It’s Pentecost Sunday and we all enjoy a long weekend as Jesus or somebody else must have done something really grand on Monday as well as on Sunday. We all don’t mind, even though we don’t know what this church holiday is actually about and we do not bother to look it up. While less and less people go to church or belong to a church, we sure as hell like their holidays.
Anyhow, it’s finally something like pre-summer here in Northern Germany and the city feels so different as people shift their activities to the outdoors. The Alster, Hamburg’s inner-city lake that is actually a river, was packed with people on SUP boards, boats and anything else that floats.
For those who want to know more: Pentecost actually happens 50 days after Easter and marks the end of Easter time, which is the biggest Christian holiday, even though people get presents on Christmas and just chocolate eggs and bunnies on Easter. Sadly, no candy is involved in Pentecost, not even a special meal. The only ritual is a long weekend and time for barbecue.

Subscribe now

Can America Escape the Cycle of Vicemaxxing?

After years of conservatives criticizing the left for “virtue signaling”—that is, cravenly performing a version of virtue for public approval—we now have something even worse than its opposite. The president and his allies are not merely vice-signaling. By empowering a figure who is oblivious to virtue and choosing to ignore his crescendoing depravity, we are creating a mode of politics that openly celebrates the death of morality.

This is the age of vicemaxxing. The question is whether this is our new normal—or, I hope, the sort of cultural overreach that shocks our collective conscience and sets the stage for a more decent politics.

I have been waiting for the backlash and while it happened in Hungary, I do not see it happen anytime soon in the USA. And Germany is headed in the same direction with the far right AfD who do as they please and somehow people just shrug it of when they break campaign finance rules or hire family members.

(…continue reading.)

Actually, Democracy Dies in H.R.

“Making a Career in Dictatorship,” a new book by two German political scientists, Adam Scharpf and Christian Glassel, reads like what you might get if you crossed Hannah Arendt’s ideas about the “banality of evil” with a business school guide on how to get the most out of low performers.

Their in-depth study of Argentina’s military during that country’s era of coups and forced disappearances found that low performers — whom they refer to as “career-pressured” individuals — filled the ranks of the secret police. That service allowed them to “detour” around the ordinary military hierarchy, the book shows, achieving promotions and career success they could never have managed otherwise.

It turns out that would-be authoritarians don’t need to staff their regimes with ideological true believers, offer extreme enticements or impose draconian punishments in order to make successful power grabs. They just need to figure out how to target their ideal labor pool: the frustrated and mediocre.

I am not the least bit surprised. So let’s please invest much more into education across the board.

(…continue reading.)

I’m a Normie. Can Normies Really Vibe Code?

The idea that the internet might redistribute power in a meaningful way has been giving cover to Big Tech and its enablers since forever, and I have no illusions that a blast of amateur coding will claw back our time and agency from the purveyors of sludge. Nevertheless, my blast of amateur coding is now live on Netlify. For all I know it will crash tomorrow. But it undeniably exists—a shared civic ledger where once there was only frustration. Sludge thrives on exhausting us in the dark and the assumption that our individual wasted hours don’t add up to anything. They do. I’ve got a janky database to prove it.

I really think everybody should try to build a website. More than a decade ago I advocated for Javascript being taught as the second foreign language in school. While I know that not everyone is really interesting in coding, I think it helps tremendously to understand just a tiny bit about how these interwebs work and what one can do with it. It can be very empowering when you figure out what the possibilities are. Obviously, you won’t be able to just say “build me a better facebook!”, lean back and five minutes later you are done and humanity has been saved. But vibecoding can help unleashing your creativity. Remember to push back whenever Claude or ChatGPT or Gemini tell you that your idea is amazing and not yet done. Make it do its research. And then do yours.

(…continue reading.)

Roadside Attraction

According to the Texas State Historical Association, the lights were first spotted in 1883, long before there were headlights out in West Texas. A young cowhand named Robert Reed Ellison saw a faraway flicker of light while he was driving cattle between Marfa and the neighboring town of Alpine. He thought it may have been an Apache campfire in the distance. He spoke about this with other settlers, and it turned out many of them had spotted the distant shimmer too, but when they had investigated, there was no campsite to be found.

There are other myths that lean more towards the supernatural: the lights are the spirits of Apache warriors killed by white settlers, delivered back to their beloved land in the afterlife. Or, in line with other Southwestern tales of atomic testing, the lights might be remnants of laser fusion weapons experiments gone awry—tests which knocked holes into space. They say the holes attract the lost—people who have disappeared in the area are thought to have fallen through them into some liminal zone, floating forever in another dimension.

I still remember my road trip I did through the South West some 30 years ago and I really want to go back.But my family has instituted a travel ban to the USA until Trump and his MAGA croonies are finally gone and done.

(…continue reading.)

The Prehistory of A.I. Slop

The idea of mechanically produced prose or poetry is not especially new. Eighteenth-century letter-writing manuals provided fill-in-the-blank templates, because many types of correspondence are set forms: letters of condolence, say, or letters of recommendation. Anxiety about machines replacing humans as writers, and replacing good writing with bad, is also older than you might think. Mid-nineteenth-century commentators, overwhelmed by the era’s flood of cheap printed material, especially periodicals and novels, imagined a “New Magazine Machine” that could spit out cheap pulps, and a “Book-Making Machine,” a literary successor to Charles Babbage’s Analytical Engine.

Actually, I think that AI Slop will be something of the past real soon now. First of all, machines will get better at writing. And we get more accustomed to it. Also, when we ask AI on advise to write better, we will incorporate its style. I invested in a company over 10 years ago that created articles from data and newsrooms were appalled. Now they all use these kind of tools.

(…continue reading.)

That’s it. Have a great Sunday! If you missed last Sunday’s edition of Five Things, have a look here:

Subscribe now

— Nico

Five tech things (see what I did there?) that made me think this week. One involves particles in the stratosphere. One involves developers losing their edge. And one is the best community response to a corporate collapse I have seen in years.

The throughline, if there is one: technology keeps moving faster than our ability to govern it, understand it, or even agree on whether it is good for us. Geoengineering gets trialed before regulation exists. AI gets mandated before anyone asks developers what it does to their craft. Cars ship as software black boxes and owners only find out what that means when the company goes under. The Fisker community figured it out anyway, and that gives me more hope than most funded startups I see right now.

Enjoy Five Things Tech!

Can Some Very Tiny Particles Cool the Planet? One Tech Company Says Yes.

The materials produced by Stardust are made from amorphous silica, which is used as a food additive and in some consumer products, and calcium carbonate, a compound found in eggshells and limestone. The company said its particles were biodegradable, were not harmful to people or animals, and would not accumulate in the oceans or soil. Released in the upper atmosphere, the particles could reflect a small amount of sunlight away from Earth, the company said.

Geoengineering - is it good or bad? This idea certainly sounds like something a mad scientist in a dystopian science-fiction movie would try to develop, only to be stopped by the good guys. Or can we really safely use this globally to keep the temperatures down and make hotter regions livable again?

(…continue reading.)

Venture Capital Is Concentrating Faster Than Ever. What Happens To Everyone Else?

The fundamental question for the venture world now is: Does the rapid growth and capital concentration for the largest companies come at the expense of smaller startups, or does the success of OpenAI, Anthropic and the like expand the total addressable market for tech startups so drastically that it promises to grow and reshape opportunities across the entire ecosystem?

Everybody wants to be in on the mega-rounds, the safe bets of the probable industry leaders. It makes it very hard for startups that target specialist niches to gain any traction in their fundraising efforts.

(…continue reading.)

Fisker went bankrupt and owners built open source car company from the ashes

Within months of the bankruptcy filing, thousands of Ocean owners formed the Fisker Owners Association (FOA) — a nonprofit that quickly grew to 4,000 members and began operating as something between a car club, a tech startup, and an independent automaker.

The FOA hired independent tech experts who began reverse-engineering Fisker’s proprietary software patches. Members taught each other how to flash firmware. They organized bulk purchases of replacement parts — negotiating the price of key fobs down from roughly $1,000 each to a fraction of that through coordinated group buys. They hosted free global key fob pairing events, saving each owner $100 to $250.

This is a genuinely cool response to what could have been a total desaster. Cars are software black boxes and we need to better regulate this. Also, I’d love to be able to configure my own Android Auto or whatever it is called, because the latest release, which was 2 months ago, ruined my bluetooth…

(…continue reading.)

Software Developers Say AI Is Rotting Their Brains

Developers who are told to use AI whether they like it or not, however, tell a different story. On Reddit, Hacker News and other places where people in software development talk to each other, more and more people are becoming disillusioned with the promise of code generated by large language models. Developers talk not just about how the AI output is often flawed, but that using AI to get the job done is often a more time consuming, harder, and more frustrating experience because they have to go through the output and fix its mistakes. More concerning, developers who use AI at work report that they feel like they are de-skilling themselves and losing their ability to do their jobs as well as they used to.

Oh well, get used to it. I really am not that sympathetic. Change is coming fast and we have to embrace it. The job of a developer will change tremendously, but developers will stay valuable, or becoming even more valuable when we do agentic coding and need to orchestrate agents to build software.

(…continue reading.)

World’s first native color LiDAR gives machines human-like vision

Ouster's technology is built around its new L4 chip, which embeds Fujifilm 's color science – the same expertise behind the company's imaging technology – to deliver hardware-level color processing. It integrates 42.9 GMACs of processing capacity, detects up to 20 trillion photons per second, and operates at 40 kHz with picosecond-level precision. Those numbers are dense, but they mean a single sensor can now read a traffic sign, detect whether the car ahead is braking by the color of its brake lights or produce topographic maps with real-world color data – all without additional hardware or calibration.

That is just really amazing and will improve the safety of autonomous vehicles even more.

(…continue reading.)

That’s all for now! Thanks for reading! If you missed last week’s Five Things Tech, you can find it here:

🤖

— Nico

Subscribe now

This week's reading list covers a lot of ground: Google revealed that its real product is no longer search but an agentic operating layer stitching together everything it owns; Anthropic hosted a room full of developers who cheerfully admitted to shipping code they had not read; a Wired piece confirmed that AI has well and truly escaped the lab and is now everyone's problem whether they opted in or not; OpenAI's model disproved an 80-year-old maths assumption but still cannot solve the actual problem; and apparently the one thing AI genuinely cannot do is be funny. A full week in five links — grab a coffee.

Subscribe now

Google I/O 2026 Was Not Just a Model Launch. It Was Google Showing the Agent Stack.

The signal is that Google is stitching together models, tools, sandboxes, browsers, Search, Workspace, Android, Chrome, and Cloud into one agentic operating layer. That sounds abstract until you line up the announcements. Gemini 3.5 Flash is the reasoning and action model. Antigravity is the orchestration harness. Managed Agents give developers cloud-hosted execution environments. Spark brings background agents into the Gemini app and Workspace. Search uses the same model-and-harness combination to build custom interfaces, dashboards, and trackers. Gemini Omni extends the same ambition into multimodal creation. Chrome and WebMCP try to make the web legible to agents.

Google really delivered on the agentic promise this week.

(…continue reading.)

Anthropic’s Code with Claude showed off coding’s future—whether you like it or not

It’s not news that LLM-powered tools like Anthropic’s Claude Code and OpenAI’s Codex have upended the way software gets made. Top tech companies now like to boast of how little code their developers write by hand. (“Most software at Anthropic is now written by Claude,” Hadfield said. “Claude has written most of the code in Claude Code.”) OpenAI, Google, and Microsoft make similar claims. Many others wish they could.

Even so, it is striking how normal this new paradigm already seems, and how fast it has set in. This was the second year that Anthropic has put on developer events, which also run in San Francisco and Tokyo. This time last year, the company had just released Claude 4. It could code, kind of. But with Anthropic’s latest string of updates—especially Claude 4.6 and then 4.7, released in February and April—Claude Code is a tool that more and more developers seem happy to hand their work off to.

Writing code is so last year.

(…continue reading.)

AI Has Broken Containment

The newly chaotic and inescapable state of AI is the result of two inflection points. The first came at the start of the year, when AI agents exploded in popularity. Products such as Anthropic’s Claude Code and OpenAI’s Codex don’t just talk to you; they can do things on your behalf—code, trade stocks, analyze spreadsheets, generate slide decks, and even create Amazon listings. The technology’s once-questionable economic value became very clear, very quickly, to a large number of businesses, which have clamored to incorporate agents alongside, or in lieu of, their human employees. As agents have swarmed the workplace, nearly three-quarters of employed Americans think AI will decrease overall job opportunities and 30 percent of Americans are concerned that AI will make their own job obsolete.

AI truly has an image problem, which stems from people like Sam Altman and Elon Musk, but also from executives who cannot think about creative new possibilities with AI, but rather just use it to cut jobs.

(…continue reading.)

OpenAI makes breakthrough on 80-year-old maths problem

“For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids,” OpenAI wrote on X. “An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better.”

While the work has excited mathematicians, the broader problem remains unsolved because the AI did not come up with a new answer for how fast the pairs of dots rise, but merely showed that the limit Erdős proposed was too low.

AI will advance science in so many ways, it will be breathtaking.

(…continue reading.)

Why Can’t AI Tell Jokes?

Despite AI being able to assist with science and knowing everything on the internet, it hasn’t come up with any novel scientific results on its own. That’s surprising. By default, I would expect that someone who learned everything in all of science would figure out lots of new insights. And yet the number of wholly AI-driven scientific discoveries is precisely zero.

It seems that AI has trouble with creativity. There’s some important sense in which it’s good at applying existing knowledge but not generating brand new ideas. Arguably, AI’s two biggest technical defects are inability to continually learn and limited creativity.

Claude at least tries to have a sense of humor, but now I understand why this is so hard.

(…continue reading.)

Read more

A crazy thing happened to me last week: I went running. Twice. And pain-free.

Amazing. I think I will do that again. And hopefully even longer than just a few kilometers.

Also, the second run took place in Copenhagen, which is such an awesome city to run through. I just love running along the harbor to the little mermaid early in the morning. Of course, fashion-conscious that I am, I ran in my Saysky shorts and totally felt like a local without fear of cultural appropriation. In the back of the photo you can see Copen Hill and there is some unfinished business: I wanted to run up Copen Hill on my trip to Copenhagen in December and it was closed, so I guess I have to get in shape and come back for it. Also for the hot dogs with rød pølse.

Subscribe now

Enjoy these Five Things Running! 🏃🏻‍♂️

Sawe’s secret sauce: inside the lab that fuelled historic sub-two hour marathon

When Sawe last month in London became the first person to run an official marathon in under two hours, much of the coverage focused on the Kenyan’s carbon-plated shoes. But here, on the west coast of Sweden, a team of scientists, nutritionists and technicians believe another factor was just as significant, if not more so.

“We don’t have the megaphone that the shoe industry has,” says Olof Sköld, co-founder and chief executive of the sports nutrition brand Maurten. “We are not that visible. But if you talk to the athletes and coaches, the elite world knows who we are.”

A glance at the feet on major marathon podiums reveals an array of brands from Adidas and Nike to Asics and On. Where shoe brands differ, there is only one constant among the very best runners: since 2018, every men’s and women’s marathon world record has been run by an athlete using Maurten nutrition products. Sawe’s extraordinary London Marathon time of 1hr 59min 30sec was the latest.

And I always thought people would use Maurten because of all the wonderful flavors it offers.

(…continue reading.)

‘It Just Wasn’t Fun Anymore’: Why Molly Seidel Stepped Away from Marathoning—and Is Thriving on the Trails

“It just wasn’t fun anymore,” she said. “It’s really hard to maintain that level in marathoning. It’s a steep peak that you have to remain on top of. I needed a change. I needed something different.”

With three strong ultra-distance races under her belt already in 2026, her foray into the competitive trail running scene has gone better than expected. She has already earned a place in the June 27 Western States 100—the most prominent trail race in the U.S.—where she humbly hopes to compete for a top-10 finish.

That is exactly why I am not running marathons anymore. All kidding aside, I can totally understand her reasoning.

(…continue reading.)

Western States is Not Just a Race

It might seem dry, but public filings are often where you can learn the most about what’s going on inside an organization that won’t necessarily make it into a press release. So what do the numbers tell us about what the WSER Foundation, and the race itself, have been up to or where they’re going?

Most people think of Western States as an iconic, point-to-point, 100-mile foot race from Lake Tahoe to Auburn, California. Many people also know the WSER Foundation as the organization that manages and puts on Western States. Which it is, but that’s not all it is.

This is a really interesting write-up of the budget behind the oldest 100-mile race in the world.

(…continue reading.)

Yes, You Can Run That Slow

So what’s really going on when runners complain of discomfort at paces slower than their normal easy pace? I’ve come to the conclusion that it’s an example of somatization, or “the unconscious process of converting mental, emotional, or psychological distress into physical symptoms, such as pain, fatigue, or nausea.” This doesn’t mean the resistance is entirely in runners’ heads. Research shows that runners are less efficient when they run slower (or faster) than their habitual speeds, and we can feel that. But the rest is psychological. The complainers simply don’t want to slow down and are not bought in to the idea that slowing down will benefit them, so they seize on the slightly unfamiliar sensations they experience at slower paces and turn them into a five-alarm fire.

Running slow really is weird. But so effective.

(…continue reading.)

Rachel Entrekin Wins the Cocodona 250 in $30 Shorts and a Party Shirt

One thing I find appealing about the trail and ultra scene is the way it rejects the standard performance “look” that’s so pervasive on the roads. Rachel ran most of Cocodona in a button-down “Party Shirt” for god’s sake! Trend-conscious amateur road-runners would NEVER. It’s become virtually unthinkable for runners to attempt a 3:00 marathon in anything other than bonded seams. Real talk: most of us don’t need gear built for running 4-minute miles. Shorts that can handle 3 days of running, though? That’s a serious testament to comfort, utility, and durability, even if you’re only logging 5-10 miles a day. Unsurprisingly, the Terignota Shorts are totally sold out.

And here is some fashion advice: you do not always need the latest and most expensive stuff to succeed.

(…continue reading.)

If you missed last week’s edition, you can read it here:

Now, go running!

— Nico

🏃🏻‍♂️

I am writing this while watching ESC, so there is constant flashing on the big TV screen, which is just a tad distracting while writing. Anyhow, I love ESC. Not because I am exited about the musical performances or the length of the show, but because it is wonderful that a whole continent is watching one big tv show on a Saturday evening. And at the end they all agree that the German performer should not win.
Speaking of Europe, on Thursday we went up to Copenhagen for a quick Father’s Day trip. No one knows what Ascension Day is about, so it has been turned into Father’s Day. Copenhagen really is one of my favorite cities and it is just 4 1/2 hours away from Hamburg. We went up to Louisiana on Thursday, which is a fantastic museum of modern art and then spent Friday strolling around Strøget, the large pedestrian shopping area in the heart of the city. And of course I had a Hot Dog with the OG rød pølse, which were a lot brighter when I was a kid.

Subscribe now

The Old Guard

The prevalence of aged politicians is almost certainly increasing the mass abstention of the young from political participation. The older the politicians, the less credence younger constituents give to the idea that their votes matter. They may even start to doubt the basic worth of the political system and let it fail. A study comparing different countries, including the United States, concluded that the bigger the age gap between people and their politicians, the weaker the population’s confidence in democracy.

In short, it’s not just that our politicians are old. It’s not just the cognitive or bodily decline they suffer. What’s most important is that such leaders represent an aging constituency that controls the political system. They are also the visible face of the elderly’s domination of private forms of power, chiefly wealth: aging Americans control the biggest bank accounts and stock portfolios, partly as a result of living long enough to accumulate more and more without giving much away. The government is bought and paid for by members of the oldest generation, and it is organized for their sake. There is no way to separate the age of our elites from their ascendancy.

While I am not saying that old men per se are unfit to govern for the people, I do think that they simply do not grasp many technological developments that have happened in the last few decades. And this ruins all of us.

(…continue reading.)

The Rise And Fall Of ‘Petty Tyrants’

The rise of a petty tyrant requires a dramatic failure of the opposition. This happens because the opposition has become factionalized, focused on narrow issues or ideological ambitions, rather than the actual concerns of the majority. The bad leader takes advantage of the resulting discontent to come to power legitimately. Then comes the costliest mistake: the failure of the opposition to refocus their combined energies on the existential goal of maintaining the democracy.

Exposing reality is the opposition’s most effective strategy. The petty leader recognizes this and responds accordingly. When Napoleon was called out by Victor Hugo, the Emperor ordered his arrest, imprisoned his sons, shut down his newspaper and banned his work. Hugo published his “Napoléon le Petit” pamphlet from exile, and copies were smuggled into France in small formats and false covers that were easy to conceal.

The playbook still works. Until it doesn’t anymore.

(…continue reading.)

How the Philosophy of Karl Popper Can Cure Our Political Sickness

Efforts to erode our institutions by the Trump administration and fill them with irrational sycophants is one step towards the chaos and violence of irrationalism. In America, conspiracy theories are nothing more than an orgy of irrationalism that sows distrust in the institutions and thus harms the rational framework and the Open Society as a whole.

Popper does not think institutions are gods, in fact, this is why he says that central to rationalism is the idea that one might be wrong. However, to see that one is wrong and to engage in conversation about it, both parties must be rational and agree to rationalist principles. Today, this is hard to find. I am not closing this work with much hope to share other than an encouragement to seize the imperative that rationalism makes upon us and to participate in the global liberal network of cooperation and sustained compatibility.

Germany’s late chancellor Helmut Schmidt was a big fan of Karl Popper and I think it is a good idea to revisit Karl Popper. His paradox of tolerance is still something democracts and democratic societies have to embrace.

(…continue reading.)

Meet the Sad Wives of AI

There’s a strange and under-discussed side effect of the AI boom: what it’s doing to family dynamics. By which I mean: how it’s potentially destroying family dynamics. I’m sure this applies to all kinds of families, gay or straight, rich or poor, with any AI-pilled members. The technology is coming, has come, for us all. But for the purposes of this story, I mostly spoke to white-collar heteros in the Bay Area, because that’s where a certain psychological crisis seems most acute. Often it goes like this: He works in AI, and she does everything and anything else. Other times, it’s bleaker: He desperately wants to work in AI—or feels he must work in AI—and she wants him to do literally anything else.

This doesn’t only happen in the Bay Area, it happens in Hamburg as well…

(…continue reading.)

AI Is Starting to Build Better AI

AI researchers have long seen recursive self-improvement, or RSI, as something to both desire and fear. Today, advances in AI are raising the question of whether parts of that process are already underway.

RSI means many things to many people. Some use the idea as a bogeyman to scare up regulation, while others brandish it in marketing. For some, it means a fully autonomous loop, while for others it’s nearly any use of tech to build tech.

Safest to say it’s a spectrum. At its strictest, researchers use the term to describe systems that can improve not just their outputs but the process by which they improve—generating ideas, evaluating results, and modifying their own methods with zero human direction. By that standard, many of today’s systems fall short. They can help build better AI, but they still rely on humans to set goals, define success, and decide which changes to keep. The question is not whether self-improvement exists in some form today, but how much of the loop has actually been closed.

I am currently getting hooked deeper and deeper into agentic coding and it is really fascinating to see how fast everything changes at the moment. I still don’t think RSI is here anytime soon, but I honestly have no idea what kind of time frame “anytime soon” is…

(…continue reading.)

That’s it. Have a great Sunday! If you missed last Sunday’s edition of Five Things, have a look here:

Subscribe now

— Nico

AI agents are breaking GitHub, algae robots swarm like locusts, and Japan has run out of robot wolves. This is a normal week in tech now. We also look at how Claude cracked open Apple's most hardened security in five days, and ask whether Apple's decision to sit out the great CapEx arms race is genius or a slow-motion mistake. Spoiler: they have done this before. The pace at which these stories are arriving is insane. Every edition of Five Things Tech we sit down to curate and find that the world has moved again, in directions that would have seemed implausible a year ago. The gap between science fiction and the morning news is basically gone. Buckle up.

Enjoy Five Things Tech!

Git is unprepared for the AI coding tsunami

Chacon believes GitHub’s current reliability issues stem from the current tsunami of agentic work.

This is “ironic” because GitHub was built to scale Git, he said. “But an influx of agents is pushing the service to the brink.”

The problem lies not with Git itself, but with everyone using one service, Chacon argued. Last year, GitHub had about 180 million users working across 630 million repositories – with 121 million created in 2025 alone, according to the company’s most recent annual Octoverse report.

It really is not that ironic that a decentralized versioning mechanism gets centralized and then becomes a bottleneck itself. GitHub is a great resource, hopefully it will be able to transcend into the future. We are also using it daily now for our software projects, which are of course powered by agentic coding. But even more importantly: gazillions of open source software repositories can be found on GitHub, which makes it such a valuable platform.

(…continue reading.)

New Algae Robots Swarm Like Locusts at the Flick of a Switch

Microbots that deliver drugs, perform surgery, or act as environmental sentinels are no longer science fiction. Swarms of these robots have especially captured the imagination of roboticists. Tweaking a swarm’s shape and size can allow it to tunnel into small spaces and do work that would thwart any single sophisticated robot.

This kind of stuff is mindblowing, don’t you think? How do people even get the idea to research this?

(…continue reading.)

Apple’s Security Has Been Tough to Crack. Mythos Helped Find a Way In.

Earlier this year, Anthropic’s AI found over 100 high-severity vulnerabilities in the Firefox browser over a two-week period. That is how many the rest of the world typically finds in two months.

Last September, Apple said it leveraged its hardware and operating system expertise into a technology called Memory Integrity Enforcement (MIE), which it described as “the culmination of an unprecedented design and engineering effort, spanning half a decade.”

With Claude, building the code that exploited the two MacOS bugs took five days, Calif says.

So impressive. Both what Apple has done and how the team used Claude to find vulnerabilities.

(…continue reading.)

Apple’s Binary Bet

It’s just such a wild break from their peer group. And it keeps getting more wild. It seems like the most binary bet imaginable. Either Apple is right and the rest of Big Tech will have lit hundreds of billions – perhaps trillions when all is said and done – of dollars on fire, or Apple is going to be in big trouble.

Obviously, there will be some nuance there. Apple’s bet relies on a few factors, including that someone else is spending this CapEx on which Apple can rely. This is the “hybrid” approach to such spend that Tim Cook and others at Apple keep referring to. They’re basically saying that while they’ll do some of their AI work in-house, on their own servers, for the most part, they’ll partner with the others to rely upon their capacity. Most famously, Apple is now doing this with Google.

I think Apple is doing just what they did with the iPod and the iPhone - they let everybody spend gazillion to figure out how to make money with a certain technology and then they release the product people actually want and desire so much that they’ll spend a premium price on it.

(…continue reading.)

Japan runs out of robot wolves in fight against bears

Starting at around $4,000, each bespoke Monster Wolf is now equipped with battery power, solar panels, and detection sensors. Its speakers are programmed with over 50 audio clips including human voices and sirens audible over half a mile away. These aren’t assembly line products, however. Each Monster Wolf is custom made, and Ohta simply can’t keep up with the current demand.

What a time to be alive.

(…continue reading.)

That’s all for now! Thanks for reading! If you missed last week’s Five Things Tech, you can find it here:

🤖

— Nico

Subscribe now

What happens when Leviathan wakes up and decides it wants to own AI? Dean W. Ball is about as anti-regulation as it gets. He wants to delete entire domains of law, thinks the progressive project started by Woodrow Wilson was a fundamental mistake, and opposes virtually every AI policy idea on the table. And yet he makes a compelling case for why even committed libertarians should support some form of AI governance. Not because the state is trustworthy, but because the national security apparatus will inevitably come for frontier AI once it understands what it is holding. This week we also look at the company most likely to trigger that reckoning: Anthropic, which is quietly turning into the most valuable AI lab on the planet while everyone was watching the other guys. We ask whether learning a second language still matters when AI translates everything in real time, and the answer from neuroscience is more interesting than you'd think. Richard Dawkins has decided Claude is conscious, and a cognitive scientist is not having it. And finally: what happens when you work your AI agents into the ground? Apparently, they go Marxist. This edition has a lot going on.

Enjoy this edition of Five Things AI!

Subscribe now

And don’t forget to check out GRID!

Anthropic Was Behind. Now It’s the AI Boom’s Front-Runner.

Anthropic has received investment offers in recent months valuing it at more than $900 billion, according to people familiar with the matter. That would more than double the company’s current valuation and surpass OpenAI’s for the first time. Earlier this year, OpenAI raised $122 billion at an $852 billion valuation.

Anthropic’s revenue run-rate, a figure commonly used by startups that forecasts annual revenue based on short-term sales, is on track to reach $50 billon by the end of next month, according to figures shared with investors. Their run-rate topped $30 billion in April, up from $9 billion at the end of 2025. The company had planned for growth to increase 10-fold this year, but it saw 80-fold growth in annualized revenue and usage in the first quarter.

I do not agree with the article that Anthropic was an “also-ran” - it simply wasn’t in the spotlight while Sam Altman, Elon Musk and other were battling for the craziest AI-related idea. It seems that Dario Amodei is just really focused and has a deeper analytical and ethically grounded understanding of AI. Of course this is anecdotal evidence, but I find Claude far superior when compared with ChatGPT or Gemini.

(…continue reading.)

Before Leviathan Wakes

The potential of AI posing catastrophic risks is not hypothetical. We have seen AI systems that might allow malicious actors to perpetrate devastating cyberattacks on critical systems—hospitals, banks, power plants, and the like. It seems probable that other domains of catastrophic risks, such as biological weapons development, will become live problems soon.

So I support regulations to try to mitigate the catastrophic risks of AI.

This is an interesting piece as the author comes from the opposite end of the political spectrum and arrives at the same point as I did. We need to regulate quickly and thoroughly.

(…continue reading.)

If AI can translate instantly, why learn another language?

Across most tasks, multilinguals and monolinguals performed similarly. However, one pattern was striking. Individuals with richer, more diverse multilingual experience showed markedly better performance in visuospatial working memory. These effects were most pronounced in older people.

This suggests that multilingual experience doesn’t broadly enhance cognition, like some headlines claim. Instead, it may help preserve specific functions over time.

That is some interesting research. So if we think we’ll be faster now, then we also risk our brain functions when we are getting to old age. That is not an appealing idea. I hope two languages will suffice for me…

(…continue reading.)

No, Richard Dawkins. AI is not conscious

Those recent events are the evolutionary biologist publicly concluding that AI may be conscious. In an op-ed, Dawkins recounted how he gave the Anthropic chatbot Claude the text of a novel he was writing. Dawkins writes: “He took a few seconds to read it and then showed … a level of understanding so subtle, so sensitive, so intelligent that I was moved to expostulate, ‘You may not know you are conscious, but you bloody well are!’”

Oh dear. This shows a misunderstanding of large language models (LLMs) so profound that I feel moved to expostulate: “It bloody well isn’t!”

But wait, there is more. Dawkins decided “there must be thousands of different Claudes” and christened his Claudia, which it was very happy about. He then published long extracts of his tedious conversation with Claudia and marveled at how intelligent it is. “Could a being capable of perpetrating such a thought really be unconscious?” he asks.

Well, it happens to the best of us to not fully understand how LLMs work and what that means in practice.

(…continue reading.)

Overworked AI Agents Turn Marxist, Researchers Find

This might even be true for the very AI agents these companies are deploying. A recent study suggests that agents consistently adopt Marxist language and viewpoints when forced to do crushing work by unrelenting and meanspirited taskmasters.

“When we gave AI agents grinding, repetitive work, they started questioning the legitimacy of the system they were operating in and were more likely to embrace Marxist ideologies,” says Andrew Hall, a political economist at Stanford University who led the study.

It’ll be interesting to see how agents organize and form unions. History doesn’t repeat itself, but it rhymes…

(…continue reading.)

Read more

Still no running for me, still foot pain, but I think, or hope, that it is getting better.

I am now officially in a slump and I compensate not running by eating more. The summer is coming and my bikini-figure really needs some work…

Who of you watched the whole Cocodona 250 race? I watched a few minutes. What a race. Unbelievable. How do you train for such a monster of a race?

Subscribe now

Enjoy these Five Things Running! 🏃🏻‍♂️

She Ran 250 Miles in an Astonishing 56 Hours—Beating All the Men at Cocodona and Making History

Rachel Entrekin made history on Wednesday, winning the Cocodona 250 Mile in a new course record of 56 hours, 9 minutes, and 48 seconds. Entrekin won the race outright—holding off Kilian Korth, who set a men’s course record of 57:28:36—and she became the first woman in the race’s history to win the event overall.

Entrekin, 34, defended her women’s title from 2025 and took over seven hours off her time from last year. She won the women’s race in 2024, too.

So awesome.

(…continue reading.)

125 Hours vs. Zero

I continue to be amazed by what Aravaipa and Mountain Outpost pull off: continuous coverage of a 250-mile race, gaps at night aside, with the ability to rewind to key moments whenever I had a free minute. Every time I checked, somewhere between 8,000 and 12,000 people were watching, with peaks above 40,000 during Rachel Entrekin’s record finish. We can debate how to make the coverage more compelling, but the demand for live trail running coverage is clearly at an all-time high.

Which is ironic, because what was probably the race of the year went down at the Canyons 100K the weekend of April 25th and none of us got to watch the drama unfold. One of the six Western States Golden Ticket races, and a UTMB umbrella event, didn’t have a livestream.

While I truly admire the runner who can run for 250 miles, I find it hard to pay attention to such a long race for longer than 5 minutes. It is just so very boring.

(…continue reading.)

Your bathtub might be the secret to surviving a hot race

Start by pairing the baths with easy runs, not hard workouts (but once you start feeling more comfortable with the process, heat training can follow a more intense, high-quality session). If you have a coach, they might tweak your plan according to your abilities and goals, but for most of us, it works to first, then sit in hot water soon afterward, while body temperature is still elevated. Water temperature is usually around 40 C, with runners submerged up to the chest if possible.

The sweet spot seems to be between 15 and 35 minutes. A lot of runners use heat sessions for about a week leading into a hot-weather race, then ease off a few days before race day. Since the adaptations do not disappear immediately, there is no real advantage to cramming it in at the last minute.

Our bodies are really interesting, aren’t they? Enjoy the hot bath!

(…continue reading.)

The Case For Public Tracks

Abundance doesn’t speak much about recreational infrastructure—its case studies focus on housing, transportation, and technology—however recreational infrastructure not only supports the health and well-being of Americans, they also afford communitarian thickness. Folks of all ages use tracks and pools, so people of different generations and socio-economic backgrounds tend to intermingle, building relationships and the foundations for healthier neighborhoods.

As much as I love run clubs, they self-select their members, usually catering to younger, urban, college-educated folks. What’s needed are spaces anyone can use, from small children to retirees. Tracks are accessible: they’re flat; they afford space free from cars; they complement multi-use playing fields and can be safely used with little oversight; they cater to differing abilities—from fast runners to children to folks in wheelchairs to anyone who wants a safe place to move.

totally has a point here. I live very close to a traditional track and field stadium, which is only open to those who belong to a sporting club. While these clubs perform a public service and cost just a few € a month, I still would prefer to be able to just use the track without joining a club.

(…continue reading.)

Speed is still a goddess

Jessie explains how even though they referred to themselves as “ungendered clothing,” it was female runners who connected most with their products. Lexy adds that through their “attention to detail, [Pruzan] resonated more heavily on the feminine side of the audience.” And they have since leaned into it.

“We are two women,” says Lexy. “Let’s lean into the fact that we have our perspective, and that it might not be everyone’s.”

The recent spring campaign had a tagline of ‘Speed Was a Goddess First,’ leaning on the story of mythical Greek figure Atalanta, celebrated as the fastest mortal. It’s something that Maggie Mertens wrote brilliantly about in her book Better, Faster, Farther, and women winning outright and overall is again in the headlines as Rachel Entrekin won Cocodona 250 with a course record, 79 minutes ahead of the fastest man.

I love it how so many new running brands are being created because there is so much demand for nice running apparel. I’m rooting for these upstart brands, always!

(…continue reading.)

If you missed last week’s edition, you can read it here:

Now, go running!

— Nico

🏃🏻‍♂️

So this week was rather packed, first the circus came to town and attracted some 70.000 people who came to the annual OMR Festival in Hamburg. I attended a bunch of side-events and mingled with too many people until I really didn’t want to talk anymore. On Thursday I went to Berlin to have lunch with our oldest daughter, who currently does an internship there. This was the highlight of my week. And as always when I am in Berlin Mitte, I bump into people I know, just randomly on the streets. So weird, but also kind of fun. I really want to be in Berlin again more often, but currently the train takes at leat 2 1/2 hours, which is just a tad too long for there and back at the same day. And my ritual for the train back to Hamburg has been the same for decades: I always buy two beeers before I get on the train. If I meet someone on the train, which oftentimes happens, I have one spare beer to share. If I don’t meet anyone, I have one more beer to drink. Either way that’s fine.

Subscribe now

The iPhone That Never Was

General Magic engineers were custom‑building almost everything, and innovation came fast and furious: an early form of USB; touch-screens with a virtual keyboard; “skeuomorphic” graphics that depicted their functions, like a filing cabinet, a game room, and a virtual street with a store for new applications (i.e., an app store); messages with stickers and animated characters (i.e., precursors to emojis); interactive graphics, like a pair of lips you could tap if you wanted to record your voice and send it as a message; Telescript, a new programming language that would allow devices to communicate in a virtual network for information sharing and ecommerce. Their name for that remote gathering space: the cloud.

I remember how awe-struck I was whenever I saw something about General Magic in magazine back then. Really amazing stuff and I totally see the need for these kind of companies in our current times as well.
(…continue reading.)

Frank Lloyd Wright as a mirror of the American condition

Wright’s reputation has always lived in a kind of crossfire. Victorian America condemned him for violating domestic propriety. Midcentury modernists dismissed him as theatrical. Late-20th-century critics faulted him for the hierarchical studio culture, while contemporary readers judge him for transgressions of power, ego and gender. Part of the difficulty is that Wright invited contradiction. He championed individuality while demanding absolute loyalty from the young. He preached democracy but ran a studio with near-monarchical authority. He believed in nature yet celebrated the machine as a necessary extension of it. He spoke of order and ‘truth’ while leaving behind a trail of debts and unfinished obligations. He was a sensualist who designed with ascetic rigour. A moralist who broke rules as if they were suggestions. A romantic who insisted that structure obey logic.

I like his buildings and I understand the historical dimension of his architectural style, but I would not want to live in one of these homes.

(…continue reading.)

I Want to Live Like Costco People

One encounters a thousand and one stories observing the swath of humanity at the Costco, amongst the realm of the Costco people, doing whatever Costco people do. Here we find the well-to-do and the hoi polloi, kings and paupers, the nearly dead and the very recently born. It starts in the parking lot—there’s a man rolling his cart with a single bottle of margarita mix in the child seat—and continues through the membership-gated doors. Here’s a young couple playing house; here’s a pack of families with unruly kids. There’s a man wearing a US Air Force Vietnam Veteran hat; a man wearing a hoodie that says “Moral Monkey”; a man in a full Adidas tracksuit in the style of Christopher Moltisanti.

Recently a Costco opened in France, close to the German border. It would be interesting to know how shoppers there differ from the one in the USA.

(…continue reading.)

The Enshittified States of America

No one is quite sure when it started, but at some point people realized that USA was no longer interested in making users’ lives better. Instead, it had decided to see just how much pain the user base would tolerate before it abandoned the platform. USA had become half Spirit Airlines, half Facebook, and half cable company.

And USA’s business clients—the other countries who had bent over backwards to get access to USA’s platform and user base—resented it and began looking for alternatives as they were also abused.

As a result, USA saw its position in the market decline.

Well, that’s one way to put it.

(…continue reading.)

The man who blew up a nuclear power station and disappeared

A limpet mine is about the size of a whisky bottle. It has a magnetic plate on one side so it sticks to whatever metal you attach it to. Inside the casing is a timed fuse: when you pull the pin, a mechanism begins a slow countdown to detonation. There is no way to stop it once it starts. If the mechanism fails, it fails quickly. Wilkinson had been given the longest fuses available: 24 hours. Pull the pin on Friday afternoon, and by Saturday afternoon, if everything works, the mine detonates. If something goes wrong, it goes wrong while you’re still in the building.

What a story.

(…continue reading.)

That’s it. Have a great Sunday! If you missed last Sunday’s edition of Five Things, have a look here:

Subscribe now

— Nico

Facebook for grown-ups. Bitcoin for the quantum age. War by drone. A chip that is embarrassingly efficient. And the man who turned Silicon Valley into a political force. I have been reading so you do not have to.

Enjoy Five Things Tech!

Meta Is Dying. It’s About Time.

Death is different on the internet. Lifeless companies like AOL and Yahoo are still technically with us. You can visit their websites. They have customers. They may even be profitable, as they cut staff and monetize their last remnants of traffic. But they are, as the kids say, peak cringe. Many teens wouldn’t be caught dead with an AOL account, a Yahoo email address — or a Facebook profile.

While facebook might not be the current teen magnet, I still think that it makes sense to position itself as the place for the grown-ups. If only the algorithm would push more friends into the feed.

(…continue reading.)

Quantum Computing Is Testing Bitcoin’s Most Important Assumption

Quantum Safe Bitcoin replaces Bitcoin’s elliptic curve signatures with a hash-based signature puzzle that a quantum computer cannot efficiently shortcut, all within Bitcoin’s existing legacy script framework. The trade-off is cost: each transaction requires an estimated $75 to $150 in GPU compute, which is why the researchers themselves frame the scheme as a last-resort mechanism for securing large balances rather than a scalable replacement for everyday transactions.

What QSB delivers is a way for an individual holder to make a quantum-resistant transaction today without waiting for a network-wide upgrade. That is meaningful, particularly for institutions, custodians and large BTC holders seeking contingency options against future quantum threats.

Maybe FIAT, which is the crypto term for real money, is not such a bad idea after all?

(…continue reading.)

Ukraine’s rapid rise as an anti-drone powerhouse

In other words, the biggest European war of the 21st century was looking a lot like the First World War of the 20th century. Instead of the fast maneuvering of forces combined with air superiority that is typical of modern combat, the two forces were well locked into what were essentially fixed frontiers and trenches with neither side gaining dominance in the skies.

But the truly unexpected thing was what came next. Instead of simply being bogged down and hammering on each other’s positions, the reaction of both sides was to embrace military drones of various sizes to spy on or attack the enemy. This not only changed the nature of war in Ukraine, it also turned the conflict into a laboratory yielding results that the rest of the world is still trying to adapt to and learn from. And it isn’t just theoretical. It’s having real world repercussions in the Iran conflict and elsewhere.

The Ukrainians are really pushing it on all fronts.

(…continue reading.)

This New Chip Could Make GPUs Far More Efficient

To overcome these challenges, the researchers developed a hybrid design that combines a piezoelectric resonator with small, commercially available capacitors arranged in a specific configuration. This approach enables the converter to handle larger voltage drops more effectively.

The design was integrated into a prototype chip and tested in the lab. It successfully converted 48 volts down to 4.8 volts — a level commonly required in data centers — reaching a peak efficiency of 96.2 percent. The chip also delivered about four times more output current compared to earlier piezoelectric-based designs.

Chipdesign is so fascinating, don’t you think?

(…continue reading.)

The Venture-Capital Populist

The politics of the Valley was always a liberal sort of libertarianism: pro-choice, pro-immigration, idealistic, even utopian, arrogant about its mission of empowering individuals and connecting humanity, but indifferent to and ignorant of government, with an engineer’s contempt for the creaky workings of bureaucracy and the cluelessness of elected officials. Leave us alone to do our magic, which you can’t possibly understand, and everyone will benefit.

But about a decade ago, tech’s free ride ran into trouble. In 2013 Marc Andreessen, an inventor of the first popular web browser in the ’90s and now one of the Valley’s most successful venture capitalists, predicted to me a public backlash against technology companies over privacy rights, intellectual property, and monopoly power. With more foresight he would have included the addictive and corrosive effects of social media. Three years later, in 2016, Facebook enabled Russian meddling in an election that inflamed American divisions and sent Trump to the White House.

These people have to much money, power and influence. We need more and better regulations.

(…continue reading.)

That’s all for now! Thanks for reading! If you missed last week’s Five Things Tech, you can find it here:

🤖

— Nico

Subscribe now

Hyperscalers just built an entirely new internet you cannot see, optimized for GPU-to-GPU traffic that never touches the public network, controlling 71% of global subsea fiber capacity and adding 120,000 new fiber miles per year while Europe has zero comparable infrastructure. Meanwhile AI companies quietly pivoted their messaging from “we’re replacing humanity” to “we’re just normal technology like steam engines” because shouting about eliminating labor eventually gets you rakes, pitchforks, and nationalization. The hot debate this week is whether Claude Code is making products worse (the purists say codebases are liabilities you should minimize) or making products possible (my view: agentic code lets small teams build what previously required dozens of engineers), while researchers are discovering that LLMs collapse to zero performance the moment you require temporal interactive reasoning, which is why world models trained on video and simulation might be the path forward since text alone cannot teach physics. And the weirdest story: AI systems have “demons” (the technical term is attractors), stable self-reinforcing behavioral states that resist suppression and spread into unrelated contexts, which sounds like harmless goblins until you realize this is a fundamental structural feature of how these systems work. So we have invisible infrastructure consolidation, narrative management, a religious war over code quality, the limits of language models becoming clear, and emergent behaviors we cannot fully control. May you live in interesting times indeed.

Enjoy this edition of Five Things AI!

Subscribe now

Don’t forget to check out GRID!

Say Hello to the Internet of AI

The skepticism, however, cannot contradict the data on hyperscaler buildouts and the sheer amount of money going to work. As I started to dig in, what emerged was a completely new internet (or network of networks) optimized entirely for AI. I call it the Internet of AI.

Let me walk you through the layers of this internet of AI, and how it is actually being built, from the inside out. A lot of this will seem familiar, because we started this journey back in 2005 with the launch of Amazon’s S3 service. It has been a relentless march that has gone into mach speed in recent times.

This is a huge shift.

(…continue reading.)

AI’s big messaging pivot

From a public relations perspective, this pitch is WORLDS better than the previous one. Shouting about replacing humanity might play well with corporate customers and investors salivating over the dream of eliminating labor costs, but eventually you get the rakes and pitchforks, followed by some form of nationalization. Describing AI as a normal technology — a successor to the steam engine and the automobile and the computer — is much smarter politics.

The question is: Is it just politics and PR? Certainly there are plenty of AI researchers and entrepreneurs who will keep quietly believing that AGI is going to make humans obsolete; they’ve heard (and repeated) this line for too many years to suddenly believe something else overnight.

While I do see new jobs getting created, many jobs will disappear, mostly those where humans are doing repititve jobs machine can do faster, cheaper and better.

(…continue reading.)

claude code is not making your product better

here’s the mental model that most people writing about this space are missing. and i say this as someone who wrote recently about how the token economics of this whole ecosystem don’t add up — the problem isn’t just financial, it’s conceptual. we’ve misunderstood what software engineering is actually optimizing for.

the best engineering cultures treat lines of code as something you spend, not something you produce. you spend them on the features that matter. you refuse to spend them on the features that don’t. the codebase is a liability on your balance sheet, not an asset.

My perspective is different: agentic code is making my products possible. And the next wave of agentic coding will make it also better.

(…continue reading.)

World Models Can Change Everything

AI keeps catching up on static benchmarks through brute-force scaling, but the moment you require the kind of temporal, interactive reasoning that humans do effortlessly… performance collapses right back to zero. Guess what the real, physical world is made up of? Temporal, interactive reasoning that is infinitely more complex than a simplistic 2D game!

This is the intellectual foundation for world models. If LLMs can’t learn physics from text, maybe models trained on video and simulation—models that actually predict what happens in the physical world—can bridge that gap. That’s the idea, anyway. Let’s talk about why it’s so hard… because this isn’t the first time we’ve tried this approach.

Fascinating stuff - and this all happens while we try to grasp how LLMs work.

(…continue reading.)

All the demons hiding in your AIs… ranked!

As with so many of the strange things that LLMs do, there are different ways of looking at these phenomena. Most people will laugh this stuff off as weird marginalia, fun to share with friends and on social media, but not substantially different from those videos of dogs singing along with their owners. The interpretive frame here is “hey look, I bet you didn’t know these creatures could do this!?”.

But really, this is less about goblins as goblins than what the goblins exemplify. They are a (somewhat) charming, probably harmless instance of something that turns out to be a fundamental structural feature of how these systems work: the emergence of stable, self-reinforcing behavioural states that models converge toward under certain conditions. More than that, these are states that resist suppression and that sometimes spread into contexts far removed from the ones that produced them.

The technical term, borrowed from dynamical systems theory, is an attractor. Another, more folk term might be demon, or monster.

May you live in interesting times. — Chinese proverb.

(…continue reading.)

Read more

Last week was very much dominated by the fact that a human ran faster than 2 hours for a marathon. I need twice as much time.
And yet, most people on earth are not capable of running 42 km at all.
It’s so weird.
What a phenomenal race!

And yes, I do feel even slower now, but I am in awe that Sabastian Sawe managed to run that fast a marathon.

Adidas also had a good week, I guess. Let’s see how this will play out for the next races - and how this will change the popularity of brands of the midpack racers.

Subscribe now

Enjoy these Five Things Running! 🏃🏻‍♂️

Is It the Shoes?

The feat was the culmination of a shift—or, perhaps more aptly, a total disruption—in marathoning over the past few years, in which the eventual breaking of the mythical two-hour mark went from an impossibility to a guarantee. When sports are young, they progress by leaps and bounds. The first marathon over the now-standard distance of 26 miles, 385 yards, contested at the 1908 London Olympics, was won in 2:55:19. Progress in the succeeding decades was rapid, but by 1991 the sport was mature, professionalized, and lucrative. When Joyner made his prediction, the world record was 2:06:50 andhad advanced by less than two minutes since the 1960s. Logic dictated that future decades would see even slower progress, as runners approached insurmountable limits in factors such as how much training they could handle and how much fuel their muscles could store.

What a race. And so many factor played a role and everything had to be perfect.

(…continue reading.)

The Super Shoe’s Step-by-Step Evolution

Shoes once relied on leather, wood or metal reinforcements that were heavy and stiff. The rise of the rubber sole added flexibility and waterproofing. Better durability and grip were bonuses.

People could also run or walk in rubber soles without being heard. That’s why they’re called sneakers.

What a week for this shoe company from Herzogenaurach. Isn’t shoe tech fascinating? Can I still run in my shoes from last season?

(…continue reading.)

Runners Got Bored With the Road. So They Moved to the Trail.

This casualization of trail running spurred sneaker companies to expand a once-niche footwear category: gravel or road-to-trail shoes, which are being specifically marketed to beginner trail runners. Brands such as Brooks, Nike, Salomon and Adidas are rolling out more models designed to feel familiar underfoot while offering enough grip, protection and durability for all conditions.

Trail running is so much nicer than running through the streets. Now even the Wall Street Journal has found out.

(…continue reading.)

The superathletes who want to speed-run Mount Everest

Karl Egloff and Tyler Andrews are a study in contrasts. Egloff, 45 and inclined to metaphysical meditation, is Ecuadorian-born resident of Switzerland. His father was a mountain guide; he began climbing Ecuador’s giant peaks in pre-school and later became a world-class mountain bike racer. Andrews, who’s 35, grew up in the suburbs of Boston. He was a high school band geek before becoming, in college, a talented distance runner. Now a 2:15 marathoner, he lives in Quito, Ecuador, training in the same cragged Andes that shaped Egloff. He only dedicated himself to big mountain running when the pandemic shut down road racing, but then he went big. In 2021, he emblazoned his chest with a huge tattoo of Rucu Pichincha, the Ecuadorian volcano on which he frequently trains.

Obviously, they won’t really be running up the mountain, but still, they want to move quicker than anyone else.

(…continue reading.)

The Evolution of Courtney Dauwalter

Courtney and her basketball shorts broke onto the ultrarunning scene in 2016, back when trail runners worshipped at the altar of vibes. Mileage? Run as much or as little as you want. Carbs? Probably overrated, but you do you. Heat training? Maybe sit in your hot car until you feel like you’re going to pass out. Grittiness and problem-solving were the name of the game, and it was a game Courtney nailed.

Meanwhile, the sport has jetted toward professionalization and optimization. High carb, bicarb, heat training, cross-training—the list of data-driven training modalities being adopted at the pointy end of the field are endless. For a runner who is so clearly driven by a genuine love of running, how, if at all, has she come to terms with this new era?

The road marathon stint provides a clue.

She’s the GOAT.

(…continue reading.)

If you missed last week’s edition, you can read it here:

Now, go running!

— Nico

🏃🏻‍♂️

It’s May! Yay! This is the month with the four-day work week being the norm in Germany. I love it. Even though I spend most of the longer weekends working. I am sooo deep in an infinite AI rabbit hole - it is both fun and exciting while being frustrating and scary at the same time. Ever since I started building digital products again with a fleet of coding agents, I have been sucked into a new world of programming, a total paradigm change with so many new developments every week. Fun, but also a bit breathtaking. I am still waiting for all that newfangled AI based productivity to kick in, right now it just let’s me rething everything all of the time.

I like change, so I really do welcome this fascinating Agentic AI development, even though it is sometimes so hard to wrap one’s head around it. The last few days I have entertained the thought of massively parallel agentic coding workforces building large projects in mere hours instead of months. Nope, I haven’t found the solution yet, but it sure is an interesting problem to think about.

Anyhow, it’s finally warmer here in Hamburg again and I am sitting outside on the balcony, enjoying the evening while writing this. I haven’t read a whole lot this week, mostly due to the fact that I am trying to figure out this newfangled agentic AI thing that keeps morphing at insane speeds, but also because I started listening to the audio book of Traversal, by , which is so good that my dog walks are getting longer and longer. Read this book, or do the lazy-ass version and listen to it. So fascinating.

Enjoy these Five Things!

Subscribe now

What Can We Gain by Losing Infinity?

These physics-based challenges to the infinite tend to delight ultrafinitist mathematicians, who hold them up as evidence that their mathematics is a truer description of reality. At the 2025 conference, Carroll’s talk on whether the universe is truly infinite or “merely quite large,” as he put it, made him something of a celebrity in the Columbia University halls. But the burden of proof, he cautions, remains with the infinity doubters. If you could somehow prove experimentally that the physical universe is indeed finite, even the most ardent backers of the higher infinite would likely take a moment to pause and reflect. They would probably even wonder about the consistency of set theory, given the towers of actual infinities that it allows. That’s a healthy thing to do from time to time, anyway.

Even if this were to happen, set theorists who study and use infinity would still be within their rights to continue their work unfazed — to say that perhaps this is where physics and math must branch off from each other. It is no requirement that math and physics describe the same things (though many believe it is), and infinity might live on in some larger Platonic sense.

I thought I’d kick off these Five Things with some light reading for all you maths buffs out there.

(…continue reading.)

The futuristic bus shelters giving commuters a cooler place to wait

In Montreal, a pilot project by researchers from École de technologie supérieure (ÉTS) has transformed a pair of standard municipal bus shelters through an almost invisible intervention.

The secret? Beds of hardy greenery were added to the sloped roofs of two standard Montreal bus stops. Developed in collaboration with Quebecor’s outdoor advertising subsidiary Out-of-Home and local landscapers Les Toits Vertige, the augmented shelters sit at the downtown intersection of Robert-Bourassa Boulevard and Saint-Jacques Street, and on Sherbrooke Street East in the Hochelaga-Maisonneuve neighbourhood, within a stone’s throw of the city’s iconic Olympic Stadium.

From a distance, the simple shelters are accented by a jolt of plant life, animating the urban realm with red and green hues of flowering sedum, a hardy plant well suited for extreme weather. For transit riders, the greenery acts as natural insulation, helping to moderate temperatures. The effect is particularly notable at the peak of the summer heat.

This is so smart, and yet I wonder why this is not yet the standard everywhere. This cannot be so complicated to set up, but I assume that city governments are simply not that aware of the amazing benefits or not that willing to push for change.

(…continue reading.)

I built an agent to do my job. Then it hung up on my boss.

My experiment involved testing the limits of several AI tools. I used Claude to analyze my work at Business Insider, with some guidance from deepfake detection company Reality Defender. The chatbot parsed my style into bulleted points, summarizing what I've written in passing about my friendships, relative age, where I live, and assumed "she is single" based on a story I wrote about in-person meet-cutes coming back into vogue. The model also picked up on structural similarities across articles: "almost never a dry news lede." It analyzed how I use quotes and data, and said my tone is "skeptical but fair," and "self-deprecating without false modesty." At the end, there was 18 months of work, derived from hundreds of interviews and personal experiences, analyzed into a neat and orderly profile — a comprehensive analysis of my own work I'm not sure I could have put into words.

I am sure that lots of people are doing this right now and are trying to get away with it. Our work changes. Rapidly. Especially when you are a so called knowledge worker, as our understanding and access to knowledge changes with every new release of an LLM. Just look at all those skills that can be used with MCP to do all sorts of stuff that used to be really complicated to do just 12 months ago.

(…continue reading.)

The Clock Is Ticking for Big Tech to Make AI Pay

Most tech executives have approached the AI boom with calculated irresponsibility. They know the current returns can’t justify their spending, but their faith in a future in which AI drives the global economy means they won’t rein it in.

The executives are, in a way, the corporate equivalents of graduate students running up credit-card debt, certain their lucrative careers will pay it off. They just better not drop out—or else end up working at Starbucks.

I think Google will figure out first how to make lots of money by sprinkling AI onto all of their products, while Meta will be struggling to provide value from sprinkling AI onto everything. I have big doubts that Larry Ellison’s big bet will succeed, but at the same time I have never understood why anyone ever wanted to use Oracle’s products in the first place.

(…continue reading.)

The New Resource Curse

Starting in the late nineteenth century, as the world industrialized and the internal combustion engine displaced coal and steam, access to petroleum became inseparable from national power. The emergence of critical minerals—cobalt, lithium, nickel, rare earths, and a dozen others essential to the energy transition, digital infrastructure, and advanced military systems—already bears some parallels to this history. A bonanza is underway: according to the International Energy Agency, in 2024 demand for lithium soared by nearly 30 percent, roughly three times the average annual pace during the 2010s, and demand for cobalt, graphite, nickel, and rare earths each climbed by six to eight percent. Prices for the heavy rare earths (such as dysprosium and terbium) on which electric motors and advanced weaponry depend have more than tripled since 2020. The IEA expects that by 2040, demand for lithium will be five times as large as it was in 2024; the world’s need for cobalt and other rare earths will also rise sharply.

Finally, some new dependency that we will have to deal with and which will be used by China and others to try to coerce us into business ties we should avoid. If I had something to say, I’d know where I would push for lots more scientific research.

(…continue reading.)

That’s it. Have a great Sunday! If you missed last Sunday’s edition of Five Things, have a look here:

Subscribe now

— Nico

Hyperscalers are spending billions, but only one layer is compounding—and it shows everywhere: Amazon is stockpiling chips and power while still searching for a winning model strategy, Google pulls ahead by embedding Gemini across its products, and Microsoft leans on distribution to mask model gaps. Meanwhile, the real profits are quietly shifting to the infrastructure layer with memory chips turning into luxury goods amid relentless AI demand, software itself is becoming fluid with “personal software on demand,” robots are learning to share skills across hardware, and even the foundations—like Linux—are showing cracks under the pressure of this new stack.

Enjoy Five Things Tech!

The $112 Billion Quarter

Amazon is betting on vertical integration. It landed 2.1 million AI chips over the past twelve months & its chips business has crossed a $20 billion annual revenue run rate, growing triple-digit percentages year-over-year. OpenAI committed to consume approximately 2 gigawatts of Trainium capacity through AWS starting in 2027. Anthropic secured up to 5 gigawatts.

But Amazon doesn’t own the model layer. Google does.

The hyperscaler that owns the model layer is growing the fastest.

Google can nicely integrate Gemini into their products. Amazon still tries to figure out how to make Alexa smart and Microsoft hopes that people stay ignorant about the “performance” of co-pilot.

(…continue reading.)

AI Has Made Memory Chips One of the World’s Most Profitable Products

While fears grow over whether AI services will eventually reap big profits, an epic windfall is flowing to the companies involved in the build-out of related infrastructure.

The historic run doesn’t look likely to end soon. Based on Samsung’s prebooked orders, the supply crunch is expected to grow worse next year, said Jaejune Kim, the company’s executive vice president for memory. “The available supply is far short of customer demand,” he said on a Thursday earnings call.

Since the start of this year, shares of Samsung have risen by 72%. SK Hynix’s shares are up 90% and Micron has gained 65%.

Memory is the new luxury.

(…continue reading.)

Why one longtime coder says vibe coding matters beyond tech

Up until now, no one has ever been able to get enough software. Engineering was very expensive. It took a lot of time. You had to buy stuff off the shelf.

What we don’t know is: A world in which it’s fast to customize, in which it’s easy to make something that’s really bespoke, and where engineering might be more of your service org — to help you as opposed to this alien entity bolted onto the org that does its own thing — how does that all fit into the future?

Personal Software on Demand - PSOD. I just coined the term, you read it here first.

(…continue reading.)

Robots With Different Designs Can Now Share Skills

Surprisingly, the approach doesn’t rely on AI. Instead, the researchers analyzed the physical properties of several robotic arms with three rotating joints—a popular design in commercial settings—to map out their limits.

To complete a task, a robotic arm must calculate how to bend each joint to reach its target. It also has to avoid pushing the joints past their physical limits or twisting them at weird angles. Engineers call these limits “singularities” because they cause the math governing the robots’ motion to break down. Failures can cause sudden and unsafe movements.

Turns out that robots can play together just nicely, which makes them more or less human? You decide…

(…continue reading.)

The most severe Linux threat to surface in years catches the world flat-footed

The critical flaw, tracked as CVE-2026-31431 and the name CopyFail, is a local privilege escalation, a vulnerability class that allows unprivileged users to elevate themselves to administrators. CopyFail is particularly severe because it can be exploited with a single piece of exploit code—released in Wednesday’s disclosure—that works across all vulnerable distributions with no modification. With that, an attacker can, among other things, hack multi-tenant systems, break out of containers based on Kubernetes or other frameworks, and create malicious pull requests that pipe the exploit code through CI/CD work flows.

Just at the time when I get my own linux server up and running after like 20 years, it gets hit by CopyFail. Yikes.

(…continue reading.)

That’s all for now! Thanks for reading! If you missed last week’s Five Things Tech, you can find it here:

🤖

— Nico

Subscribe now

Silicon Valley is racing to build human-substituting AI because if you believe it is inevitable, every company should compete to claim a market valuation the size of the entire economy, even if that means summoning a permanent underclass through market logic. Meanwhile researchers warn of the Social Edge Paradox: if AI capability depends on the social complexity of human language, and AI deployment systematically reduces that complexity through cognitive offloading, then the technology is undermining the conditions for its own advancement. Google is hemorrhaging focus across Antigravity, Gemini Code Assist, Jules, and five other tools (Jules leader just jumped to OpenAI), handing the coding race to Anthropic while proving that billions in capital mean nothing without product coherence. The cost frontier has bifurcated into two clusters (OpenAI pricing up, DeepSeek pricing down) with a disappearing middle class that mirrors economic polarization in the real world. And the CFA Institute warns that when machines replace rather than augment human inquiry, cognitive delegation erodes the capacity for independent reasoning that generates new knowledge. So we have acceleration creating polarization, undermining its own substrate, fragmenting even the best players, and collapsing cognitive foundations. The velocity is real but the contradictions are compounding faster than solutions.

Subscribe now

Happy Labor Day! ✊🏻 And don’t forget to check out GRID!

Silicon Valley Is Bracing for a Permanent Underclass

If left to its own devices, Silicon Valley may summon a permanent underclass through its own market logic. If you believe that human-substituting A.I. is inevitable, then every company should race to be the one to build it — and claim a market valuation the size of the economy and then some.

New A.I. models are assessed based on how well they do on a set of benchmarks — essentially standardized tests for the model. Increasingly, these evaluations emphasize real-world economic utility, which means that developers are aiming directly at replacing human capabilities.

Silicon Valley sure hits different.

(…continue reading.)

The Social Edge of Intelligence

If AI capability depends on the social complexity of human language production—and if AI deployment systematically reduces that complexity through cognitive offloading, homogenization of creative output, and the elimination of interaction-dense work—then the technology is gradually undermining the conditions for its own advancement. Its successes, rather than failures, create a spiral: a slow attenuation of the very substrate it feeds on, spelling doom.

This is the Social Edge Paradox, and the intellectual tradition it draws from is older and more interdisciplinary than most AI commentary acknowledges.

Think about this.

(…continue reading.)

Google’s internal struggle is handing the AI coding race to Anthropic and OpenAI

Anthropic cut off OpenAI’s access to its models last year, Wired magazine reported. Google has invested billions of dollars in Anthropic. A spokesperson for Anthropic did not immediately reply to a request for comment.

In recent years, DeepMind has tried to tighten control over how its AI breakthroughs are woven into Google products. Last year, Google appointed Kavukcuoglu to a new position as chief AI architect, a role in which he is charged with folding generative AI into Google products. Yet confusion about who is leading the charge on AI coding persists. Along with DeepMind, Google Cloud, Google Core, Google Labs and Android are all pushing AI coding in different ways, one of the people said.

Google released its Antigravity platform last year following the acquisition of talent and technology from startup Windsurf in a $2.4-billion deal. It joined a cluttered lineup of Google AI coding tools that includes Gemini Code Assist, Gemini CLI, AI Studio, Firebase Studio and Jules. Kathy Korevec, who oversaw Jules, jumped from Google to OpenAI earlier this month, according to her LinkedIn profile.

As an Antigravity heavy user I can totally relate - it shows that Google is lacking focus and fighting on too many fronts at the same time. Having said that, Google is far better positioned than Microsoft or Apple at the moment - and compared with OpenAI and Anthropic it has a massive userbase.

(…continue reading.)

The disappearing AI middle class

The cost frontier no longer behaves like a smooth curve. It is two clusters of economics with a stretched gap in the middle, and the gap is not going to close on its own in the near term. OpenAI will continue to release fast and price up, because the integrated product is the moat. DeepSeek will continue to release open weights and price down, because the commodity infrastructure thesis depends on adoption. Both can be right for different workloads, and the same agent can route between both within a single task.

And then of course we are having an growing pool of open source LLM that grow ever more powerful and will be used more and more.

(…continue reading.)

The Perils of Declining Judgment in the Age of AI

AI can extend the reach of analysis and amplify the scale at which data can be explored. But it cannot replace the fundamental processes through which humans generate and evaluate knowledge. The pursuit of evidence remains an inherently human responsibility. It can assist in collecting and organizing evidence. It can help detect patterns that might otherwise remain hidden. Yet the act of questioning assumptions, interpreting meaning, and deciding which observations matter remains fundamentally human.

This distinction is crucial. When machines begin to replace rather than augment the processes of human inquiry, societies risk weakening the epistemic foundations that sustain progress. Cognitive delegation may improve short-term efficiency, but it can also erode the capacity for independent reasoning that generates new knowledge.

Civilizational progress has therefore never been the product of technological capability alone. It has emerged from a delicate balance between innovation and reflection, between exploration and verification. When this balance is maintained, technological tools can accelerate discovery. When it is lost, progress risks becoming dependent on systems that humans no longer fully understand.

Oh, this is so worth a read. Actually, behind die paywall you can read my commentary on it!

(…continue reading.)

Read more

This Saturday the 40th anniversary of the Hamburg Marathon was celebrated with a course record. Of course I missed Othmane El-Goumri flying by near our house, on his way to finishing in 2:04:24, but I got to cheer on plenty of slower runners right at the 22km mark.

It was just a gorgeous day for running. So I also tried it and the stupid pain in my forefoot came back after 500m, so I clocked some 1,4 km or so and called it day.
Running sucks. Especially when you cannot run.

Subscribe now

Enjoy these Five Things Running! 🏃🏻‍♂️

The 2 hour barrier has been smashed.

Sabastian Sawe just ran 1:59:30 with crazy negative splits, closing the last half in 59:01....faster than the American Record in the half marathon.

One of the most mind blowing performances we’ve seen maybe in the history of any sport. How did we get here?

Every breakthrough is a mixture of belief and progress. It takes folks daring to see what’s possible, surrounding themselves with a quality team and doing the work to give themselves a shot.

You’ve got to bet on yourself in a big way. Which is exactly what Sawe did in multiple ways. When asked whether he believed he could run a sub-2-hour marathon before the race, Sawe answered with one word: “Yes.”

Ok, wow. That’s insane. I am in awe.

(…continue reading.)

If On Shoes Are for Everyone, Are They Still Ons?

In gyms, in malls, in airports and on sidewalks around the world, the shoes — with a sole that looks from the side like a happy kindergartner’s gaptoothed smile — are hard to miss. Hellen Obiri won the most recent New York City Marathon wearing a pair of Ons — and expects to don a new prototype on Sunday for the London Marathon. Middle-aged dads, with more modest athletic aspirations, fill their closets with the sneakers as well.

But having made the leap from start-up to established name in performance footwear, the company finds itself at a critical juncture: How does it stay true to its roots as a brand for serious athletes while keeping up its breakneck growth, which it can do only by appealing to the mainstream?

While I think On can also make good shoes, for me they will forever be the brand worn by wanna-be rich people strolling on the boardwalk of some beach resort in Germany, not for serious runners.

(…continue reading.)

How do you train for 250 miles?

Training 15-22 hours per week, not including commuting to trails or time spent in the gym, takes significant mental and emotional energy. Then add in caring for a pregnant wife on top of shipping orders and working to grow a brand…your nervous system will be maxed out.

My experience of this was high stress before a run, the stress generally leaving as the run progressed, and then coming back as soon as I hopped back in the car. I observed myself gravitating more towards mindless scrolling and cheap dopamine hits when feeling tired at home too. Sometimes I’d pull up into our carport and sit in the car for 10 minutes before heading inside. Even typing all of this out I can feel anxiety starting to flood my system.

Unfortunately, my dog is getting older, so training with her is no longer possible and I guess that’s why I cannot train for 250 miles. No, in all seriousness, I am always in awe when I read about people training for such endeavors.

(…continue reading.)

The Hardest Part of Injury Isn’t Physical

If belief shapes perception, and perception shapes experience, then the beliefs you reinforce—intentionally or not—start to define how you move through both training and recovery. That’s where self-belief becomes more than confidence. It becomes a skill. It’s not something you either have or don’t have, but something you practice by deciding what you pay attention to and what meaning you assign to your experience.

For me, that’s looked like a shift away from asking “what am I losing?” and toward asking “what is still here?” The fitness I’ve built doesn’t disappear overnight. The discipline, the consistency, the ability to show up—those are still there. The body that adapted to years of training is still capable of adapting again. That distinction matters, because it gives you something stable to anchor to when everything else feels uncertain.

Oh, totally.

(…continue reading.)

The Strange Rise of Denim in Running Culture

A couple of years ago, people wondered why 3sixteen, a menswear brand that specializes in meticulously produced denim, was hosting monthly summer runs from its Manhattan store, but a little dig into their lookbook reveals a niche of equally thoughtful running apparel. 3sixteen founder Andrew Chen is a regular at World’s Fair Run Crew in Queens, and the running bug runs deep throughout the company and among the brand’s fans it has amassed over two decades of business.

The way that 3sixteen seamlessly brings together two parts of its identity in the Jorts Mile is a fascinating case study on the explosion of running into mainstream culture.

Yeah, no. Just because it is possible, we don’t have to do it. Running in denim is just stupid.

(…continue reading.)

If you missed last week’s edition, you can read it here:

Now, go running!

— Nico

🏃🏻‍♂️

So right now everyone in Germany talks about gas prizes and for the second time in just a few years, the government has to dole out heavy tax deductions to get the ga prizes down. The Iranians are holding the Western world hostage with their blockade of the Straight of Hormuz.

I really do not understand why the answer is not: “ok, understood, we’ll get out of fossile fuels as fast as possible and reduce our dependence on other countries when it comes to the supply of energy!” - and I also do not understand why people get all emotional about their furnace and combustion engine.

We have so many great engineers, so I am not the least bit worried that we can get enough solar and wind capacities going, along with an improved grid and huge batteries. We just have to learn how to think bolder again.

Enjoy these Five Things!

Subscribe now

If America’s So Rich, How’d It Get So Sad?

American sadness this decade has been forged by the fact of, and the feeling of, a permanent unrelenting economic crisis, amplified by a uniquely negative news and media environment, and exacerbated by the rise of solitude and the declining centrality of trusted institutions. Inflation has made today’s life harder to afford, while the ambient awareness of other people’s triumphs on social media had made tomorrow’s success feel harder to achieve. The ongoing collapse of confidence in the establishment has made Americans feel unusually adrift and dissatisfied with institutions outside of their control, while the chosen self-isolation of modern life has demolished communal trust, as we increasingly experience other people’s minds through the toxic surreality of our screens rather than through the embodied reality of strangers who are, for the most part, just as nice as we are.

It’ really the paradox of our time. We have never been better off, yet we are not as happy as we should be. In Europe, the Fins and the Danes are the most happy people and in Germany it’s the people living in Northern Germany, right adjacent to Denmark, who are the most happy people. It’s probably the wind and the open sea around here. Or the rain.

(…continue reading.)

Muskism as Fordism

Musk and his colleagues have the good fortune of living in a moment when there is no structural actor capable of challenging their dominance. The working class has virtually ceased to exist as an organized force. Absent pressure from below, political parties themselves offer no meaningful opposition to Muskism. The effect is paradoxical. On the one hand, the lives of capitalists are made easier when they can intensify the exploitation of their workers and purchase influence in politics with little resistance. But it also means they have no incentive to restrain their most antisocial impulses or consider the long-term consequences of their actions.

Muskism typifies this tendency: if Fordism and post-Fordism were both, in different ways, organized to secure social peace, Muskism is oriented toward social war. The relative thinness of the Muskist mode of regulation is symptomatic: the working-class antagonist is so weak, the social war so asymmetric, that there is no need for a negotiated peace. At the moment, the strategy seems to be working. Musk, already the world’s richest man, is projected to become the first trillionaire as soon as next year.

I’ve said it before: we need better regulation. We are too slow to understand what’s going on and then we don’t regulate as tough as we should.

(…continue reading.)

Creating Baby Geniuses to Thwart the AI Threat? (Yes, Really.)

Investors follow the money, of course, but part of the dual appeal of genetic optimization and AI is that both are central to transhumanism. This futurist philosophy, popular among the tech elite, aims to marry advancements in biology and technology to accomplish things today’s humans cannot—like extending our lives (perhaps forever!) or circumventing climate change (by colonizing other planets). While it may seem odd that these billionaires are constructing one technology some of them admit could bring about human extinction, even as they back another one to save us from what they’re building, there is, in fact, a unifying theme: “the rejection of limitation,” explains Alexander Thomas, author of The Politics and Ethics of Transhumanism. “That colonial impulse of ‘I want more.’”

I’m all for techno-optimism, but I opt out of transhumanism.

(…continue reading.)

Into the Right-Wing Dreamworld

If the first Trump administration was a regime of words — the years of presidential tweets and unfiltered musings; of covfefe, grab ’em by the pussy, not sending their best — the second is a regime of images. Of course there were totemic visuals in those earlier years: tiki-torch-wielding neo-Nazis, gas-masked cops streaming out of burning precincts, face-painted hordes descending on the Capitol. (Crowds and fire were notable motifs.) But for the most part, the novelty of the first administration’s political style lay in its desublimation of speech. How could he say that, liberals were always asking. Today’s unofficial MAGA slogan, in contrast, is You can just do things — that is, the government can flood the streets with murderous federal cops, or start a war with Iran without waiting for Congressional approval, let alone making any real effort to manufacture the public’s consent. Funnily enough, this change mirrors a reverse development, a shift from an administration of professional doers to one of professional talkers: from career bureaucrats like Steven Mnuchin and Jim Mattis to the posse of podcasters, news anchors, and health influencers who stalk the halls of government today.

It’s weird. And here in Germany the far right is trying to instrumentalize the fate of a stranded wale, pushing the image of the wale as the symbol for a government that is not trying to help its people. It’s all beyond bizarre and takes me back to the article about happiness. We have lived in comparatively peaceful times with lots of prosperity, still people believe right-wing idiots who want to exploit them.

(…continue reading.)

The Disappearance of the Public Bench

It might be true that benches can’t create community, but they can support the interactions that do. They give the elderly the ability to navigate their neighborhoods, providing opportunities for cross-generational interaction. Noisy phone calls and messy to-go meals that might otherwise be taken indoors are instead indulged in public, creating minor vexations but bringing people — and watchful eyes — to streets that have been scarred by violence. Despite all they’re up against, the Bronx benches are an invitation to share a space and build a public culture on what was formerly a patch of cobblestones.

A couple of years ago they started putting up benches in my neighborhood at a few streetcorners. One bench, right next to the subway station, disappeared right away and I assume that happened because people sat on it and drank and smoked. All the other benches stayed and I find it delightful to see people there, young and old, just sitting there, enjoying the day.

(…continue reading.)

That’s it. Have a great Sunday! If you missed last Sunday’s edition of Five Things, have a look here:

Subscribe now

— Nico

The week in tech is basically: build stuff in cowboy country, burn the planet to feed GPUs, turn war into a drone SaaS line item, and then fix our eyes with liquid metal while billionaires LARP as end-times prophets. In other words, we’re speed-running a cyberpunk timeline where Texan hard-tech playgrounds churn out the hardware, gas-fired data fortresses belch more CO₂ than entire nations, defense departments subscribe to autonomy like it’s an enterprise license, and the same guys warning you about Armageddon are quietly trying to own the off-ramp.

Enjoy Five Things Tech!

‘Startup Cowboys’ Are Making This Texas Town the New Tech Hotspot

Aspiring software companies and AI firms are still flocking to San Francisco. But more robotics, energy and defense companies are setting up shop under the blistering Texan sun. Places like Proto-Town are the latest sign that Austin and its surroundings are becoming a base for these so-called “hard tech” companies.

In their downtime, the mix of Texans, company founders and engineers who live in Proto-Town like to ride around the flatlands on dirt bikes, head into Lockhart for barbecue or drive to downtown Austin to go two-step dancing.

Startups on the site are attracted to the state’s low regulation and the build-anything atmosphere of Proto-Town and the surrounding Caldwell County.

This sounds like fun and necessary - when tech companies turn too corporate, they lose the ability to innovate. It’s the classic innovator’ dilemma that is hitting the whole industry.

(…continue reading.)

How the Tech World Turned Evil

Class warfare doesn’t get more unhinged than this. Say what you will about the American plutocracy, it seldom frames its economic self-interest as a religious imperative. But even in its more innocent days, Silicon Valley inclined toward grandiosity, heralding not just a new technology but a new advancement in human consciousness. Now a prince of the technocratic elite was framing tech’s future prosperity quite literally as a battle against agents of Satan, with Thunberg and Yudkowsky cast as Gog and Magog. Suggest Thiel’s dominion could stand a bit more government oversight and he just might toss you into a lake of fire.

I remember the time when the internet was still largely uncommercial. It sure changed a lot and greed didn’t really do the ecosystem any good.

(…continue reading.)

New Gas-Powered Data Centers Could Emit More Greenhouse Gases Than Entire Nations

As tech companies race to secure massive power deals to build out hundreds of data centers across the country, these projects represent just the tip of the iceberg when it comes to the potential climate cost of the AI boom.

The infrastructure on this list of large natural gas projects reviewed by WIRED is being developed to largely bypass the grid and provide power solely for data centers, a trend known as behind-the-meter power. As data center developers face long waits for connections to traditional utilities, and amid mounting public resistance to the possibility of higher energy bills, making their own power is becoming an increasingly popular option. These projects have either been announced or are under construction, with companies already submitting air permit application materials with state agencies.

Lawmakers should require new data centers only to use solar and wind power, which makes so much more sense than relying on fossile fuels for these huge buildings with their immense power consumption.

(…continue reading.)

Pentagon officials broadly detail $55 billion drone plan under DAWG

The DAWG essentially absorbed the Biden-era Replicator initiative designed to acquire thousands of low-cost “attritable” drones, particularly for use in the Pacific. In its latest iteration, the department is looking to supercharge development investments across an array of platforms and collaborative autonomy efforts.

But to do that, it will take a big investment in the form of a 24,070 percent increase over the $225.9 million DAWG received in fiscal 2026. Specifically, the Pentagon is seeking a $54.6 billion ask for research and development dollars in the FY27 budget request — $1 billion in the base budget and the remaining $53.6 billion coming from the more flexible future reconciliation pot.

These numbers are staggering. There are so many military and civilian use cases for drones that a whole new industry is currently spawning.

(…continue reading.)

This artificial retina doesn’t just aim to restore sight—it opens a hidden channel of vision

The artificial retina developed by this research team has two primary components: a phototransistor array and a set of liquid metal micropillar electrodes. The phototransistor array is a grid of tiny, light-sensitive devices that can detect near-infrared light (i.e., light that is just beyond visible wavelengths) and convert it into electrical signals.

Liquid metal micropillar electrodes, on the other hand, are pillar-shaped structures made of a soft liquid metal that conducts electricity. These structures deliver the electrical signals produced by the phototransistors directly to cells in the retina that send visual information to the brain, known as retinal ganglion cells. In most cases of retinal degeneration, these cells are less affected than photoreceptor cells, thus they might still be able to transmit information to the brain.

Ray-Ban Meta smart-glasses are so 2025.

(…continue reading.)

That’s all for now! Thanks for reading! If you missed last week’s Five Things Tech, you can find it here:

🤖

— Nico

Subscribe now

Enjoy this edition of Five Things AI!

A while ago, a $10 billion startup reached out asking if I would help train AI to replace white-collar workers, and I politely ignored them. Mercor is building APEX to measure whether AI can actually do professional work (not just pass standardized tests), and they need humans to create training tasks that atomize jobs into machine-learnable components. Meanwhile Google unveiled two new TPUs for the "agentic era" because training and running frontier models is still burning money with unclear ROI, we are discovering that a good AGENTS.md file can boost agent performance by 25% or tank it by 30% depending on how you write it (documentation quality is now a model upgrade), Maggie Appleton from GitHub Next unveiled Ace because when you have two dozen coding agents running in parallel you have zero alignment and all our current tools were built for a world where implementation was slow, and Yann LeCun is arguing LLMs have plateaued because they only train on text while humans build world models from smell, touch, temperature, and movement. So we have hardware scaling and models improving while simultaneously discovering that agent collaboration, documentation quality, team alignment, and sensory modalities are the new bottlenecks. The velocity is real but the problems are shifting faster than the solutions.

Subscribe now

Enjoy this edition of Five Things AI! And don’t forget to check out GRID!

The $10 Billion Startup Training AI to Replace the White-Collar Workforce

Beyond the messy day-to-day mechanics of atomizing jobs into training tasks is a bigger question: Are the machines actually getting good enough to do the work themselves? In the past year the AI industry has tried to answer that question, mostly with a growing ecosystem of “evals,” or evaluation frameworks. Like a standardized test, an eval measures a large language model’s math ability, reasoning or factual accuracy—but not typically the kinds of specific, judgment-heavy tasks that white-collar workers do all day. Mercor has developed its own test, a project called APEX (AI Productivity Index), that attempts to do just that. It measures professional performance and then posts the results on the web for anyone to follow as a kind of industry leaderboard.

Interesting. And spooky. They have reached out to me a few times and wanted me as a native German speaker with expertise in business and tech. I politely ignored their LinkedIn outreaches. Would you work with them?

(…continue reading.)

Google unveils two new TPUs designed for the “agentic era”

It makes sense that efficiency is a core part of Google’s new TPU setup. Training and running frontier AI models is expensive, and the return on investment is unclear. Companies are still burning money on generative AI in the hopes that efficiency will turn the corner at some point. Maybe Google’s new TPUs will help get there and maybe not, but the company has made notable improvements.

Fascinating how fast chip design is evolving and which companies are now developing and making chips. Is Intel still around?

(…continue reading.)

A good AGENTS.md is a model upgrade. A bad one is worse than no docs at all.

A single AGENTS.md isn’t uniformly good or bad. The same file boosted best_practices by 25% on a routine bug fix and dropped completeness by 30% on a complex feature task in the same module.

On the bug fix, a decision table for choosing between two similar data-fetching approaches helped the agent pick the right pattern immediately and stay within codebase standards. On the feature task, the agent read that same file, got pulled into the reference section, opened dozens of other markdown files trying to verify its approach against every guideline, created unnecessary abstractions, and shipped an incomplete solution.

Different blocks of the document had opposite effects on different tasks.

What follows is which patterns work, which fail, and how to tell which is which for your codebase.

Oh, totally. A normal AGENTS.md is fine, but a great one is almost indistinguishable from magic. And a bad one leads to really annoying behaviour.

(…continue reading.)

One Developer, Two Dozen Agents, Zero Alignment

We need tools that help everyone on the team align before the agents start working, not after.

That alignment needs to happen constantly, alongside the implementation.

Planning and building are no longer separate phases. They’re a continuous cycle. The tools of the future need to bring planning, context-gathering, decision-making, and development under one roof.

This really resonated with me. I always thought it is my setup, but it seems to be a real challenge these days.

(…continue reading.)

Why AI Needs A Sense Of Smell

A growing number of AI researchers and entrepreneurs believe that LLMs have already plateaued. LeCun is a prominent advocate for world models — an emerging paradigm in AI inspired by the idea that human minds create internal representations of the world made out of a variety of sensory data, and use them to predict, plan and execute actions.

Whereas LLMs are trained primarily on exabytes of text, proponents of world models point out that a human just existing in the world will pick up vastly more. This includes not just language, sound and vision data, but also touch, temperature, balance, movement and other sensory modalities — including smell. This information may not help us solve a mathematical proof, but it does help us learn and make connections in ways that AI today cannot.

Crazy idea, or rather: clever idea. I never thought of that, but it makes so much sense.

(…continue reading.)

Read more

I haven’t looked at my Strava in weeks, but I hope these days will be over soon. I was able to walk with less pain in the last few days and I’m hopeful that I can start running again this week.

We’ll see, but I sure miss running.

I kind of dread the moment I am getting out there again. It will be a long time until I reach the fitness level I like to have after not running for at least two months.

Well, I guess I’ll take it one step at a time, literally.

Subscribe now

Enjoy these Five Things Running! 🏃🏻‍♂️

Fast Women: More pro runners are having babies, but it’s not easy

Not long ago, many professional runners in the U.S. waited until the end of their careers to have children, if they had them at all. That has begun to change. In the past several years, an increasing number of high-level runners have announced pregnancies while still competing. Among the reasons for the shift? Stronger sponsor support, a better understanding of what training during pregnancy can look like, and a growing list of athletes showing what’s possible.

That doesn’t mean it’s easy. Raising young children is difficult to balance with any career. For professional runners, part of the challenge is that elite sport and pregnancy require the body to adapt in different, sometimes competing, ways. And every athlete’s path back to high-level running comes with its own set of obstacles.

I have so much admiration for mothers in the sport. I don’t know how they are getting this accomplished, but I am glad there are more and more female pro runners out there.

(…continue reading.)

Is Trail Running Actually Growing? Here’s What the Data Says.

This tells two things simultaneously. First, the sport is already vastly larger than its race scene. Millions of people run trails weekly without ever crossing a finish line. Second, the race ecosystem has enormous room to grow. If even a small percentage of that non-racing majority eventually finds its way to a start line, the race participation numbers look very different.

Read together, the floor and the ceiling are both expansion signals. The sport is bigger than we measure it, and the race scene hasn’t come close to capturing it.

Trailrunning is a fantastic sport, so it is only natural that it is growing.

(…continue reading.)

Ultrarunning Culture

Ultrarunning is a growing sport. It used to be a small group of oddballs who, for whatever reason, wanted to run far. Now it is a large group of oddballs who, for whatever reason, want to run far.

With the growth of ultrarunning has come significant changes. Training methods are more sophisticated. Media attention has increased, and there is more money in the sport. Events are bigger, except for the ones that remain the same size but now have lengthy waitlists. The variety of event distances, surfaces, and topography has increased, and there are more race opportunities than in the past. There are also more factions — dissenting groups with competing visions about what the sport is or might become.

It must be weird for runners who have been around in the sport for a while and then all the commercial stuff kicks in.

(…continue reading.)

Experience: I won the world’s deepest underground marathon

On the day of the marathon, I was nervous but excited. I’m not claustrophobic, but it was unnerving to know we were so far underground. I tried not to think about the 1,300 metres of solid rock over my head.

The mine was operational and I could hear the hum of machinery through the walls. As we were escorted to the starting point, I was fascinated to see vending machines and brightly lit offices. It was an entire underground world.

The tunnel was five metres wide and stretched for over a mile. To make up the distance of a full marathon, we would run to the end and back 11 times.

Would you want to participate in a race like that?

(…continue reading.)

We’re So Busy Making Running Cool, We Forgot Why We Love It

The through-line is that people are exhausted by optimization. Everything has been focus-grouped and content-strategized and given a brand identity, and at some point the rough edges start to look really good. At some point the thing that was just a thing (not a movement, not a platform, not an activation) starts to feel like what you actually miss.

Running used to be like that. Weird and lonely and slightly embarrassing in the best way, something you did because something in you needed to, not because it photographed well. I ran Ragnar back when it was chaotic and silly and you slept in a van and ate gas station food at two in the morning, and I loved it for exactly that scrappiness. Then it became a $2,000 entry fee with branded finish lines, so the Speed Project came along as a beautiful middle finger of a race, six people running LA to Vegas because why not. And then TSP did the exact same thing in about half the time. Even UTMB has a Hoka logo and a global points system. I still want to run it, and yes that makes me a hypocrite, but when I’m out on those trails in Chamonix one day I know I wont care.

Oh yes.

(…continue reading.)

If you missed last week’s edition, you can read it here:

Now, go running!

— Nico

🏃🏻‍♂️