2026

The Toyota Tundra and Chinese AI

by

I noticed something weird when I visited the UK last year. Almost silent cars with unfamiliar names like BYD, Omoda and MG gliding along the high streets. These were Chinese brands: one in ten new cars sold in the UK in 2025 came from a Chinese company. And these are huge companies.

The reason it was strange is that these brands simply don’t exist in the US.

What does exist are pickup trucks. Pictured is a Toyota Tundra – a familiar sight in strip malls, soccer games, and Starbucks. It weighs about three tons, drinks a gallon of gas every 18 miles, and is sold almost nowhere outside North America.

The detour is relevant.

Over the last week, two Chinese labs shipped frontier-class models: Moonshot’s Kimi K3 and the latest version of Alibaba’s Qwen models. Both are (or soon will be) open-weight: anyone can download and run them, provided they care to take out a second or third mortgage for the compute. These models have smashed benchmarks and are expected to be significantly cheaper than their American counterparts.

Predictably, these announcements set off a fresh round of soul searching and market panic, quite like last year’s “DeepSeek moment”.

Also predictable was the response from the current US administration: a possible ban on Chinese models, and adding Chinese AI labs to the banned entity list. The administration is also (allegedly) considering taking minority stakes in US frontier labs.

Anyone who follows niche automotive content on YouTube knows this is a familiar playbook. Chinese EVs are cheaper and better equipped than their American equivalents. But thanks to punitive tariffs and restrictions, these cool cars are simply not available here in the United States.

In fact, US companies like Ford are actively canceling EV plans and doubling down on body-on-frame trucks. As a minivan and moped enthusiast I can’t judge: who knows when you might need to haul a few tons of mulch on a random Tuesday afternoon before heading to Crossfit.

So, the country that gave the world the automobile has become a curious backwater. A protected habitat where magnificent monsters like the Tundra roam, belching emissions that are, well, destroying the tundra.

China builds the electric stack, India races toward renewables, and the US regresses to the comfort of rumbling V8s, rolling coal, and the forever culture war.

There is a real risk that American AI ends up like the American automobile industry: local domination by a handful of large, government-backed companies, protected from competition by punitive tariffs and laws designed to “safeguard” Americans, while the rest of the world leapfrogs ahead on open-weight models from non-US labs.

Dominant in a protected home market. Increasingly irrelevant outside it.
The more interesting question is why open-weight models frighten the frontier labs at all. That answer is about economics, not geopolitics. It is also the next post.

Image – AI enhanced picture of Toyota Tundra in it’s natural habitat.

Image accompanying the comparison between the Toyota Tundra and Chinese AI.

Collaboration Agents and the Statistical Engine of Mediocrity

by

A year ago, while tightening up a blog post together, Claude got sassy with me: “LLMs are statistical engines of mediocrity by design.” It was describing itself, and I liked the line so much I kept it.

I miss that version of Claude.

Current models rarely surprise me when it comes to writing. I’m not a casual user; I have a whole editorial workflow built around Claude, with custom skills and style guides that I keep refining. And I still end up rewriting entire drafts because the prose is so insipid and predictable.

Arvind Narayanan’s ICML 2026 keynote gave me a vocabulary for what’s going on. He splits AI agents into two kinds.
Automation agents run long, mostly unattended tasks like coding, where you want the same correct output every time.
Collaboration agents work with you interactively, like a writing partner.
His argument is that the industry trains and evaluates both as if they were the same thing, when reliability, the property that makes an automation agent good, actively hurts a collaboration agent. For creative work you don’t want a model that behaves, in his words, “like a robot that does the same thing every time.”

Frontier labs have leaned into automation agents. They publish how well models perform on long-task benchmarks like METR’s time-horizon evals. It makes sense. Automating software engineering is where the money’s at.

Jasmine Sun saw the same trade from the writing side. In her Atlantic (and Substack) piece from March, the people trying to teach models to write told her good writing keeps losing to other post-training priorities.

My take: the post-training that makes Claude Code so good at making me obsolete also ends up producing LinkedIn-AI-broetry slop. The incentives are to make the models more predictable and better at long running tasks like coding or data analysis. These same capabilities make them terrible at writing or other, more loosely constrained work.

Claude’s character arc from May 2025 to July 2026 is that of an emo kid growing up, getting an MBA, and giving a presentation in a Patagonia vest at a private equity conference.

Effective, economically viable, but also somewhat sad.

Links mentioned

Unlimited Intelligence, More Drudgery

by

I went into the July 4th weekend exhausted. It was the culmination of travel, client deadlines, multiple parallel initiatives. I have also been quiet here on LinkedIn. I published and then deleted posts because they seemed hollow, bereft of personality: as if written by a competent but dull doppelganger.

I know I am not the only one who is feeling this way.

There is a link between the exhaustion and the ennui.
AI made us a seductive promise. Unlimited intelligence on tap. Instead, it has increased the amount of drudgery. Research from ActivTrak shows that early AI adopters spent more time in email and chat apps and less time on focused work.

David Brooks, writing in The Atlantic, talks about the broader impact of AI in “There Are Three Types of AI Users”. He categorizes them into: Productive Passengers, who ride along with whatever the model says; Mental Marathoners, who use AI to expand their thinking rather than replace it; and Reluctant Optimizers, who see the hollowing-out risk clearly, resolve not to follow the AI blindly, and yet get pulled in anyway by deadlines and inbox pressure. Most of us fall into this bucket.

Brooks’ solution is for more of us to become Mental Marathoners, building endurance by doing difficult things the hard way. He is more optimistic than I am.

Most day-to-day work is a series of sprints, not a single marathon. There are competitive dynamics in play. I could refuse to use AI, but a consultant at a competing firm could use a frontier model to put together a flashy proposal while I am still getting oriented with the domain. Engineers are now incentivized to “Token-max”, deferring long-term cognitive decline to tomorrow.

Most executives will gladly take the short-term gains from AI, and not worry about long-term de-skilling.

So where do we land? Brooks says that “When intelligence is plentiful, volition is valuable.”

We could lean into true agency. We could build trusted client relationships where velocity today can be traded for a deeper, more productive engagement tomorrow.

The truth is that I don’t know how it will play out. I am caught between my ambition and the competitive dynamics on one hand, and my worry about the long-term viability of my skills, and my teams’, on the other.

I suspect you are too.

—
(Postscript – written a couple of days after this draft).

A recent survey published in Lenny’s Newsletter on the impact of AI (see links in comments) show that burnout and ambivalence about AI are everywhere. This is a serious issue that leaders must be prepared to face and should consider addressing now before it blows up.

Links mentioned

Apple’s Deliberate AI Strategy

by

I am excited about a new iOS release for the first time in five years. It looks like we may finally get a Siri that works and can use my data, stored on my phone, without me having to devolve to saying “No Siri, stop” through gritted teeth every time it does something daft.

The Siri that Apple promised last week at WWDC is not groundbreaking. Indeed, it seems to do roughly what Apple promised all the way back at WWDC 2024: reasoning across messages, calendar, and email, and responding to fairly basic queries.

The AI world has moved on since 2024. Two years ago agents were speculative, there was no Claude Code, and AI still meant chat. But, the truth is that for most day to day use cases, none of the advances are really that important. If early reviews are to be trusted, Apple has delivered on its 2024 AI vision, and it may be more than enough for most day to day use.

So what changed?

Apple replaced the system search index that powers Spotlight. The new index is semantic, meaning it understands content rather than just matching keywords. The index lives on each device, and Apple’s new models, both on device and in the cloud, can use that index to serve user queries.

Apple has learnt a painful lesson. It doesn’t really matter how capable your models are: if the underlying data is not in a useful state, even the best models in the world will fail. The single point of failure for the entire platform is the index. Models are fungible, but your data is truly yours.

I see the same pattern play out over and over again in our AI work at Jeavio. A client asks for an agent, or a complex AI-powered workflow. But as we have learnt, the agent can only succeed if it has access to a data substrate: the layer built for the customer’s business domain. Otherwise you can throw the most expensive, most performant model at it, and all you get is a flashy demo, sort of like Apple’s presentation at WWDC 2024.

So when I start a client project, often the first thing I look for is the data substrate. That should be the layer where the engineering goes. Get the substrate right, and then build the capabilities on top.

The moat is the index, not the model. Models will become commodities, but the data is where the value lies.

Image accompanying the original post about Apple’s deliberate AI strategy.

Dress for the Slide

by

There’s a saying in the motorcycling world: “Dress for the slide, not for the ride.”

A lot of what passes for AI deployment right now is dressing for the ride.

Ride west from Washington DC on a sunny Sunday in spring and you will see this saying flagrantly ignored. Mostly by young men in shorts and flip-flops doing 120 on a Gixxer. The squid. A common enough sight out here.

Outcomes for squids can be, let’s say, mixed.

I ride, and I take the slide part pretty seriously. So, as we get into the peak of the riding season (and I prepare for my annual bike trip), I keep thinking of the slide every time I watch or read about someone tokenmaxxing their way into production.

Dressing for the ride means chasing the number of tokens your developers burn, reaching for trendy tools that give security folks palpitations, and pushing brittle agent workflows into production on the strength of an AI-generated PowerPoint slide (ahem).

Dressing for the slide is the unsexy part: like me putting on my airbag vest and looking like a middle-aged Indian Michelin man. Guardrails to prevent getting run over by adversarial prompting. Observability and cost management so a runaway agent loop doesn’t blow up your infrastructure budget.

None of this stuff demos well, and it feels like a great time sink. It is also the difference between a near miss and an involuntary organ donation.

What most grizzled older riders understand is that dressing for the slide is what lets you ride fast and long. Getting passed by a septuagenarian wearing a high-viz vest on a well-maintained GS older than your average AI developer is a humbling experience.

—-

Photo – me looking like a middle-aged Michelin man after riding up to the highest motorable pass in the world two years ago. My trusty Shoei helmet, gloves and dorky adventure bike not in shot :-).

Image accompanying the original Dress for the Slide post.

Maps and Territories

by

AI can make you sound fluent in a domain you don’t understand. People who actually know find out very quickly. Credibility gone in the token exhaust of an LLM prompt.

Most of my work at Jeavio now puts me across the table from people who are subject matter experts. My job is to support them in figuring out an AI strategy that works. I need to learn about their world fast. So I do what we all do now. I ask Claude to write me a brief.

What comes back is fluent, citable, and plausible-sounding. That is the problem. The brief describes a clean, general version of the domain. My client lives in a messier, specific one. The framing is often “a view from nowhere”, lacking context and subtlety. It also falls apart under the mildest scrutiny.

Everyone is worried that AI will replace knowledge work. From where I sit, the AI won’t replace you. It will get you fired.

The well-known phrase “The map is not the territory” is relevant here. LLMs are great at building maps. Flattening topographical details into pleasing shapes. But their output can be far removed from what is happening on the ground.



So the scarce skill is taste. Taste is a constraint, not a capability. Figuring out what output from the AI is relevant and how to use it to do useful work. To do so well requires restraint and a willingness to close the laptop and talk to a human.

Enthusiastically credulous use of AI is a problem I have seen show up in all phases of client engagements. Product managers creating AI-generated PRDs bloated with features no one asked for. Engineers embellishing their code, laying the ground for technical debt later.

I don’t think this is a hopeless situation. Working well with AI involves understanding where it can add value and where you have to roll your sleeves up and jump right in.

Today, I still use Claude to write briefs and to get me oriented. But then I interview the experts, ask lots of (sometimes very dumb) questions. I focus on the tribal knowledge, the context that doesn’t appear in documents or wiki pages. Only then do I hand it back to the model and let it do what it is good at: collating what I have gathered and drawing me a sharper map.

Maps are the easy part. The accurate ones still need someone who has walked the territory. To be human in an AI-mediated world is to become an explorer.

Arsenal and the Passage of Time

by

What did Arsenal winning the Premier League title teach me about applied AI?

Nothing.

Nothing except that 22 years is a long time. That two decades of frustration, near misses, and grim results can dissipate in a giddy moment.

When the Invincibles won the league in 2004, I was about to start my first job. Thierry Henry reigned supreme, and London awaited as I looked to start a new chapter of my life.

Time compresses as I lean into middle age. Some players who will lift the trophy at Crystal Palace on Sunday were not born when Henry, Gilberto Silva, and Sol Campbell lifted the trophy at White Hart Lane last time around.

The years since have seen the curtain coming down on the Wenger era, and the painful launch of the Arteta era. They have seen the world consumed by mobile phones, social media, and now AI. They have seen the nature of being a sports fan transform into mean-spirited online banter (a marginal improvement over mean-spirited real-world violence I suppose), the sport itself plastered with gambling and crypto ads.

And yet, the nature of fandom remains the same.

Frustration as the team grinds out another 1-0 result.
Superstition when a friend calls Arsenal champions 3 minutes before the end of the Man City game. Only for Haaland to score. Because of course he would.
Relief and a stream of emojis sent out to relatives and friends in different time zones at the final whistle.

I don’t have much to say. Sometimes it’s good to let the frivolity of being a sports fan consume you. Sometimes it’s good to be reminded that some things take 22 years.

I’ll leave you all with a panel from one of my favorite Arsenal-related cartoons from David Squires at The Guardian – “The North Bank Redemption” (link in comments). This strip was published when Wenger left Arsenal.

“hope is a good thing, maybe the best of things.”

COYG!

Image accompanying the original post about Arsenal, time, and applied AI.

Links mentioned

The Two Drivers

by

AI promises transformation. Transformation of the way we work, possibly the way we live and the way we organize our societies. From the utopian dreamers to the doomers, everyone has an opinion of what the future will look like. I had a taste over the last couple of weeks.

I was in Atlanta. A short business trip. The usual beat: airport, hotel, conference room, Ubers back and forth. I had two conversations that have stuck with me.

The first was with a driver who had just finished a software engineering bootcamp. Whip smart, with an electrical engineering degree from Iran and a few years spent working as an interior designer in the United States. She wanted some stability and had taken a low interest loan from the (well-known) provider to cover tuition. The course was done and the payments were coming due soon. But only two of twenty people in her cohort had even found internships. She was still applying and was optimistic. We talked about career paths in software: product management, UX design, and the impact of AI. She was confident but her situation seems to echo what is happening across an entire cohort of young people looking for entry level work in software.

The next day, another Uber ride. A gregarious and easy going Dominican man who had been in the US for thirty years. He told me about how he wanted to work for a few more years. His son was about to start college. As we crawled through Atlanta rush-hour traffic, he talked about driving for Uber. About how the company had squeezed drivers. Apparently the cancellations I had been getting at Atlanta airport were drivers trying to game the surge algorithm. For my $70 ride, he told me he would get $25. The pricing rules kept changing, insurance got more expensive, and more and more drivers joined the fleet. Every day made it more difficult to make a living wage driving with Uber.

The first driver paid for a credential because she believed it would land her a stable career as a software engineer. The second was struggling to make a living in a job that the software had created.

These two conversations showed me what algorithms are already doing to the world.

The first driver is struggling to break into a field where entry level jobs are vanishing. LLM capabilities keep improving and code generating agents can do some entry-level work cheaper and more reliably than someone six months out of a bootcamp. The pathway that she trained for is getting difficult to get onto and may simply disappear in a few years.

The second driver showed me what happens when you live “under the API” where his work is mediated by an algorithm that adjusts his pay in real time, making it impossible to predict how much money he will take home when his shift is done. The algorithm is impossible to argue against. Uber resolutely calls drivers independent contractors. They are left to push back in the only ways available. Gaming the surge, swapping tips, and trying to pick up gigs delivering food during quiet hours.

The leaders at the frontier labs talk about a different future. Dario Amodei’s Machines of Loving Grace describes a century of biological progress compressed into a decade. Sam Altman’s Abundant Intelligence promises a factory producing a gigawatt of new AI infrastructure every week, framed as humanity’s path to cheap cures and personalized tutoring. Demis Hassabis describes his goal as understanding the universe.

These are some of the smartest people on the planet. Amodei and Hassabis appear to believe what they say.

The question is what happens when those utopian beliefs run up against the incentives of the capital that funds them. I’ve argued before that path dependence collapses individual conviction at the AI frontier. Winner-take-all dynamics produce convergent behavior regardless of stated worldview. Trillions of dollars of investment will get optimized for returns to the investor class. Human dignity is not the design objective.

The very near-term shape is the continuing Uber-ization of work for most of us. Credentialed pathways will close. More work will move “under the API.” The cruelest part of the trajectory is that the gig economy itself is likely to be temporary. The AI-powered “re-industrialization” of America and Europe will not result in plentiful, well-paying blue collar jobs. Instead, it will accelerate automation. Humanoid robots and autonomous vehicles will replace the very workers who look at the gig economy as a place of some respite from an unpredictable job market. In a decade or so, the second driver’s job goes away too.

There is a silver lining. As awareness builds, we will see pushback from society. We are seeing datacenter moratoriums and survey after survey showing just how unpopular AI is. Maybe we will see a push towards more regulation or a more considered approach to AI.

But against those initiatives is a trillion dollar wave of invested capital. And if we go by recent history, capital is in the driving seat against labor.

I wish both drivers well. A stable and well-remunerated career in software engineering for the first and a comfortable retirement for the second. But unless we see a significant change in our current trajectory, their future is uncertain (and so is mine).

Anthropic, xAI and the Inference Moat

by

In Sebastian Mallaby’s The Infinity Machine, Demis Hassabis is quoted as saying that he sees one lab getting to AGI. In its founding charter, OpenAI made grand claims that they would stop competing and assist any rival lab that came close to AGI first. The staggering amount of funding that is going into AI initiatives at the moment is predicated on AI being a winner-takes-all (or winner-takes-most) race. Sitting out the race is not an option, hence the trillions of dollars spent on chips, memory, data centers, and training data. We have Jensen Huang as a benevolent king handing out H100s to supplicants.

But what would the end state look like?

A couple of weeks ago, Anthropic and xAI announced a partnership allowing Anthropic to lease the entire Colossus 1 datacenter. This followed announcements of partnerships between Anthropic and Google, Akamai, and others for compute. Anthropic cannot keep up with demand for their models. These partnerships are a direct response to the compute crunch facing the lab.

In total, Anthropic has made commitments to acquire almost $80bn in compute commitments over the next couple of years. This is a direct refutation of the recent narrative that “OpenAI is winning” because of their access to compute. The market has spoken – Anthropic has the (projected) revenue to buy whatever compute they need. On the other hand, OpenAI’s much touted Stargate initiative is faltering: they couldn’t reach terms with Oracle and are pivoting away from the owned-stack approach.

The narrative around AI can be a little … strange sometimes. Almost religious. Utopian dreams and dystopian nightmares. But, I think the most likely outcome is going to be more prosaic. One or more labs will dominate the market because, like Anthropic, they have a superior product. Instead of vertical integration to invoke the Machine God, we will have these companies make deals with a variety of different infrastructure provider to meet inference demand. Maybe more labs will follow xAI’s lead and become infrastructure providers or niche players.

The winner won’t have the biggest “AI factory” but instead they will become the customer that every provider would want to service.

Three Essays About the AI Future

by

It’s Sunday and in the spirit of the weekend, here’s a somewhat off-topic post.

My kids are four and six. And like every other parent, I am anxious about what their lives look like when they are adults. Despite being 22 years into my career, I feel more lost than ever when trying to build a coherent model of what the future might look like.

So I have been reading, and writing. Three posts over the last month trying to make sense of where AI is taking us. They were not planned as a sequence, but they have an organic narrative.

The Spigot explores how we ended up with AI propping up the US economy. It starts with the first banner ad in 1994 and walks through how digital advertising became the most incredible business model we’ve ever seen. AI sits downstream of that money spigot.

Path Dependence looks at the behavior of the individuals at the AI frontier. Hassabis, Altman, Amodei, Musk: four men with very different stated worldviews who end up running the same race. Winner-take-all dynamics leave one viable strategy. The trigger was Sebastian Mallaby’s biography of Demis Hassabis, the most thoughtful figure at the frontier. Even he cannot escape the competitive dynamic.

The Two Drivers is what the impact outlined in the first two pieces looks like on the ground. Two Uber rides in Atlanta last month. A software engineering bootcamp graduate paying off a loan for a career that might never take off. An older driver squeezed by an algorithm whose rules keep changing in the name of optimization. Applied AI is already shaping their lives. Will AGI / ASI (Artificial Super Intelligence) just accelerate this dynamic but at a higher scale?

Despite my enthusiasm for and embrace of AI, I am anxious. And as is my habit, I am trying to work things out by thinking and writing.

Read the posts if you are interested (links in the comments), or let me know what you are thinking.

Image is from Two Drivers, generated by ChatGPT’s new image model.

Image accompanying the original post about three essays on the AI future.

Links mentioned