following week, and so on.

“In biomechanics, in human physiology, what kills you is routine,” Bermon says. “Whether it’s your bone, whether it’s your muscle, whether it’s your tendon, you have to challenge yourself every day with different things.” In shoes, just as in life, “There is nothing worse than monotony.”

Write to reports@wsj.com

Health Report

Read the full report

The Constant Beeps in Hospitals Drive Everyone Crazy. There’s a Better Way.

Natural-Fiber Clothing Is in Demand—but Some Scientists Say It Needs More Scrutiny

The Race to Save Florida Orange Industry From Total Collapse

Are Patient Portals a Blessing or a Curse? WSJ Readers Weigh In

Your Mouth Could Hold the Secrets to a Longer and Healthier Life

They Were Diagnosed With Autism as Adults. It Has Changed Their Lives.

Yes, You Should Stand Up Straight—for All Sorts of Reasons

What’s the Price Outlook for Weight-Loss Drugs?

Copyright ©2026 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8

Justin Cormack, who worked on MirageOS and Unikernel Systems back in the day, has been noticing what I've been noticing: people are discovering (or rediscovering) unikernels again. He's running a series of conversations on the topic for his newsletter, and I was first up. He emailed me, and five minutes later I was talking to him from the pub with a stein of beer in hand. We went deep on Mirage, Orleans, Haskell, Nix, Cursed and a whole lot more.

Here's the gist, written up properly, with chapter links at the bottom if you want to jump to a specific part of the conversation. Justin's edited transcript is over at Ignore Previous Directions.

Unikernels were hard. Key word: were. Now we have AI.

what a unikernel is, and why they were hard

I first ran into unikernels around 2015. I'd staffed up a team of Haskellers and gone pretty deep on functional programming. Where there are Haskellers, there are OCaml programmers, and from there you find MirageOS. Great idea. I played with it back then.

A unikernel is the idea that your application is the operating system. There's no userland. If you want a web server, DNS, or to send an email, there's nothing you can fork or spawn. You have to write those things as libraries in your application.

That was the friction. Justin remembers it well: when they were building Mirage, they had a TCP stack and an HTTPS stack, but there was almost nothing for storage. They were pulling drivers out of NetBSD because they could run them in userspace. It was hard back then.

There's a lot of dogma in our industry. Nix is hard. Bazel is hard. Unikernels are hard. Yes, they were. These hard concepts are now in the model weights. All you've got to do is prompt for them and, cognitively, get rid of the dogma that they're hard.

the operating system is design debt

Every application that isn't a unikernel was built on the assumption that there's an application, and then there's an operating system underneath. Why do we even have an operating system? Because forty years ago there was a human operator. I've done IBM 5250, AIX, Solaris, and mainframes. The multi-user operating system exists because a person sat in front of it, and then we put the application on top.

I consider that design debt, and here's why it matters right now. Applications get popped. They were getting popped before AI. Someone pops the userland application and gets a shell. That shell is a VIP butler service for exfiltration.

With a unikernel, the attack surface is much smaller. If the functionality isn't in the application (the operating system), then the attacker is screwed. There's no next hop.

Justin pushed back here, and fairly. Attack-surface reduction is something people are very fuzzy about. You can remove the shell from a Linux container, but almost every Linux environment still has something that's effectively an interpreter. You can execute a new program without a writable filesystem. You've still got memory safety and gadgets to worry about.

All true. But look at what we've been doing for twenty-six years. The earliest adage I remember from the SunOS and cgi-bin days was "don't put the compiler on production." Then came build containers and production containers. Then Chainguard. We keep chipping away at the attack surface, instead of going in the opposite direction and ensuring there is no attack surface.

And this is the bit people miss: if there's no shell and no interpreter, there's nothing in the model weights that knows what to do next. That turns a drive-by (pick your framework's RCE of the week and you've got a shell, and the model weights know what to do with shell access) into a targeted attack that needs your source code.

you can just port the missing libraries now

The classic objection: your unikernel needs to talk to Stripe, and OCaml doesn't have a Stripe library. Before AI you'd sigh and write it. Now? Run a loop to port the Go library to OCaml. Here you go: Stripe in a unikernel.

Justin had a great example of the same thing. He'd been building minimal Linux OS images for appliances, which is halfway to a unikernel anyway, since you're only running one application as PID 1. He needed to make an XFS filesystem. Rather than drag in xfsprogs and everything it brings with it, he sat down with an agent and had it write mkfs.xfs in Rust, producing byte-for-byte identical output, with every flag interpreted. It reverse-engineered the on-disk formats one by one, with tests across block sizes. It took a few hours.

That works because the original tool is a golden oracle. Generate filesystems at different sizes with both implementations and diff them. Port the tests across. Automate it.

Storage was the other big gap. Most workloads these days are cloud-shaped, even on-prem, so take the turbopuffer approach: S3 as your primary, infinitely growable storage, with a local NVMe block cache and an LRU (or whatever caching algorithm you like) for the hot bits. Justin is a massive "S3 for everything" fan too. As long as latency isn't the constraint, you get infinite storage with multi-user access, and you can build everything on it.

nix machine tests and overlays

Justin had been experimenting with Nix too, and was surprised that the first time he got an agent to prototype an OS, it built all the tests into flakes.

Nix the language sucks. Nixpkgs is great. NixOS machine tests are the bee's knees: you write a test that spins up a fleet of machines and exercises the interaction between your network rules and your application. It's the thing people don't know about.

And when something upstream is broken, or there's a supply chain problem in your dependencies, that's just an overlay.

If you have to tool-call a human, also known as "Dear Maintainer", who might be on holiday or might have abandoned the project, and wait a day, two days, or even five minutes, that's not AGI.

We're building recursive products here. Agents need the ability to modify the world as first-party source, not as third-party bundled binaries. We're going back to the contrib folder and Unix patches.

if you care about security, you have two choices

Justin asked what unikernels still need for people to discover them. Honestly, it's this. We've raised two generations of developers who don't even know they're a thing.

If you deeply care about security, there are really two choices:

1. You're sending satellites into space, and you should probably use seL4, a formally verified operating system. (I'm still a bit salty about the Australian government disbanding that team.)
2. Everyone else should seriously consider unikernels. Stop trying to harden something that is very hackable. Invert it and design from the other direction.

The other classic criticism is that many early unikernel designs ran everything at a single privilege level: your application in the same ring as the OS. In 2026, that's a prompt away from being fixed if you want ring separation. It's certainly more secure than praying to god your systemd cgroup configuration is right.

Think about how much time enterprises spend patching the world every time something new drops in Linux. Upstream now expects you to patch your kernel weekly. The week we recorded, someone popped KVM (essentially Firecracker, the core primitive we all thought was good sandboxing) and collected $50,000 from Vercel and a few other vendors. That is not much money for something that could root every managed cloud provider in the world.

spaceleans: a distributed unikernel operating system

About seven months ago, I went deep on unikernels to check whether my mental model was right. I showed Justin my Mirage folder, which holds all the functionality I needed to add.

• There was no way for a unikernel fleet to keep time, so I took an NTP client from another language and ported it. Then I built an NTP server based on RADclock and borrowed ideas from how TigerBeetle handles time: not one clock source but many, packaged as a library.
• Network stack, DNS, HTTP clients and servers, structured logging, OTel, Anthropic and OpenAI clients, and payments via Airwallex.
• A generic retry library for handling back pressure over HTTP, plus some PPX metaprogramming for fun.
• A PII wrapper at the logging boundary, so secrets and PII never leak through the logging subsystem. Every project should have one. It's pluggable; just use a functor.

And then the most cooked thing, which I'd never shown anyone before: Spaceleans, Microsoft Orleans ported to OCaml, running as a unikernel.

Orleans is a distributed actor system with transactions. You take many physical machines and merge them into one addressable heap. An actor always exists: await GetCustomer(), and if it isn't in memory, it gets rehydrated from a pluggable storage provider. You collapse your n-tier architecture into actors and stop caring whether something lives on machine A, B, C or D. The runtime handles it as an infrastructure primitive.

So in a weird sense, I built a distributed unikernel operating system out of actors, with a filesystem on top. Justin called it Erlang-esque, and he's right. I did all of it in a week. I'll probably never release it, but it falsified the idea that unikernels are hard.

sampling history

Being a little older means you can sample history, like an experienced DJ such as Carl Cox, who's been in the scene long enough to pull from previous repertoire and bring it forward. All of these ideas existed in the eighties. The models have read the papers. They've got TAPL and the most advanced type theory in their training data.

The thing that's lacking is people's curiosity and ambition to do these unhinged things, and the knowledge that previous records exist that can be sampled from.

How do we get people to try this stuff? We just do it. If you've got a turbo Lamborghini alien space rocket that's more efficient and more secure, good for you; you've got a leg up. Do cool things, attract curious newcomers, mentor them, grow. Same as it's always been.

Meanwhile, everyone else will be trying to Chainguard their Ruby on Rails application and managing AWS with fifty AWS-certified engineers, when two people with Nix and Hetzner would do. Eventually, it comes down to money. Higher-powered tools are more efficient, and efficiency wins, especially as AI collapses margins.

ocaml, rust, haskell and back pressure

Has OCaml's time come again? It's still going strong. A certain trading firm is using it very well. When I caught up with Yaron at the start of the year, I asked him whether OxCaml exists so their language extensions end up in the training data and lift the whole company. I got a very "no comment" smile.

For agents, OCaml is lovely. Functors between modules are beautiful. The .mli files, a typed header explaining how a module should work, are really efficient context for agents. opam and Dune are legitimately good. Hindley–Milner. And compile times are fast. I see no reason to do F# these days.

Justin has mostly been writing Rust, and agents are good at Rust. But compile time is the tax on back pressure. LLMs hallucinate, and when compilation is slow, each hallucination is expensive because you get fewer attempts per minute. Justin's S3 clone is about a million lines of Rust; with four agents compiling at once, they fight over disk and CPU. You end up spending more on fast machines than on tokens.

Haskell's type system is great, and the models do it really well. But I don't feel good running it in production: a space leak lives in the runtime state space and only shows up in production. Justin pointed out that the linear-type ideas in Rust came out of Haskell papers trying to solve exactly that. Then there's Zig's approach: allocate everything up front and never allocate again, which is what game devs did in the eighties and nineties. It's hard to persuade an agent to do that in Rust, though, because constant-memory programs aren't in its training set.

I think dependent types are the winner for next-generation languages. Anything that lets you codify more into the type system is more back pressure. You probably won't be surprised to hear I've got a fork of the Rust toolchain with dependent types. You can just do things now.

languages for agents, and what cursed taught me

The pace of language development has been held back by how fast humans can learn new concepts. Operator chaining is essentially sugar for humans. If agents write the code, we can lean on forty years of academic PLT research, as long as you know how to sample it.

The industry codified "do not make breaking changes" after Python 2 to 3. Justin knew companies with hundreds of people on that migration for years. I think that rule is no longer true. Ship a skill pack with the breaking change and let agents auto-migrate.

Justin asked what it actually costs to make a new language successful now. Go was the last language a company spent real money on, and it took a long time. I can answer that one.

Cursed was built with Sonnet 3.5 and 3.7, a deliberately underspecified prompt, and three months of running it in a loop. I started in C (not enough back pressure; I wasted too much time in Valgrind as the agent clobbered its own updates), then Rust, then Zig. Zig was a mistake; it would work today if I'd stayed with Rust. It cost roughly US$6,000, and I did it three times over. Compare that to what Go cost.

Now the real bit, the part that still scares me. If you allocate the context window correctly — a lookup table of the lexical structure and grammar — the model can program in a language that isn't in its weights. It's brute force and inefficient, but it works.

Think T-diagrams (tombstone diagrams). Lock down your grammar and lexical structure, reach a stage-two self-hosting compiler, ship a sensible standard library, and start the next training run. From there, you can reach a Roslyn-style self-hosted compiler with language services stupidly fast. Justin asked whether fine-tuning an open model would help bootstrap a language like this. It's not needed.

That was true a year and a half ago with much weaker models. It'll take just one programming language designer going all in with the good models to shock the world.

what next

As the pub was shutting, Justin asked what was on my mind. If you haven't read my latest post, go read it. If you manage people, create the space and time for them to experiment now, because within six months, leadership will ask you to put people on a vitality curve.

Yes, the labs trained on the commons. I hate that, and I get it. But you trade time and skill for money; employers have minimum standards, and those standards have changed faster than ever before in our industry. Be curious, learn how to build an agent, and go create beautiful stuff. We're in a renaissance.

It's a time-compression device. The more experience you have, the more you can sample. Not everything ships; some of what I showed Justin may never see the light of day. I use these projects as katas and redo them when the models get better.

But if you want to build something secure, seriously consider unikernels.

chapters

• 0:21 — discovering unikernels: Haskell to OCaml to Mirage
• 0:55 — what a unikernel is, and why it was hard
• 2:45 — hard was past tense
• 3:33 — why do we have an operating system?
• 4:17 — the shell is a butler service for exfiltration
• 5:12 — attack surface reduction is fuzzy
• 7:26 — don't put the compiler on production
• 9:13 — porting Stripe into a unikernel
• 9:54 — minimal Linux and mkfs.xfs in Rust
• 12:22 — S3 for everything
• 13:48 — Nix, machine tests and overlays
• 15:28 — tool-calling a human is not AGI
• 15:53 — seL4 or unikernels
• 16:56 — privilege rings
• 18:24 — weekly kernel patches and the KVM escape
• 19:38 — demo: the Mirage folder, NTP and time
• 22:05 — Spaceleans: Orleans in OCaml as a unikernel
• 25:31 — sampling history
• 26:52 — how to get people to try this: just do it
• 28:57 — OCaml and OxCaml
• 31:45 — Rust compile times and back pressure
• 33:06 — Haskell, space leaks and dependent types
• 35:13 — memory strategies: Rust, Zig and game devs
• 36:10 — languages designed for agents
• 37:41 — breaking changes and skill packs
• 38:53 — Cursed
• 42:36 — what Cursed taught me
• 46:47 — what's next: AI use is mandatory

Keep curious.