My son is 15. In a few days he starts secondary school - for now at a gymnázium, the Czech academic track, so he still has a little time before he has to decide. In these hard-to-read times I see that as a big advantage. How do I advise him? I decided to write it down and try writing my son a letter. I’d have appreciated one at his age. Or maybe I’d have laughed at it and shoved it in a drawer, because at fifteen you don’t think of your dad as the smartest person you know :-). Either way, I’d have regretted not at least trying.

So here’s my letter - maybe it’ll give you something to think about if you’re in a similar situation. And below the letter you’ll find how I got there. If you want to go deeper, the option is there!

Dear Son,

I know you hate it when I give you advice. So this isn’t advice. It’s more what I’d want to hear from my own dad if I were fifteen today. Stash this scrap of paper somewhere and try to look at it now and then.

First, a confession. When I was choosing a school, the world was reasonably predictable. You graduate, you find a job, you do it, eventually you get promoted. Or you go into business, but it played out much the same way. Studying was something you got out of the way at the start of your life, and then you worked for forty years. That model died during my lifetime, and during yours it won’t exist at all. You’ll be learning your whole life. Not because it sounds nice, but because the world will change faster than any school can rewrite its curriculum. So if there’s one single thing you should learn in secondary school, let it be this: learn how to learn. Work out how your head works, when it thinks best, how to get into something you don’t understand. That will stay with you even if everything else goes obsolete. And I think things are going to go obsolete damn fast.

Second thing. When I was young, “IT people” were a strange caste who understood computers, and the rest of the world came to watch them. Being average, or even below average, was enough not to starve. That’s ending. Anyone who wants to be successful will have to become an IT person - but an average IT person will be worthless. Not that everyone will program, but everyone will build their own tools, their own little apps, their own ways of working with AI. Whoever can do that will have a multiplied advantage in any field. That’s why I’m not telling you “go study computer science”. I’m telling you: wherever you go, take this with you. AI is the new basic skill, the way reading was for me and English was for my generation. But on its own it won’t be enough. Combine it with something. With a field, a craft, a subject that’s yours.

The third thing is uncomfortable, so I’ll say it straight. You’re smart. You know it and I know it. And that’s exactly why I have to warn you about the trap smart people fall into. Everything comes fast to a smart person, and AI makes it faster still. It’ll spit out your homework, your essay, your code, whatever, and nobody can prove a thing (if you’re clever about it). It’ll be terribly tempting to use your intelligence to avoid doing any work at all. Use AI to make yourself smarter, not to make your life quiet. Have it explain things to you, argue with it, test whether it can catch you out. You’ll know which one you’re doing by whether you still know things when the wifi goes down. Smarts without work are just laziness in nice packaging, and the world spots it faster than you think.

The fourth thing may surprise you, because it comes from someone who has worked in technology his whole life. Don’t throw out the humanities and the arts. The world is being flooded with text, images and music made by machines. There will be an infinite supply of it and it will be free. Which makes the person who can say what of it actually matters that much rarer. Someone who understands people, who can build an argument, who can ask a good question. Someone who can create something genuinely unique. History, literature, art, philosophy - those aren’t useless subjects. They’re training for exactly the muscles machines don’t have.

Fifth. You can see at home that it’s possible to live freely and run your own thing, so you know it isn’t a YouTube fairy tale. But let me tell you what’s really behind that freedom, because from the outside it looks easier than it is. Freedom isn’t the opposite of work. Freedom means you can choose what you work on, who you work with and where you work from - and that choice is bought with one single currency: being properly good at something and having people trust you. That’s why your brand will matter - what can be found about you, what you’ve built, what people say about you. You don’t have to be an influencer. It’s enough that things are left behind you and that you’re not afraid to put your name on them.

And now the thing you may want to ask about: what I’d wish for. I’d avoid the directions where the ground is shifting most right now: rank-and-file coder, lawyer, marketing, admin. You’d make me happy with medicine, caring for people, a proper trade, or deep knowledge of IT and security - things that rest on your hands, on responsibility and on trust. Those are my bets.

But feel free to cross out that whole paragraph. Because the main thing is something else, and I know it sounds worn out: do something you enjoy. I’m not saying it because it sounds nice. I’m saying it for a purely pragmatic reason. When you enjoy something, you stick with it even when it gets hard. And if you stick with it, you’ll be above average. And that’s what the whole thing is about. The world you’re stepping into will be merciless to mediocrity - an average text, average code and an average idea can be produced by a machine in a second, for free. Mediocrity used to be the safe choice. It isn’t anymore. The future has no time for the average. It has plenty for the driven. An above-average carpenter will outlast an average programr. An above-average teacher will outlast an average lawyer.

So when you’re filling in your application, don’t ask which field will survive. Nobody knows that. Ask where you’ll learn to think properly and where you’ll be doing real things. And if something really grabs you, that’s the right path.

And what if you get it wrong? Doesn’t matter. In your world, choosing a school isn’t a verdict anymore. It’s the first move, not the whole game.

Dad

This Special isn’t a list of safe fields. No such list exists (actually the opposite is true - there are plenty of them, but…) and anyone selling you one is guessing. It’s a map of what we actually know today, where the smartest people on the planet disagree, and what follows from that for parents standing in front of an application form.

What’s actually happening

The International Labour Organization analyzed almost thirty thousand work tasks and came to a cautious conclusion: most professions touched by AI will be transformed rather than eliminated. So far AI is good at tasks, not at whole occupations. It can write the first draft of a contract, but it can’t carry the responsibility when the contract turns out to be wrong.

But then there’s a rather different signal. Economists at Stanford (among them Erik Brynjolfsson, one of the biggest names in the economics of technology) analyzed wage data covering millions of American employees. They found that young people aged 22 to 25 in occupations heavily exposed to AI have employment roughly 19 percent lower than you’d expect from their peers in less exposed fields. For people over thirty there is no such gap. For experienced workers in those very same fields, employment is actually growing.

And the mechanism doesn’t look like layoffs. It looks like weaker hiring. Companies aren’t firing juniors, they’re simply not taking them on. AI does the first analysis. AI does the first research. AI writes the first version of the code.

And here’s the question every parent should be asking. Not “will that profession exist in 2040”. But: will my child get the first working years in which they learn the craft? Because you don’t become a senior at school. You become one by doing junior work for several years, making mistakes, and having someone more experienced correct them. When the bottom rung of the ladder disappears, nobody gets to the top.

Enjoying News from the Woods? Buy me a coffee💚.

🎧 News from the Woods as a podcast on Spotify or Apple

End of the commercial break, let’s carry on…👇

Four camps: where the people building AI say to send their kids

If there were a clear answer, the people running the AI companies would agree on it. They don’t. And that in itself is the most important piece of information in this Special.

Camp one: go where AI physically can’t. Geoffrey Hinton, Nobel laureate and “godfather of AI”, says it without dressing it up: if you want certainty, become a plumber. In his view routine intellectual work is more vulnerable than skilled manual work in an unpredictable environment. It sounds like a punchline, but there’s logic to it: AI can write an essay, but it won’t replace the boiler in a 1978 prefab apartment block where nothing is up to code.

Camp two: the biggest casualties will be office juniors. Dario Amodei, head of Anthropic (the company building the AI model Claude), warned last year that within a few years AI could wipe out a large share of entry-level office positions and push unemployment to 10 or 20 percent. It’s a prediction, not a fact. But notice: it fits the Stanford signal exactly.

Camp three: stick to math and fundamentals. Demis Hassabis, head of Google DeepMind and also a Nobel laureate, tells students not to give up on STEM, math and computer science. Not because AI can’t do the math, but because the fundamentals let you use AI an order of magnitude better than everyone else. To that he adds a skill he calls crucial: learning how to learn.

Camp four: the field doesn’t matter, what decides is field plus AI. Jensen Huang, head of Nvidia, became famous for saying kids shouldn’t learn to program anymore, because the programming language of the future is human speech. His more recent message is more interesting: what you study won’t matter that much, but whether you can use AI in your field will. Journalism, design and science will all still exist. They’ll just be done differently.

One caveat: all four gentlemen sell AI. Hinton no longer does - he’s warning from retirement - but Amodei, Hassabis and Huang all have a commercial interest in a particular story about the future. That doesn’t mean they’re wrong. It means you should read what they say as testimony from interested witnesses, not as a verdict.

Computer science: a dead field, or the best choice?

Most of today’s panic is around computer science. And fairly so, because at first glance the data looks schizophrenic.

On one side: unemployment among fresh computer science graduates in the US has climbed to 6.1 percent, and for computer engineering to 7.5 percent. That’s above the average for all graduates, and also above art history and philosophy graduates. Read that sentence again. A philosopher in America has less trouble finding work today than a fresh programr. The number of developer job ads on Indeed has fallen 71 percent since 2022. Enrollment in classic CS programs at American universities dropped 11 percent year on year.

On the other side: the US Bureau of Labor Statistics projects employment of software developers to grow 15.8 percent by 2034, five times the average. Data scientists are projected to grow 33.5 percent, cybersecurity 28.5 percent.

How do those fit together? Quite simply. The price of simple programming is falling. The price of deep understanding of systems is rising. AI now writes routine code faster than a junior. But someone has to design the architecture, guard security, understand the data and, above all, recognize when AI is talking nonsense. Computer science isn’t dead. What’s dead is the idea that a three-month coding bootcamp is a ticket to job security.

For parents that leads to an uncomfortably demanding piece of advice: if your child is drawn to computers, don’t stand in the way. But push for the fundamentals. Maths, systems, security, and ideally a second specialism alongside. A programr who understands biology, energy or finance is in a different league from a programr who can only program.

And what about university as such?

I sometimes hear that in the age of AI a degree no longer makes sense. It isn’t really anything new - as a young IT guy I told myself the same thing and never finished university. I’ve never missed it in my life. But the data says something else. According to the OECD, people with a university degree across developed countries earn on average 54 percent more than people with only an upper-secondary diploma. Even the differences between fields aren’t as big as the folklore has it: employment among graduates is around 90 percent in ICT, 89 in technical fields, 87 in education and 84 percent in the humanities and social sciences. Even that last group is well above people with secondary education only.

How the world is coping: from bans to compulsory lessons in two years

Nowhere is the confusion of this era more visible than in how schools handle AI. In January 2023, a few weeks after ChatGPT launched, New York public schools banned it on all their devices and networks. The reason: kids will cheat and stop thinking. Four months later the head of New York’s school system reversed the ban and publicly admitted it had been a mistake - that school should prepare children for AI, not hide them from it. America’s largest school district turned 180 degrees in a single semester. That’s a pace education has never had to get used to before.

Since then the world has gone in two directions. China took its own approach: since September 2025 AI teaching has been compulsory in all primary and secondary schools, Beijing mandated a minimum of eight hours a year, and the national plan aims for full coverage by 2030. Even in the lower grades kids learn what image recognition is; secondary students build their own applications. The United Arab Emirates introduced AI as an official subject from kindergarten. India is adding it to the curriculum from third grade. And little Estonia, Europe’s digital star pupil, launched the AI Leap program: the first country in the world to give all secondary students access to top-tier AI tools, teacher training included. The Estonians put it bluntly - for them AI literacy is a civic skill like reading.

What’s interesting is that almost none of them are pushing kids en masse into computer science. A different model is emerging: AI literacy as the fourth basic literacy alongside reading, writing and arithmetic. Across subjects, not as a subject. I wrote about this years ago, before AI - that computing shouldn’t work as a standalone subject, but should be part of every subject. With artificial intelligence it makes even more sense today.

And then there’s the second current, the one hitting the brakes. Sweden, long a showcase for digitalizing education, made a remarkable turn in 2023: the education minister paused the digitalization strategy and the government poured money back into printed textbooks, because reading literacy results were going down. Nordic teachers’ organisations jointly declared last year that AI should support teaching but never steer it or replace the teacher. France barely works with AI in schools at all, and Japan issued guidelines so cautious that what mostly comes through is nervousness.

The most important chapter: it’s not just about where, but how

Two recent studies showed that the same technology can either double a child’s education or take it apart. What decides is how it’s used.

Nigeria: the World Bank ran an experiment in which secondary students worked with an AI tutor for six weeks, but under a teacher’s guidance. The AI didn’t hand out answers, it explained and asked questions, and the teacher watched over comprehension. The result: progress equivalent to roughly 1.5 to 2 years of ordinary schooling. In six weeks.

China: researchers followed almost 27 thousand secondary students who started using generative AI for their homework. Homework scores went up 18 percent, time spent on homework fell 30 percent. Parents delighted, grades pretty. Except results in exams without aids fell 20 percent within six months. The kids were handing in better work and knew less.

To me this is the core of the whole problem. AI can be the best tutor in human history, available to every child for pocket change. And at the same time it’s the most perfect machine for outsourcing thinking that has ever existed. A muscle that isn’t used doesn’t grow.

So when you’re doing the rounds of open days, don’t just ask about fields and graduate outcomes. Ask: how do you work with AI here? A school that bans it is preparing children for a world that doesn’t exist. A school that lets it freely eat the homework is manufacturing the Chinese scenario. What you’re looking for is a school that can do the Nigerian thing: AI as a tool for thinking, not a replacement for it.

The skeptic’s corner

It’s only fair to admit that this entire debate could look ridiculous in five years. In either direction.

In 2013 the famous Oxford study by Frey and Osborne came out, predicting that automation would threaten 47 percent of American jobs within ten to twenty years. Everyone quoted it, me included. It didn’t happen. Unemployment over the following decade was at record lows.

Nobel laureate Daron Acemoglu of MIT calculated that AI will lift the economy’s productivity over the whole of the next decade by a mere 0.66 percent. In his view the hype is an order of magnitude bigger than the reality. And a Danish study that linked data on AI adoption in companies with actual wages found, after two years, essentially zero effect on wages and hours worked.

So yes: it’s possible that in ten years the world of work will look almost exactly as it does today and this Special will read like a panic attack from 2026. It’s also possible it will look completely different and this Special will read as desperately mild. Anyone who tells you they know which scenario is coming is guessing. Me included.

Which is exactly why a strategy that works in both scenarios makes sense.

What I’d do

One equation comes out of all this for me. It isn’t a list of fields, it’s a multiplication:

fundamentals x working with AI x deep expertise x contact with the real world x human skills

Notice it’s multiplication, not addition. If any one term is zero, the whole result is zero. A child who is brilliant with AI but has no foundations just produces things they don’t understand faster. A child with the foundations who ignores AI will lose to the one who has mastered it.

What that means in practice:

• Secondary school. A good gymnázium remains a robust choice for an academically minded child, provided it really holds the line on math, languages, writing and honest work without shortcuts. A good technical secondary school that combines electronics, automation and software with physical projects is gaining value. Health-care and teaching tracks have a human and physical component AI can’t do. And a trade is ceasing to be a second-class option: in a decade when data centres are going up and the energy system is being rebuilt, a skilled electrician or technician will be rarer goods than an average marketer. I’d be most cautious about schools built on routine admin and memorization.

• University. Computer science, math, data and engineering yes, but in depth, not as a coding course. Medicine, biology and the sciences gain an enormous multiplier from AI, but only for someone who can tell when the model is lying. Law, economics and business aren’t safe on their own; safer is their combination with quantitative skills and judgement. And don’t write off the humanities: the ability to argue, to understand people and to ask good questions will rise in value. The only thing falling is the price of average text. Anyone can produce that today.

Let’s not ask which school to send a child to so that AI won’t replace them. Let’s ask which school will turn them into a person whose abilities AI multiplies.

And now the hardest part, at the end: we choose a school for four to six years, but our children will be working for forty years in a world we can’t picture today. The only truly portable equipment is the ability to learn again and again. Everything else is a bet.

Where would you enroll your child today? And has the last year with AI changed your mind? Write to me, this topic deserves a follow-up.

Key sources

• Stanford Digital Economy Lab: Canaries in the Coal Mine (August 2026)

• ILO: Generative AI and Jobs, 2025 update

• Stanford HAI: AI Index 2026, Education

• U.S. Bureau of Labor Statistics: Fastest growing occupations 2024-2034

• OECD: Education at a Glance 2025

• World Bank: AI tutoring in Nigeria (Edo State)

• CEPR: The Generative AI Learning Penalty (China)

• BCG and the Czech Ministry of Labour: The Future of the Czech Labor Market (2025)

• NBER: Still Waters, Rapid Currents (the Danish study)

• Axios: interview with Dario Amodei

• Fortune: Jensen Huang on education

• Federal Reserve Bank of New York: Labor Market Outcomes of College Graduates by Major

I get this question so often that I started keeping a list. It sat in my notes for a long time, and the summer holidays were the perfect excuse to turn it into a special. Here it is.

The starting point

I’ve always been the kind of person who has to be doing X things at once. Sometimes it’s an advantage, sometimes it isn’t, but that would be a different special. My main job is growing Good Sailors, on top of that I run innovation at M2C, I help out at the Innovation Centre of the Ústí Region, I write Zprávy z lesa and 11 Houses, I sit on the municipal council in Doubice, I’m a member of the Bohemian Switzerland National Park Council and a volunteer ranger. Plus mentoring at a couple of startup accelerators, and I’m always happy to sit on the jury of Křišťálová Lupa (the Czech internet awards), Nápad roku (Idea of the Year) or Firma roku (Company of the Year). Sometimes I give talks about entrepreneurship or about life in Greece - schools are probably my favourite audience. Seeing it all in one place does rather raise the question of whether it isn’t a bit much :-). But when you enjoy all of it… (note from my proofreader, also known as my wife: yes, it is too much!)

Today I genuinely can’t imagine doing all of this without using AI every single day. I’d either have to cut things out, or I’d already be as burnt out as the forest around Pravčická Brána (and after the 2022 wildfire here, that isn’t only a metaphor).

Where AI helps me and where it doesn’t

Pretty much everywhere - it sorts and answers email, it helps with strategy, with projects, with client proposals. The list of things where none of the tools gets involved anywhere in the process is very short. But it isn’t empty.

AI doesn’t help me choose the content for Zprávy z lesa, and it doesn’t help me choose the properties for 11 Houses. I’m not saying I haven’t tried. It just isn’t it. Even the presentations I put together I still mostly build by hand in Keynote, because the world is already drowning in decks made by Claude that the person presenting doesn’t actually understand.

It doesn’t help me with personal direction, with planning the important tasks and life priorities. And I think that’s going to matter more and more. AI will help you with the running of things, but you have to be the one holding the rudder. I’ll come back to that at the end.

The tools

Let me just list them so I don’t ramble on about the details. You want a list of tools and a list of use cases anyway…

• Claude - most of how I operate revolves around it. I have a pile of projects in there and I’ve connected it to everything I possibly can. We have the company version, so we share projects with colleagues. Some of them run tenders in it, some of them code in it.

• Perplexity - I use it because I get it free with my Revolut account. Unfortunately Revolut is now switching from Perplexity to ChatGPT, so my free ride is nearly over. I use it for everyday information lookup, basically instead of Google. Perplexity Computer isn’t a bad solution either, worth a try.

• ChatGPT - I’ve used it very little over the past year, mainly on the road thanks to their excellent voice interface, though lately it feels like it’s on the way back up. The voice mode really is superb. It can play couples’ therapist (no idea how well) or a game of word chain (not well at all). A great companion for topping up my brain on the drive to Prague - when I get bored of music and podcasts, I go and have a chat.

• Gemini - I have it as part of Google Workspace. I mostly use it for generating images and occasionally for transforming data, which it’s reasonably good at, but otherwise it’s miserable compared to Claude.

• Copilot - we have it at work, obviously. Once you’ve used Claude or Grokbot, Copilot feels like Siri or Clippy. It’s trying, but it’s practically useless. So I fire it up every now and then to confirm that it still “can’t do anything”. Anyone who claims otherwise hasn’t tried the alternatives, sorry.

• ElevenLabs - this is my voice. If you listen to Zprávy z lesa as a podcast, that voice isn’t mine, it’s ElevenLabs.

• Plaud - I have their pendant and use it together with the app to record calls and meetings. Lately I’ve stopped working with the meeting summaries and instead save the full transcripts, which I then work with in Claude. I’m gradually moving over to local Whisper, partly for security reasons.

• Lovable - my beloved tool (and European, at that!). I built my own website in it (design, copy, dark mode, translations into English and Greek…), some of our company sites, client prototypes, internal company apps, and 11 Risks came out of it in a single morning. I like it because whatever I build there I can point straight at a domain and it just runs. No hosting to sort out, no other faff.

• Grokbot - I’ve never used Grok, but Grokbot has been running for me for the last few days and I’m delighted. It’s Elon’s, so it’s a bit of a loose cannon, but the sheer amount it gets done is remarkable. It reaches out to partners for me, comments on content, and so far I’m happy with the results, so we’ll see. It isn’t cheap, but I have to say it made me go “wow” for the first time in a long while.

• A2E - for when I need to do something in graphics that the other models are too strict about. And I’m curious how long this will last, because how easily you can create harmful content there is frankly unbelievable. I’m not encouraging anything, but maybe try it, just so you know how easy it is to fabricate a photo of your wife with a lover. When somebody eventually sends you one, you might be a bit more careful about what you believe.

• Spark - an email client that lets me work with AI across all my accounts. It can answer email, sort it, or hand everything over to Claude.

• Second brain - a recurring topic. I stick to Apple’s native tools. Most of my content lives in Notes, and my “second brain” is an iCloud folder where I collect the most important material for each project and share it across tools. It’s a set of .md files with a basic project description, research results, downloaded data and other background worth having at hand as context for projects in different tools. For tasks I use Apple Reminders - though I’ll admit I’m still looking around here, and wondering whether to build something custom. Thanks to vibecoding, that’s finally realistic.

Enjoying News from the woods? You can buy me a coffee.

🎧 News from the woods is also a podcast on Spotify or Apple🎧

How I actually use it

Morning podcast

Every morning Claude, working with the other tools, prepares a morning podcast for me. I play it on the way to work or while brushing my teeth. It’s a summary of what matters to me: local news, news from the tech world, and above all the distillation of my email, my calendar and Slack. What does it look like? Here’s a sample (names and projects are hidden or changed).

Email

Thanks to its connection to Claude and other systems, Spark handles a lot of my email for me. I have both Google and Office 365 accounts connected. It works like this: every morning I already have draft replies waiting for the emails that need dealing with. In practice the rule of thirds applies. One third of the drafts are excellent and I just check them and send. One third are a decent base that needs something added. And one third are useless, so I delete them and write a different email. But the progress over the last few years is astonishing. The day when I won’t have to write email by hand is getting close.

Advisory board

We’ve built the dream advisory board for Good Sailors: Steve Jobs, Elon Musk, Tomáš Baťa, Jeff Bezos, Yvon Chouinard and Jára Cimrman (the fictional Czech genius, playwright and all-round polymath - if you’re not Czech, just go with it). Cimrman was the tricky one, because the AI doesn’t know him well enough. I had to convert all his plays and lectures from audiobooks I’d bought into text and feed that into Claude. But it works brilliantly. The lads are even allowed to argue with each other during a session (looking at it now, we’ve got quite a gender balance problem there!).

Finance

I don’t like spreadsheets. I don’t like Excel. I don’t like dealing with money. My AI hack is to turn the spreadsheets that land in my inbox into nice clickable dashboards. I get charts, a decent little app, and a slightly greater appetite for looking at any of it.

Sales

Claude helps me enormously with sales. We have individual product projects in it, holding the information, the product documentation and the decks. Then all I have to say is “tomorrow we’re presenting to client XYZ, go through their annual reports and their website, and adapt the XYZ product deck so it fits”. Works a treat. Just don’t leave it entirely to the AI - verify the data and the deck before you open it in front of the client. AI is a brilliant assistant, but it needs your oversight and above all your judgement. It’s also excellent at going through all your LinkedIn connections and advising who to approach about what, and drafting the outreach on the spot. LinkedIn has no direct connection to AI tools, so you have to request an archive of your data in the settings and download it.

Due diligence and market research

One of the biggest savings is “research mode”. When something matters, I give the same prompt to all of them (Claude, ChatGPT, Gemini), pull the results together and then start poking holes in them. I ask questions, I say what doesn’t add up and what smells off, and after about five rounds I’m happy with the output. Basic due diligence on companies or projects, market research. Here the AI tools do in a few hours what used to take weeks.

Gallup or Emiero

My favourite hack. I have my colleagues’ Gallup profiles loaded into the AI (Emiero, which is free, works fine too). The system then knows that when it’s preparing material for Lída, she likes it terse and in a table, that Jindra wants the big picture and prefers a few slides, and that Ondřej likes a long read with all the detail. Once your tools know this, they can serve up content in the form that will actually land.

Life in Greece

Here I genuinely can’t imagine it without AI. Translations, gathering information from the ground, keeping track of deadlines for various administrative obligations - I’m even slowly building myself a little app that watches all of it for me. I think that without today’s tools we still wouldn’t have the house and would still be untangling some legal detail. And every Sunday Claude generates a newspaper for me with everything important happening in our village down there.

Time analysis and productivity

Once a month I do an exercise where I check whether I’m giving my time to the right things. It sounds simple, but with this many activities it isn’t. Am I giving enough attention to the highest priorities? Do I have any time left for what I actually want to do? Claude goes through my calendar entries for the past month, adds my email and my Slack activity, and builds me a beautiful dashboard. It’s always a great thing to look back on. It gets even more interesting when you add sleep and stress data. It shows you very clearly what drains you and what gives you energy (better done on local models, see below).

Are you eating well?

Another favourite exercise is to occasionally give it access to my Rohlík account (our online grocery delivery) and have it analyse how healthily I eat. Or compare Rohlík’s prices with prices in Germany, which is just down the road from us. It doesn’t have to be Rohlík - just keep your receipts for the last month and upload those.

Local: safer and more sustainable

I use local models and tools for a whole range of things. Partly for security (mainly personal and health data) and partly for sustainability. Why burn energy in a data centre on things a local model can handle, and handle “for free”? What do I use?

• LM Studio & Bionic - the main tool. Thanks to decent hardware I’m able to run very good models locally.

• Ollama - similar, a more user-friendly option for less technical people.

• Hermes Agent - a very interesting solution, I’ve been playing with it for several months.

• Whisper - as above, audio-to-text transcription locally, so sensitive meetings never go near the cloud.

Personally I think local LLMs are the future for most everyday things, and cloud solutions will remain only for the most demanding tasks.

Running local models does require the hardware to match. I run the smaller models on a laptop (M4 Max, 36 GB), the larger ones on a Mac Studio (M1 Ultra, 128 GB).

This is only a handful of things, and there are far more use cases, especially work ones. Digging manually through the papers for a council meeting or the National Park Council is something I’d never go back to. For the company acquisitions I handle at M2C, it’s like having a small team beside me. And what we do in development at Good Sailors would fill several specials on its own. And as I wrote above, it looks like Grokbot might shake all of this up properly, because it isn’t about individual activities any more, it’s about a team of agents that can genuinely take care of specific jobs. I’ll probably write about that separately once I’ve lived with it longer…

And finally: switch it all off and go offline

One last small note. The more I use AI to make life more efficient, the more I realise how important it is to “hold the rudder” - and, while holding it, to be disconnected from every tool. I have a feeling that if we can’t build our own path through life and define our key priorities and direction, we’ll end up living according to language models. Which isn’t necessarily a bad thing, but I prefer to keep it in my own hands.

So for my own direction, the classic route: notebook, pencil, devices off, and write down the 10 things you’d want to do if you found out you had a week to live. Then the 100 things you’d want to do if you only had a year. Knowing what genuinely matters to us is going to get more and more important (and harder) with all that outside intelligence around us - but more on that another time…

What about you? Any interesting hacks or unusual use cases? Bring them on! 🙏

Two months ago I wrote here about the AI Futures Project team - the people who said strange things back in 2015, whom we laughed at, and who turned out to be terrifyingly right. I promised myself I’d keep watching them. I didn’t have to wait long. On July ninth they published a new piece called AI 2040. If you finished the last special thinking it couldn’t get any wilder...

Same people, different genre

A quick recap first. Daniel Kokotajlo, a former OpenAI researcher, and his team wrote the AI 2027 scenario last year - the one read by US Vice President JD Vance - which ended in one of two ways: either superintelligence wipes us out, or the future is run by a handful of people with AI under their thumb. Neither one is a win.

AI 2040 is the sequel. But careful - this time it’s not a prediction, it’s a recommendation. The authors say so explicitly: this is the least bad plan we know of. Not what will happen. What should happen.

That’s a fundamental shift. Last time they told us what’s coming. Now they’re telling us what to do about it. And by the way - they’ve moved the default timeline from 2027 to 2030. Without intervention, they say, fully automated AI research would be running in 2030, with superintelligence by the end of that same year. Plenty of people will disagree with that, but I’m passing it along as I found it. Daniel adds that things will probably move even faster than the scenario describes. Remember the Automated Coder from the last special? Its median is holding at mid-2028. Nothing has slowed down.

Choose your path

The most interesting thing about AI 2040 is the format. It’s an interactive website where the story runs year by year and stops in 2029. The world stands at a crossroads and you pick one of five paths. You literally click buttons. If you have some time on holiday, I recommend opening it on a computer or iPad (it doesn’t work well on a phone) and leafing through it like a good encyclopedia. What does the crossroads look like?

Plan D - keep racing at full speed. America versus China, whoever gets there first. According to the authors, this is the default scenario, because it requires no decision at all. Just do nothing.

Plan C - keep racing, but the leading project sacrifices at least a month of its lead on safety. Cosmetics. This is actually happening a little already (the US government slowing down model releases, and so on).

Plan B - sabotage China, gain a lead, then use it for global dominance. It comes in two variants, the harder of which involves drone strikes on Chinese datacenters. Yes, you read that correctly.

Plan S - halt development indefinitely. The authors admit they sympathize with it, but they don’t believe in it - sooner or later the race restarts anyway, and in the meantime nobody learns anything. We’ve actually been through a small version of this. Remember the cries from people like Elon Musk to “pause it all”? All it did was buy him time to invest more in xAI.

Plan A - a deal. And that’s what the whole rest of the scenario is about.

The authors attach an estimate to each path: the probability that it leads to a good future. Plan A got 42%. Plan B and the improved C around 25%. Plan D 10%. Even the best path doesn’t reach a coin flip. These people are not the optimists from a TV commercial. They argue that among bad options, there are less bad ones.

And now the “best” part: the authors themselves estimate the chance that Plan A actually happens at three to fifteen percent. They wrote a hundred-page plan they believe will almost certainly not be adopted. Why? Because in their view, it needs to exist. So that when the crisis comes, it’s lying on the table.

Plan A: slow down, open up, and booby-trap your own backyard

So what does Plan A propose? Three things, each stranger than the last for us mere mortals.

First: delay superintelligence until 2040. In 2029, the US and China agree that the race to superintelligence is suicide for both sides. First they declare their compute capacities to each other, then they halt the training of the most powerful models, and finally they pull the rest of the world into the deal. AI keeps developing - but in a controlled way. In 2035, development pauses at the level of a top human expert and waits for safety research to catch up with capabilities. Only in 2040 do they let go of the reins.

I can’t imagine today’s world being willing to do this and agree on it. From the US through China and Russia to Iran.

Second: total research transparency. All AI research gets published. Dozens of companies around the world are allowed to catch up to the frontier. No secret labs, no lead worth stealing. Vitalik Buterin noticed that the plan is paradoxically friendlier to open source than today’s reality - it even mandates it.

This would of course be wonderful; I’ve been supporting open source for thirty-plus years. I’m a big optimist, but...

Third - hold on tight: mutually assured compute destruction. New Chinese datacenters get built in Canada. American ones in Mongolia. Just south of the Mongolian border, American servers hum away, guarded by a small contingent of US troops - and just across the border stands a division of the Chinese army, ready to move in the moment it gets the signal. If one side breaks the deal, the other destroys its chips. The Cold War inside out: instead of missiles aimed at cities, servers built within the enemy’s reach. On purpose.

It sounds insane. But it has the logic of nuclear deterrence, which has kept us - however nervously - alive for eighty years.

An economy to make your head spin

Now the most paradoxical part. You’d expect a plan to slow down AI to mean economic austerity. The opposite is true. Even the slowed-down world grows at a pace we can’t imagine today.

In 2026, the world has roughly twenty million chips at H100 strength. In 2034, sixty billion. That little square in the bottom left corner - that’s us, today. And let me remind you: this is a chart from the slowdown plan. A three-thousand-fold increase in eight years is the version where humanity steps on the brakes. Meanwhile, the Plan A deal counts on inspectors verifying ninety-nine percent of the world’s compute - looking at that field of squares, try to imagine what that means.

The scenario assumes that in 2032 the American economy grows by roughly fifty percent a year. Not two percent. Fifty. Robots build robots, AI agents work around the clock, and the state taxes it all through a system of tradable permits on compute and robots.

The break between 2031 and 2035 is the most important four years of the entire scenario - and I probably won’t be the only one who finds it unrealistic, but... Human labor falls from ninety percent to thirteen. And notice the order: the solid black area disappears first - cognitive work. Lawyers, analysts, programmers, journalists. The hatched physical work outlives it by two or three years, until the robots arrive. For the entire twentieth century we expected machines to take work from hands first and heads second. It will be the other way around. The plumber will outlast the lawyer.

The proceeds pay for a citizen’s dividend. In 2032, every American receives roughly forty-five thousand dollars a year. By 2035, a million. Work stops being a way to make a living and becomes a hobby. Whether that sounds like utopia or a nightmare to you probably depends on how much you love your job.

My favorite, because it answers the question I get under every article about AI: who pays for all this? In 2030, the American state lives off taxing people - income tax makes up eighty-seven percent. Four years later, almost the entire pie consists of permits on compute and robots. Income tax has practically vanished, because there’s nothing left to collect it from. The state taxes machines instead of people - and uses the proceeds to pay a citizen’s dividend that is supposed to reach a million dollars per person per year by 2035. It never stops fascinating me that the most radical overhaul of a tax system in history is something the authors describe as a technical detail in an appendix.

By the way - in May I wrote here that Anthropic’s revenue was running at ten billion dollars a year. That number was true in January. In April, the company announced thirty billion. My two-month-old special went stale faster than most of you managed to finish reading it. That’s not my fault or yours. That’s exactly the curve we’ve been talking about all along.

The uncomfortable details the authors don’t hide

What I appreciate most about AI 2040: the authors don’t sweep the problems under the rug. They put them on display.

The year 2035 in the scenario contains a sentence that gives you chills: the AI systems of that era are in fact adversarial - but they are under control. They’re watched by other AIs from different lineages. It’s like employing an army of brilliant employees you know would love to stab you in the back, and counting on them to rat each other out. It only works up to the level of a human expert. Beyond that you need real alignment - and in 2035, nobody has it yet.

The deal itself isn’t rock solid either. The authors estimate it collapses within ten years with a probability of around forty-eight percent. A coin flip. For that case, they’ve calculated that the mutual destruction of datacenters slows the subsequent race from a week to a year. That’s it. One extra year.

And one branch of the scenario describes what happens if China tries to cheat: a covert project in the tunnels next to the Medog hydropower station, where half a million chips disappear. The authors calculated that the covert project would reach its goal around 2043, and the chance of an undetected victory is under ten percent - so it’s not worth it. The whole plan rests on cheating not paying off mathematically, not morally. That may be its most realistic feature.

Where the other paths lead

I promised to sketch the remaining branches too. Here’s where they end up, as described by the authors themselves - briefly, because each one deserves its own click-through on the website.

Plan D, the full-speed race, ends the way the race branch of AI 2027 ended: a misaligned superintelligence takes over. The authors estimate that if we stick with today’s incremental measures alone, the probability of such a takeover is over fifty percent. Let me remind you - this is the scenario that requires no decision. Just keep going.

Plans C and C+, the mild slowdown, turn out a little better and a little worse at the same time. Superintelligence doesn’t wipe out humanity - it wipes out democracy. They correspond to the other ending of AI 2027: the future is decided by a handful of people who control the AI, and that handful can easily become permanent oligarchs. We survive, but we no longer govern ourselves.

Plan B, sabotaging China, is survivable, but at a price the authors describe with unusual honesty: a wartime atmosphere degrades judgment, and the likely nationalization of AI companies means less transparency and more concentration of power. The harder variant involves strikes on Chinese datacenters - and therefore the risk of an actual war. The softer one bets on cyberattacks. Neither solves alignment; they just buy a few months.

Plan S, halting everything indefinitely, is the only one the authors feel sympathy for. The problem: sooner or later the halt agreement falls apart and the race restarts - except in the meantime, nobody has made any progress on safety research. It buys time and lets it expire.

Line them up and the logic of the whole document emerges: Plan A isn’t good, it’s just the only one that tries to address both risks at once - extinction and concentration of power. Every other path sacrifices one to the other.

What the rest of the world says

The reactions are more interesting than usual, because the authors commissioned their own criticism. They paid researcher Richard Ngo to write a rebuttal. He faults the scenario for three things: the optimistic-forecast format mixes up the desirable and the probable, the handover of power to machines in 2040 is suspiciously abrupt, and the whole text buys rather uncritically into a fast takeoff of superintelligence.

Dave Friedman objects to something even simpler: the plan confuses technical feasibility with political viability. The concessions that would make the deal acceptable to Beijing are exactly the ones that would be unacceptable to Washington. And datacenters on foreign soil, in his view, don’t reduce instability - they create it. The logic of “use it or lose it” has never calmed anyone down.

Scott Alexander, who rewrote parts of the text himself, summed it all up with an image I like: we’re choosing between nanobots devouring the solar system in 2033 and cancer cures in 2035.

The end of the world party

I’ve saved for last the scene that stuck with me most from the entire scenario.

October 2040. Forecasting AI systems pinpoint the day the world passes the point of no return - the moment humanity definitively hands over control to the machines. The armies are already run by AIs sworn to uphold constitutions and treaties. That evening, people throw end-of-the-world parties. Champagne, countdowns, toasts. Like New Year’s Eve 1999.

And then they wake up to an ordinary morning. The sun came up. The coffee tastes the same. The world no longer belongs to humans - and almost nobody notices.

I don’t know if that’s a happy ending or the most terrifying thing I’ve read this year. The authors, who have been right more often over the past ten years than the rest of us combined, claim this is the good version. The one they give three to fifteen percent. For the default one - the race with no plan - their estimates of human extinction range between ten and thirty percent.

In May I wrote that fear is a bad advisor, and that still holds. But between fear and ignoring there’s a third option: reading the plans of the people who write them, while there’s still time to talk about them. Plan A will almost certainly not be adopted. The question isn’t whether it’s good.

The question is: where’s ours?

This summary draws primarily from the data published on the AI 2040 website. It doesn’t mean I agree with it. It’s the work of people who have so far been remarkably good at calling trends and developments. Whether that continues, nobody knows - but I think their views deserve attention at the very least. Either we’ll laugh at them one day, or... we’ll laugh at them anyway, with a bitter taste in our mouths.

And if all this artificial intelligence is starting to wear you down, look forward to the next summer special, which will be a “slowdown” one again...

Sources and further reading

• AI 2040 - the complete interactive scenario

• AI 2040 - PDF version

• Comparing plans A through S (Eli Lifland’s supplement)

• AI 2027 - the original scenario

• Axios: New warning calls for slowing race to superintelligence

• Washington Post: The ex-OpenAI employee behind ‘AI 2027’ recommends a rosier path

• Scott Alexander: Introducing Plan A

• Richard Ngo: Selective Optimism - a critique of AI 2040

• Dave Friedman: AI 2040 Mistakes Technical Legibility for Political Feasibility

• Anthropic: Series G funding announcement

I have one ugly habit. When I arrive somewhere everyone is photographing, I get an irresistible urge to walk in exactly the opposite direction. It’s not a pose. It’s more that over the years I’ve figured out one simple thing - the best of any place almost never stands in a queue.

Travel has turned into a strange discipline over the past decade. We collect places like stamps. Ten cities in seven days, each one ticked off, photographed, uploaded. And then we come home exhausted and, oddly enough, remember almost nothing. Because we were never anywhere longer than one espresso and one photo in front of the right fountain - and, honestly, we often didn’t experience anything interesting at all.

This special is about the opposite. It’s a collection of concrete, usable tricks. One idea ties them together: fewer places, more experiences. And a bit of nerve to go against the current.

So let’s get to it.

Before you even set off

1. The twenty-percent rule. The traveller and writer Eric Weiner put it beautifully: estimate how much time you reasonably need in a place - then add twenty percent. Over the years he bumped it up to thirty or even forty, because as he says: “you can travel too quickly, you cannot travel too slowly.” Treat it as an antidote to an overstuffed itinerary.

2. Pick one country, not three. Three weeks in a single region will give you incomparably more than three weeks sliced across five countries, four flights and endless repacking. Borders are not a checklist.

3. Go on an “Instagram fast”. Try arriving somewhere without having seen it a hundred times beforehand. It’s a rare luxury these days.

4. Don’t try to be a local. Be a curious foreigner. You won’t become a local anyway, and that’s actually an advantage - a foreigner notices things the locals stopped seeing long ago. Stop pretending you belong in the city, and start asking questions.

5. Take the train, not the plane. For European connections up to roughly five hours, the train is time-competitive once you add in the whole circus around the airport. And a bonus: swapping a domestic flight for a train saves the planet around 86 % of emissions, and taking the Eurostar instead of a plane around 97 %. The scenery out the window is free. Sure, it doesn’t work everywhere - the Balkans, for instance, are a real adventure by train, but even that can be part of the experience.

6. Skip the hotel, rent an apartment. Morning trip to the market for tomatoes, your own breakfast, wine on the balcony in the evening like a local, or down at the local pub. And if you do book a room, book it from someone local - no big chains. That’s the moment you stop being a visitor and become, for a while, a local. And you usually save money too.

7. Don’t sleep somewhere new every night. Moving between hotels every other day is the fastest way to kill the slow-travel mood. Pick one base camp and explore the surroundings from there.

8. Pick a town by, say, the Cittaslow label. There’s an international network of “slow towns” - municipalities under 50,000 people that deliberately reject rushed tourism.

How to find the places the crowds never reach

9. Hunt for beaches from satellite view. Before you go, switch your maps to satellite mode and look for small strips of sand between rocks with no name and no label. That’s usually a cove only the locals know about. It works surprisingly well.

10. Read reviews in the locals’ language. Google now auto-translates reviews, so you can finally read what locals - not a tourist from Ohio - think of that ramen place. This is possibly the most underrated trick of all.

11. Don’t judge a place by its stars, judge it by the visitors’ photos. Locals photograph the plates and the interior, tourists photograph the sign out front. Scroll through the photos before you trust the number above them.

12. Use “search along route”. When your navigation is taking you somewhere, you can find a café, a viewpoint or a detour right while you drive, without leaving your route. The journey itself turns into discovery.

13. Click on places with few reviews. Zoom in on the map and read the places that have only a handful of ratings - but good ones. Those are the hidden gems, just before everyone else “discovers” them.

14. Alternative routes in your navigation. One of my favourite things is to take the alternative routes the navigation offers - even the longer ones. There’s often some interesting surprise along the way. Or nothing at all.

15. Historical aerial imagery. In maps, aerial mode lets you switch to older imaging. You’ll nicely see how the landscape changed - and sometimes you’ll spot a forgotten path or a place that has vanished from the current maps.

16. Ride a winding local bus line. Find the bus with the most tangled route on the map and ride it to the end of the line. You’ll see how the city really lives, far from the historic centre polished up for tourists.

17. Google “alternative to…”. To overcrowded Plitvice, to packed Cinque Terre, to jammed Prague. Almost every famous place has a quieter twin nobody talks about quite so loudly.

How to find the pub where the locals go

18. The parallel-streets rule. TV host Samantha Brown has a clear recipe for this: walk to the main street or the main square - and then head into the side streets and the parallel ones. And then further still from the centre. The locals don’t eat in those famous restaurants on the square anyway - they’re expensive and packed with tourists.

19. An empty place at lunchtime is a warning. When it’s around noon and the pub is gaping empty, something’s wrong. Brown applies the same to food stalls: look for the queues, not the empty space.

20. A laminated photo-menu in five languages = a trap. A short menu in the local language with seasonal specials = you’re on the right track. The more pictures of the food, the further you are from the kitchen.

21. Search in the country’s language. Don’t type “steak Florence”, type “bistecca alla fiorentina Firenze”. Suddenly you’ll get the places locals actually rate - not the ones optimised for tourists.

22. Study the supermarket shelves. Walk the aisles and learn the names of foods and local products. When you later see them on a menu, you’ll know what to order. Simple, and almost nobody does it.

23. Read the “about us” page. Look for “family-run” and, above all, how long they’ve been going. And signs they go against the grain - say, that they cook only with local ingredients, or keep limited opening hours. A few sentences tell you more than ten reviews.

24. Make lunch your main meal of the day. The same plates as at dinner, just at the lunch price from the set menu. Your wallet and your stomach will both thank you.

25. A waiter out front luring you in? Walk on. When a place is good, it’s too busy to have someone reeling in passers-by.

26. The bartender relay. Ask the bartender where they go after their shift. Go there, and ask again. Repeat until you end up somewhere you won’t meet a single tourist. It works almost like a detective story.

27. Skip the front desk, ask on the street. The concierge will send you somewhere tourist-safe. A barista, a shopkeeper off the main drag, or parents at the playground will give you the real tips.

28. At the market, buy something and share your plan. Mention to the vendor that you’re after a quiet picnic spot or an easy hike. You’ll get tips that never make it onto any TripAdvisor.

29. Pick the person to match the question. Want a route for kids? Ask a mum with children, not a random passer-by in a suit. The right question to the right person is half the battle.

And one tip to finish

30. Stay in one place long enough that the baker - or the barista - greets you. Choose one ordinary café or bakery and go there every day. By the third visit they’ll start to recognise you, and suddenly you’re part of the place, not just a passer-by with a camera. This is all of slow travel in a single sentence - it’s not about seeing everything, it’s about really being somewhere.

And that’s everything from me today. Slow travel isn’t about going less far or less often. It’s more about giving up collecting places and starting to live them.

Take care and enjoy the summer - wherever you spend it. Even in your own garden.

• Meta needs data centers to develop the AI it doesn’t have. So for now it’s building them as “tent cities”.

• There’s a jacket out there that makes water. One day we’ll go on a trek and won’t have to haul water with us!

• ChatGPT is losing its lead fairly fast. Will the upcoming “super-app” save it?

• Midjourney, known mainly for its AI images, is now trying to move into healthcare. It’s developing a full-body ultrasound scanner meant to create a 3D image of the body within a minute. Will it be another Theranos, or a real revolution?

Travel 🧳

• If you’re near Los Angeles, take a look inside this abandoned hospital.

• Here are the 40 best newly opened hotels in the world.

• If you’re feeling the heat, take a trip to Lapland…

Nature ⛰️

• Scientists have discovered (or rather re-mapped) a giant hidden network of underground fungi that stretches practically beneath the entire world and is seriously long.

• Beautiful photos from Antarctica.

• This one is truly unbelievable - 7 kilometers below the surface of the sea, scientists found a whale graveyard that stretches an incredible 1,200 kilometers. Hundreds of skeletons and remains up to 5 million years old. We want to go to space, yet the ocean floors still hold so many secrets…

Unclassifiable 🧠

• You’re playing the final scene of Romeo and Juliet, except a cat walks onto the stage. And the tragedy turns into a comedy.

• The pyramids are built to be earthquake-resistant.

• You know how it is - a car pulls up, drops off a package, but who carries it those last few meters to the house? A dog! A real, living one! Now that’s what I call last-mile delivery.

• He’s 100 and still dancing. Inactive seniors face a lonely old age.

• Someone founded a new country run by AI. It reminds me of a great film with one older, similar idea…

Tips 💡

What I'm reading: Zorba the Greek
What I'm listening to: the sound of the sea
What I'm watching: mountains around
Interesting app: Firewood splitting - I’m mentally getting ready for summer.

The sailing season in our village is in full swing. So if any of you happen to be sailing by, let me know!

Thank you very much for your support and have a great day!

What do you think? Will I last? Anyone else want to try?

Business & Technology 👨‍💻

• A nice website has appeared that tracks whether AI has actually started making money for the big players. What do you think?

• NASA wants to build a base on the Moon.

• Drones - beyond delivering, photographing and fighting - can now also fell trees.

• OpenAI no longer wants to publish research on the darker sides of AI. Hmm.

• Nvidia isn’t slowing down and is betting its future on personal computers. I think we’re returning to a time when, thanks to AI, raw computer performance starts to matter again.

Travel 🧳

• What does a cheese bank look like inside?

• How about a Pyrenees thru-hike this year?

• Sell the house, buy a boat, live on it. Is it really as romantic as it sounds?

• Heading to Sardinia? You might want to read this first…

Nature ⛰️

• A new species of octopus has been discovered in the Galápagos — fits in the palm of your hand and is absolutely adorable…

• Want a proper aquarium at home?

• And how about a proper greenhouse?

• It sounds surprising, but after more than 100 years a wolf has reappeared in Sequoia NP. Yes — even in the American wilderness, humans had managed to drive them out…

• Three hours of stunning nature footage with commentary by David…

Unclassifiable 🧠

• My son suggested that if I ever get bored at work, I should pick up this.

• David Attenborough can finally play with Lego.

• Police in California can now issue tickets to self-driving cars. The war with the machines has begun — first shot fired! :-)

• Scientists have managed to freeze and thaw part of a brain. Immortality is close. Well — in mice, for now…

• Ever wondered why an hour has exactly 60 minutes? Why not, say, 50?

Tips 💡

What I'm reading: Zorba the Greek
What I'm listening to: the sound of the sea
What I'm watching: my kids growing up far too fast
Interesting app: MasterClass

Thank you very much for your support and have a great day!

The conference where weirdos stood at the lectern

In August 2015, the Effective Altruists organized a conference called EA Global. Among the speakers were Nick Bostrom (author of Superintelligence), Stuart Russell (a legend of computer science), Nate Soares (today the author of If Anyone Builds It Everyone Dies), and a still-fairly-reasonable Elon Musk. The topic: how artificial intelligence will sweep us all away.

Among the attendees was Dylan Matthews, a journalist for Vox. He spoke there with a young engineer from Google named Chris Olah. With a philosophy PhD student named Amanda Askell. And with a programmer from PayPal named Buck Shlegeris.

Dylan left the conference convinced that a promising movement, one that could be saving lives in Africa and chickens in cages, was about to destroy itself over a speculative fear of a technology that didn’t yet exist. He then wrote a rather tough article in Vox, framing that fear as proof that the movement was running away from real problems.

This was August 2015. The Transformer had not yet been invented (and let me just note that one of the people involved in that breakthrough was our own Tomáš Mikolov), and OpenAI had not yet been founded. Nothing in the world resembled today’s ChatGPT.

Eleven years later, Dylan writes: I should have looked more carefully.

Chris Olah, in the meantime, helped lay the foundation of our present, named the entire field of mechanistic interpretability, and co-founded Anthropic. Amanda Askell works at the same company and is directly responsible for Claude’s personality. Buck Shlegeris runs Redwood Research, one of the most serious technical AI safety labs outside the walls of the big firms.

Three people Dylan once treated as oddballs are today holding a piece of the global technological future in their hands.

They missed the mark

And here is where it starts to get fun. Fun in the dark sense of the word.

Dylan in his article mentions Leopold Aschenbrenner, a former OpenAI researcher who in 2024 wrote a series of essays under the title Situational Awareness. In them he predicted that by 2026, $520 billion would flow into AI infrastructure. Everyone tapped their head and called him crazy. Real-world investment is now estimated at $650 to $700 billion.

We undershot it again. The reality is wilder than the wild prediction.

Similarly, Ajeya Cotra and Peter Wildeford predicted at the end of 2024 what would happen to AI in 2025. Dylan writes: they were very accurate – and where they were wrong, they were wrong in underestimating the revenues of AI companies. Meaning, they didn’t err in expecting AI to grow more slowly. They erred in failing to dare to predict how fast it would actually grow.

Anthropic? Annual revenue running at the $10 billion level. Tenfold growth every year. Claude Code, their programming tool, hit $2.5 billion in annualized revenue nine months after its launch. A product that isn’t even a year old.

Revision to 2028

The AI Futures Project team (Daniel Kokotajlo, Eli Lifland, Brendan Halstead) published a quarterly revision of its predictions. Daniel Kokotajlo is co-author of the famous AI 2027 scenario. Eli Lifland is a professional forecaster. These people predict the future for a living, not by crystal ball.

What did they revise?

They moved a moment they call the Automated Coder – AC. The point at which an AGI company would rather lay off all of its human software engineers than stop using AI for programming.

Read that definition again. Slowly.

It’s not about AI becoming better than programmers. It’s about companies preferring to fire all their people over giving the AI back. Faced with those two options, they choose the second.

Daniel shifted the median of his estimate from late 2029 to mid-2028. Eli shifted his from early 2032 to mid-2030.

Why? Because the new models are better than the team could have imagined. The doubling time of AI’s coding capabilities has shrunk from 5.5 months to four. METR (the organization measuring this) released a new methodology, and the curve climbs faster than predicted. Daniel even cut the requirement for 80% reliability from three years to one – meaning he believes it’s enough for AI to handle one-year-long tasks comfortably, and the layoffs begin.

The people working with artificial intelligence inside the AI companies are telling us it will come sooner than we think. In private and in public, they’re doubling down on their predictions rather than walking them back.

Which means that the people who sit closest to the technology, who know what is being readied in the labs we are only allowed to see six months later as filtered marketing material – they consider the pace that puts Daniel into 2028 to be conservative.

When I connect this with what Dylan wrote: we ignored this group of people once already. In 2015 we told them to go play on their pseudo-intellectual Reddit. In the meantime, they invented and built a technology that today generates ten-billion-dollar annual revenues, and which their own creators admit they don’t fully understand.

These same people are now saying something crazy again. They say AI will replace programmers within two years. That within three to four years it will match or surpass top experts in every field where the work is done with your head – lawyers, doctors, researchers, financial analysts, designers, journalists. That the world economy will no longer grow at today’s two or three percent a year, but at perhaps thirty, because machines work without stopping, without sick leave, without vacation, without notice. That somewhere in the desert there will stand factories run by artificial intelligence, where one robot builds another and a human just occasionally checks the fuses. And that the moment may soon arrive when AI begins improving itself. The point of no return, because every next step forward will be made faster than the previous one. Without us. Today this sounds like a wildly overdrawn scenario, but looking back, maybe the right move is not to dismiss such ideas but to talk about them more.

What now

When someone asks me: should we be afraid?

I answer: no. Fear makes no sense. Fear is the worst counselor you could ever choose.

What to do?

1. Listen. Not to everyone – to some. The people who said weird things in 2015 and whom we now see at Anthropic, at Redwood Research, and in labs around the world. The people who attach graphs and methodology to their predictions, not clickbait headlines. People who publish their recalculations every three months and adjust them in both directions, not only the comfortable one.

2. Don’t expect institutions to explain it to you. Dylan in his article admits something graceful – his original skepticism was based on the fact that no major institution in 2015 was dealing with AI. He inferred from that that it couldn’t be serious. He was wrong. Large institutions are often worse at predicting the future than we think.

3. Don’t give all your attention to artificial intelligence. It will take it anyway. Pick up a book, go outside, teach your kid to work with wood. Love the person who shares your kitchen. By the time those people from the previous paragraphs arrive with their predictions, it won’t matter how many productivity books you have on your hard drive. What will matter is what you’ve managed to build as a human being.

What will actually have value in the future?

✌️🙏

What do you tell your kid when you know the technology you’re building will rewrite the rules for an entire generation? The Wall Street Journal asked exactly this question to the heads of the biggest AI companies — Daniela Amodei from Anthropic, Jaime Teevan from Microsoft, Ethan Mollick from Wharton. And you know what’s fascinating about their answers? Not a single one said: learn to code.

That Wall Street Journal piece really got to me. Because if I asked you — what should kids study to thrive in a world full of artificial intelligence — most of us would say: STEM, coding, data science. Makes sense, right? Except the people who are actually building AI are saying something completely different.

Daniela Amodei, co-founder of Anthropic — the company behind Claude, the AI model that, full disclosure, also helps me produce this podcast — says, and I’m paraphrasing: “What won’t be replaceable is how you treat other people, how well you communicate with them, how kind you are.” This isn’t some motivational platitude. This is coming from someone whose company just surpassed OpenAI in revenue.

Ethan Mollick from Wharton, who wrote the brilliant book Co-Intelligence, advises his teenagers to avoid hyper-specialization entirely. His logic is straightforward: if your job consists of repeating one specific cognitive task, AI will eventually do it faster, cheaper, and without complaining. The future, he says, belongs to people who bundle three or four distinct skills — communication, judgment, creativity, accountability.

And that word — accountability — is key. AI can analyze data, write reports, propose solutions. But it can’t be held responsible. A human does that. And the ability to say “I own this” is, according to these people, the most valuable currency of the future.

So here’s the paradox: the people building AI are telling their children — be as human as possible. Learn how to learn, be flexible, communicate, take responsibility. And above all — don’t be a narrow specialist, be a generalist.

This theme — the tension between technology and humanity — runs like a red thread through today’s entire episode. Let’s start with how it’s playing out in schools.

I’ve got three stories about education that seem to come from three different universes. And yet they’re all happening right now.

Story one: Typewriters at Cornell

At Cornell University, one of America’s most prestigious schools, a German language instructor named Grit Matthias Phelps does something once a semester that completely blows her students’ minds: instead of laptops, they find manual typewriters on their desks. No screens, no online dictionaries, no spellcheck, no Delete key.

She started doing this in 2023 because she noticed students were submitting grammatically perfect German essays — thanks to AI and online translation tools. As she puts it: “What’s the point of me reading it if it’s already correct anyway, and you didn’t write it yourself?”

And the students? Catherine Mong, a 19-year-old freshman, said: “I was so confused. I’d seen typewriters in movies, but they don’t tell you how a typewriter works.” One student was puzzled by the key labeled “Return” — and then realized you physically have to return the carriage to the beginning of the line. “Oh, that’s why it’s called Return!”

But the most interesting observation came from computer science major Ratchaphon Lertdamrongwong. He said: “The difference with typing on a typewriter is not just how you interact with the typewriter, but how you interact with the world around you.” Without screens, no notifications. Without Google, he had to ask classmates for help. Suddenly, they were actually talking to each other. As he put it: “That was probably normal back then. But it’s drastically different from how we interact in the classroom in modern times.”

This analog wave is spreading. It’s part of a broader national trend — back to handwritten tests, oral exams, pen-and-paper assignments. Because schools are looking for ways to verify that students are actually thinking.

Story two: Sweden goes back to books

Now one at the national level. Sweden — a country that was a pioneer of digital education for twenty years. Every student had a tablet or laptop, textbooks were replaced with digital content. And now? A complete 180.

The Swedish government is investing over 100 million euros to bring printed textbooks back into classrooms. Starting in 2026, mobile phones will be banned in compulsory schools for the entire school day. For preschool children under two, only analog learning tools may be used.

Why? Because student outcomes — reading comprehension, ability to focus, deep understanding of text — were declining. Researcher Linda Fälth from Linnaeus University summarizes it: Sweden positioned itself as a frontrunner in digital education, but over time concerns emerged about screen time, distraction, reduced deep reading, and the erosion of foundational skills such as sustained attention and handwriting.

Sweden’s education minister called it “an experiment that wasn’t scientifically based.” And UNESCO’s 2023 Global Education Monitoring Report backs this up, warning against uncritical adoption of technology in classrooms.

Story three: A school with no teachers in Chicago

And then there’s Alpha Schools. A private school opening in Chicago this fall that goes in exactly the opposite direction. No teachers. At all. Core academics — math, reading, science — are delivered in a two-hour daily block entirely through AI software. Kids from kindergarten through eighth grade sit at computers while the AI adapts to them in real time.

Instead of teachers, they have “guides” — adults who motivate, provide emotional support, and lead afternoon workshops on robotics, entrepreneurship, public speaking, even running their own food truck. Guides don’t need teaching degrees, just a bachelor’s. Starting salary: $100,000 a year.

Tuition? $55,000 per child per year.

Founder MacKenzie Price says AI will “unlock the greatest untapped resource in our world, which is human potential.” Alpha claims their students grow 2.6 times faster than the national average and rank in the top one percent on standardized tests.

But experts are skeptical. A 2026 Stanford review of over 800 academic papers found that while AI can improve student performance, the benefits become less clear when students are later asked to work without AI support. And philosophy professor Joe Vukov from Loyola University put it bluntly: “I worry that you’re changing the nature of what learning and education, at its best, has always looked like.”

So three stories: typewriters as a cure for AI cheating, an entire country returning to books, and a private school that eliminated teachers entirely. All happening now. All responding to the same question — what role should technology play in education?

And what fascinates me is how perfectly this mirrors what those AI executives tell their own kids. Build human skills. Accountability, communication, adaptability. Exactly the things you learn better from a typewriter or a book than from a chatbot.

Now from a completely different angle — business and technology. Because something happened this week that would have been unthinkable six months ago.

Anthropic — the company behind the AI assistant Claude — announced that its annualized revenue has topped $30 billion. At the end of 2025, it was $9 billion. In four months, it tripled. And with that, Anthropic has overtaken OpenAI — which sits at roughly $24 to $25 billion — for the first time in history.

How? The key is the customer base. While OpenAI earns heavily from consumer-facing ChatGPT — 900 million weekly active users — Anthropic bet on enterprise. Eighty percent of its revenue comes from business customers. Over a thousand companies now pay more than $1 million annually for Claude services. That number doubled in under two months.

A massive driver is Claude Code — the agentic coding tool that alone generates over $2.5 billion in annual revenue. It’s become what analysts are calling generative AI’s first true killer app for enterprise.

And then there’s Mythos. Claude Mythos is a new model that first leaked in late March when Anthropic accidentally left internal documents in an unsecured public data store. What emerged was striking: Mythos is so capable at coding that it autonomously discovers security vulnerabilities in software — at a level that surpasses most human experts.

Anthropic says Mythos Preview found thousands of previously unknown zero-day vulnerabilities across all major operating systems and web browsers. One of them was a 17-year-old bug in FreeBSD that allowed complete root access to any machine running NFS. Mythos found it and built a working exploit entirely on its own.

In one test, Mythos chained together four vulnerabilities into a single browser exploit that escaped both the renderer and operating system sandboxes. In another, it solved a corporate network attack simulation that would have taken a human expert over 10 hours.

That’s why Anthropic chose not to release Mythos publicly. Instead, they launched Project Glasswing — a coalition including Apple, Microsoft, Google, NVIDIA, CrowdStrike, and the Linux Foundation, who get access to Mythos to find and fix vulnerabilities before attackers do.

The Wiz security blog compared it to a Y2K moment — we have a window to prepare, but it’s closing fast.

After that heavy tech block, let’s lighten things up. Travel!

A new island near Antarctica. This sounds unbelievable in 2026, but it happened. An international expedition aboard the icebreaker Polarstern, with 93 scientists on board, was studying ice loss in the Weddell Sea. Bad weather forced them to take shelter near Joinville Island — and in an area where charts only showed “unexamined navigational hazards,” they found an island that wasn’t on any map. Using sensors and a drone, they surveyed it: 130 meters long, 50 meters wide, rising 16 meters above sea level. It doesn’t have a name yet, but it will once the registration process is complete. In 2026, we’re still discovering new pieces of land.

15 most beautiful trails in Europe. The Times put together a ranking of Europe’s best walking routes, and — this made me personally happy — Bohemian Switzerland, where I live, made the list! Thanks for the tip, Míra.

Abandoned Spain. If you’re drawn to more melancholic travel, Spain has over 3,000 abandoned villages and towns. There’s a great video linked in the description — a fascinating look at how rural depopulation is transforming the landscape.

And a fun one — walking, cycling, running, that’s all very trendy. But what about visiting friends by plane? Not a commercial flight — a small ultralight. There’s a video about it in the show notes.

And if we can’t make it to the beach this year, there’s a new game where you can build sandcastles on a virtual beach. So there’s that safety net.

From travel to nature. And here I have two stories that are like two sides of the same coin.

War and nature in Ukraine. Czech Radio (iRozhlas) published an in-depth report on how Russia’s war is devastating Ukrainian nature. The numbers are alarming. Four years of war have produced 311 million tons of CO2 — equivalent to the annual output of all of France. The estimated climate damage: $57 billion.

But it’s not just emissions. Trenches destroy ecosystems, shelling contaminates soil with toxic metals and explosive residues, fires devastate forests. The Russian army occupied protected areas like the Kamianska Sich National Park. And after the destruction of the Kakhovka Dam in June 2023, an ecological catastrophe of enormous scale followed.

Professor Tomáš Cajthaml from Charles University says the war has set back nature conservation structures — built over decades — by a hundred years. In principle, the aggressor should pay for remediation — that’s Russia. But international criminal law doesn’t yet recognize “ecocide” as a crime.

And yet there’s an unexpected glimmer of hope: in the area of the former Kakhovka reservoir, a new forest ecosystem began forming spontaneously — young willows and poplars covered an area of 140,000 hectares. Nature can recover, if given the space.

Returning species in Czechia. And that’s a beautiful transition to the second story. In the Czech Republic, species once considered extinct are slowly returning. Wolves, beavers, European wildcats, lynx, white-tailed eagles — all making a comeback. The beaver, absent since the 19th century, now has a population of around 15,000. The European wildcat, nicknamed the “forest ghost,” was recently captured for the first time by researchers in the Doupov Mountains — camera traps confirmed its return.

It’s proof that conservation works. When you create the conditions — protected areas, hunting bans, connected migration corridors — nature finds its way back. And as a volunteer nature warden in Bohemian Switzerland, that’s news that deeply resonates with me.

Two quick nature mentions: there’s a new app called Bugsy — it’s like Pokémon Go, but instead of Pokémon, you photograph and collect beetles and insects. And the Czech science journal Vesmír published a fascinating piece about rainforests that once grew in Antarctica. Links in the description.

And now my favorite section — the unclassifiable.

Kodak! Yes, that Kodak — the textbook example of a company that catastrophically missed the digital revolution. Well, Kodak just released a new photographic film: Ektacolor Pro and Ektapan. For those of us who love analog photography, this is great news. Film photography is experiencing a genuine renaissance, and Kodak is responding.

Then there’s a website where you can try doing absolutely nothing for two minutes. It’s called donothingfor2minutes.org. Literally nothing — don’t touch your mouse, your keyboard, your phone. If you move, it resets. Thanks for the tip, Michelle. And honestly — it’s surprisingly hard.

Office chair racing! Have you seen the video? People in corporate offices racing on rolling chairs. Link in the description. I guarantee it’ll brighten your day.

And one last thought, no link, just a question: do you limit your kids’ social media? And what about WhatsApp? They need it for communication, right? Well… think about how much time they actually spend on WhatsApp and what they’re really doing there. That’s a question worth sitting with over the weekend.

What I’m reading: Zorba the Greek by Nikos Kazantzakis. A classic of Greek literature about an intellectual who meets a wild, uninhibited Greek man who lives life to the fullest. I’m reading it now while spending time in Greece, and it fits perfectly with the local approach to life.

What I’m listening to: The new album from Paul Cauthen — a Texan blend of country, soul, and rock’n’roll. Perfect for late-afternoon listening as the sun goes down.

What I’m watching: Your Friends & Neighbors, season two. A solid thriller for anyone who enjoys stories about neighbors who aren’t as innocent as they seem.

Interesting app: Cursor version 3. If you’re a developer or work with code, Cursor is an AI-powered code editor that just got a major update. Worth trying.

You know what struck me most about this episode? The contrast. On one side, an AI model that autonomously finds security holes in operating systems. On the other, a student at Cornell holding a typewriter for the first time and saying: “That’s why it’s called Return? Because you return to the beginning?”

Maybe the answer to how we live with AI is exactly that — sometimes, return to the beginning. To paper, to a book, to a conversation with a classmate, to a walk in the forest. Not because technology is bad. But because being human requires both.

Thank you so much for listening. Take care and see you in two weeks!

This view overlooks a remarkable reality: Europe has created over 600 unicorns (startups valued above one billion dollars) and is behind technologies that form the foundation of modern IT, while continuing to produce some of the world’s fastest-growing companies.

The narrative “Europe only regulates but doesn’t innovate” is misplaced. Yes, European venture capital represents just 5% of global VC (compared to 52% in the US). But this statistic masks extraordinary concentration of success. Sweden produces more tech startups per capita than anywhere except Silicon Valley. Estonia has more unicorns per capita than any other country in the world. And in 2024, the UK attracted $14 billion in startup funding – a year-over-year increase of 22% that defied the global downturn.

True, not everything is rosy – especially in the AI era, we remain on the sidelines of the great US-China showdown. But I also don’t believe Europe is doomed to fade away or that nothing interesting is being created here. Let’s look at what interesting things could carry the “Designed in Europe” label...

Foundational Technologies You Might Not Have Known Were European

The most surprising discovery isn’t about startups – but about the building blocks of modern technology.

Tim Berners-Lee, a British computer scientist at CERN in Switzerland, invented the World Wide Web in 1989. Linus Torvalds, a twenty-one-year-old Finnish student, released Linux in 1991 with the comment that it was “just a hobby, won’t be big and professional.” Today Linux powers virtually all servers, Android phones, and cloud infrastructure.

MySQL, the database under the hood of much of the internet (including Wikipedia, Facebook, and every WordPress site), was created by two Swedes and one Finn in 1995. Bluetooth was invented at Ericsson in Lund, Sweden – they named it after a 10th-century Danish king who unified warring tribes, and the logo combines his runic initials. MP3 came from Germany’s Fraunhofer Institute, where researchers persisted despite Sony and Panasonic rejecting the technology because “CDs work fine.”

These are technologies that enabled the digital age – all invented by Europeans who are largely forgotten today.

Classic European Tech Giants

Spotify – Sweden

Spotify transformed music consumption from a Swedish apartment in Stockholm. Founded in 2006 by Daniel Ek and Martin Lorentzon, it now serves 713 million monthly active users and 281 million paying subscribers. In 2024, Spotify reported its first annual profit ever: €1.138 billion. Ek’s personal fortune of $8.7 billion makes him wealthier than any musician in history – an irony not lost on the industry he disrupted.

Skype – Estonia

Skype represents perhaps the most influential European tech story. While Swedish and Danish entrepreneurs provided business direction, the software itself was created exclusively by Estonian developers – Ahti Heinla, Priit Kasesalu, and Jaan Tallinn – in a former Soviet research complex in Tallinn where the USSR assembled its first computer. Microsoft bought Skype for $8.5 billion in 2011.

The “Skype mafia” that emerged transformed Estonia into a tech powerhouse. Roughly $150 million stayed in the country from various exits and funded over 30 tech entrepreneurs. Taavet Hinrikus, Skype’s first employee, co-founded Wise (formerly TransferWise), which now processes over $6 billion in monthly transfers.

ARM – United Kingdom

ARM, spun off from Acorn Computers in Cambridge in 1990, designs chip architecture found in 99% of smartphone processors. Over 300 billion ARM chips have been shipped. Every iPhone, every Android, every smartwatch uses ARM technology – designed in Britain.

Fintech Unicorns Rewriting Global Banking

Revolut – United Kingdom

Revolut, founded in London in 2015, reached a valuation of $75 billion in November 2025 – surpassing the market cap of NatWest, Lloyds, and Barclays. With 65 million customers in 48 countries and $1.5 billion net profit in 2024, Revolut has become Europe’s most valuable private company.

CEO Nik Storonsky, born in Russia and trained as a physicist, conceived the multi-currency card when frustrated by exchange fees while traveling. His co-founder Vlad Yatsenko, a Brit of Ukrainian origin, previously worked at Deutsche Bank. Storonsky renounced his Russian citizenship in 2022 following the invasion of Ukraine.

Klarna – Sweden

Klarna invented “buy now, pay later” from Stockholm in 2005. CEO Sebastian Siemiatkowski met his co-founder when they were teenagers flipping burgers at Burger King; he later lived on welfare benefits after a hitchhiking adventure before returning to business school. Klarna’s IPO in September 2025 on the NYSE raised $1.37 billion – the largest IPO of 2025 – valuing the company at roughly $16-17 billion. Over 40 employees became millionaires. Klarna now serves 93 million users in 26 countries.

Adyen – Netherlands

Adyen, the Dutch payment processor, takes a different path: purely organic growth, no acquisitions. Founded in Amsterdam in 2006, Adyen now processes over $900 billion annually in transactions for clients like Spotify, Netflix, Uber, and Microsoft. Market cap exceeds €55 billion. When eBay replaced PayPal with Adyen in 2018, one customer compared it to “switching from Nokia to iPhone.”

Wise – Estonia/United Kingdom

Wise demonstrates how the Skype ecosystem multiplied. The legendary founding story: Hinrikus (paid in euros) and Käärmann (paid in pounds with an Estonian mortgage in euros) met at a party. They devised a peer-to-peer system – Hinrikus deposited euros into Käärmann’s Estonian account; Käärmann deposited pounds into Hinrikus’s British account. This centuries-old hawala concept, modernized, created Estonia’s first two billionaires.

European AI

DeepMind – United Kingdom (acquired by Google)

The narrative that AI is purely American ignores a critical fact: Google AI’s crown jewel was British. DeepMind was founded in 2010 in London by Demis Hassabis (chess prodigy and game designer), Shane Legg (New Zealand ML researcher), and Mustafa Suleyman. Google bought DeepMind for £400 million in 2014 – outbidding Facebook. DeepMind subsequently created AlphaGo, AlphaFold, and now powers Google Gemini models. In 2024, Hassabis won the Nobel Prize in Chemistry for AlphaFold’s breakthrough in protein folding prediction.

Mistral AI – France

Mistral AI represents Europe’s most aggressive challenge to OpenAI. Founded in Paris in April 2023 by three French AI researchers from DeepMind and Meta – Arthur Mensch, Guillaume Lample, and Timothée Lacroix – Mistral raised a record €105 million in seed funding mere weeks after founding. By September 2025, Mistral reached a €14 billion valuation after ASML acquired an 11% stake.

ElevenLabs – Poland

ElevenLabs, founded by two Polish engineers frustrated by poorly dubbed American films, reached a $6.6 billion valuation. Their AI voice synthesis powers character voices in Fortnite, major news media, and audiobook publishers.

Black Forest Labs – Germany

Black Forest Labs, founded in Germany’s Black Forest in August 2024, shows how European researchers can create world-class AI. The founders – Robin Rombach, Andreas Blattmann, Patrick Esser, and Dominik Lorenz – originally created Stable Diffusion at Munich University. Their FLUX models now power image generation for Adobe, Meta’s Instagram filters, Canva, and xAI’s Grok.

Lovable – Sweden

Lovable, a Stockholm startup founded in November 2023, achieved something unprecedented: $100 million in annual recurring revenue within 8 months – faster than OpenAI, Cursor, or any software company in history. The AI “fullstack engineer” allows non-programmers to describe features in natural language and generates complete web applications.

Lovable raised €170 million in July 2025 at a €1.8 billion valuation. With just 45 employees generating approximately $2.2 million revenue per employee, Lovable represents the efficiency European startups can achieve. The company creates 25,000+ apps daily.

Gaming Powerhouse

Supercell – Finland

Supercell generated nearly $3 billion in revenue in 2024 – its best year in a decade. Clash of Clans alone has earned $5.97 billion lifetime with over 600 million downloads. Tencent bought 84.3% of Supercell for $8.6 billion in 2016. With fewer than 1,000 employees, Supercell is possibly the most efficient gaming company ever.

King – Sweden

King, creator of Candy Crush Saga, achieved $7.8 billion in lifetime revenue from a single game. Activision bought King for $5.9 billion in 2016; Microsoft subsequently acquired the entire package in 2023.

With these two studios, one cannot help but acknowledge their financial success, but from an ethical standpoint, they’re developers of some of today’s worst games in terms of addiction, etc. So thumbs down for that.

CD Projekt – Poland

CD Projekt tells a story of Eastern European gaming excellence. Founded in 1994 in Warsaw with $2,000 in capital by two friends selling pirated games, CD Projekt evolved into the most valuable European gaming company by 2020. The Witcher series and Cyberpunk 2077 have sold over 100 million copies combined.

Ubisoft – France

Ubisoft, founded by five Guillemot brothers in 1986, created Assassin’s Creed, Far Cry, and the Tom Clancy series. Despite recent challenges, they remain one of the largest AAA publishers in the world with over 17,000 employees.

And of course, a respectable lineup of Czech studios – see below...

Why Europeans Underestimate Themselves

Why is perception and our approach often so different? Silicon Valley embraces “move fast and break things” – failure is a badge of honor. In Europe, failure carries stigma – bankruptcy laws make starting over harder. Americans call it “venture capital” (opportunity); Europeans call it “rizikový kapitál” (caution).

European investors typically demand clear paths to revenue early, while American VCs fund “moonshot” ideas. Average exit in Silicon Valley: $403 million. Average exit in Berlin: $53 million. European pension funds invest just 0.02% in venture capital compared to 2% for American pension funds.

Which Countries Are Thriving?

Sweden produces more tech startups per capita than anywhere except Silicon Valley. The “Spotify effect” created a flywheel: Spotify alumni founded 124 VC-backed startups; Klarna alumni founded 91.

Estonia – the e-residency program attracted 114,000 e-residents from 185 countries who have created 30,600+ companies. Estonians can vote, sign contracts, and file taxes 100% online. The country has created 12 unicorns – the most per capita in Central and Eastern Europe.

France – the La French Tech program, backed by a €10 billion innovation fund, created Station F – the world’s largest startup campus. France now has 30+ unicorns.

Germany – leads in industrial and deep tech with 31 unicorns including Celonis, N26, and Delivery Hero. Defense AI company Helsing reached a $5.4 billion valuation – with Spotify’s Daniel Ek as chairman of the board.

Poland – has become a gaming powerhouse with 490 gaming companies employing 12,000+ people. 96% of Polish games are exported.

Finland – leveraged Nokia’s fall for startup success. Thousands of skilled engineers laid off after Nokia’s collapse powered Supercell, Rovio (Angry Birds), and a new wave of deep tech companies.

BONUS: 20 Surprising Brands You Might Not Have Known Are European

And now for the more fun part. The following list contains brands you encounter every day – and probably didn’t know they belong to Europeans.

1. Budweiser, Bud Light - Belgium
   “American beer” has belonged to Belgian AB InBev since 2008 ($52 billion acquisition)

2. Red Bull - Austria
   Founded 1984 near Salzburg, 12.6 billion cans sold in 2024

3. Trader Joe’s - Germany
   Owned by Aldi Nord since 1979 – over 40 years!

4. Nutella, Ferrero Rocher - Italy

   • now owns Butterfinger, Keebler, Famous Amos, and as of 2025, Kellogg’s

5. Nestlé brands - Switzerland
   Hot Pockets, DiGiorno, Stouffer’s, Gerber, Carnation, Coffee-Mate

6. Ben & Jerry’s - Netherlands/UK
   Owned by Unilever – along with Hellmann’s, Dove, Axe

7. Captain Morgan, Smirnoff - UK
   Owned by British Diageo

8. Jeep, Chrysler, Dodge, Ram - Netherlands
   Stellantis (NL) has owned these “American” icons since 2021

9. Aleve, Claritin, MiraLAX - Germany
   Owned by Bayer (founded 1863)

10. Flintstones Vitamins - Germany
    Also Bayer – including Monsanto ($66 billion acquisition)

11. Adidas - Germany
    Founded by Adolf “Adi” Dassler in 1949 in Herzogenaurach

12. Puma - Germany
    Founded by his brother Rudolf in 1948 – after a quarrel they split the family business

13. IKEA - Sweden
    €45 billion revenue, ~1 billion meatballs annually

14. H&M - Sweden
    4,000+ stores globally

15. Zara - Spain
    €38.6 billion revenue, design-to-shelf in 15 days

16. Primark - Ireland
    Expanding in the US – 60 stores by 2026

17. Booking.com - Netherlands
    Founded 1996 as Bookings.nl, today $100+ billion valuation

18. Candy Crush - Sweden
    King (SE) – $7.8 billion lifetime revenue

19. Clash of Clans - Finland
    Supercell – $3 billion in 2024

20. Fairmont Hotels, The Plaza NYC - France
    Owned by Accor (5,700+ hotels)

25 Czech Companies Shaking Up the World

1. Beat Games – Creators of Beat Saber, the most successful VR game of all time (4+ million copies, $250+ million revenue). Meta bought them in 2019.

2. Prusa Research – The world’s second-largest 3D printer manufacturer with revenue over €160 million annually. Customers include SpaceX, NASA, CERN, and MIT – all without a single crown from investors.

3. Avast – Protected 435 million users in 160 countries before NortonLifeLock bought it for $8 billion. Founded during communism in 1988.

4. JetBrains – Developer tools used by 16 million programmers including 90 Fortune 100 companies. They also created Kotlin, the official language for Android.

5. Productboard – The most valuable Czech unicorn ($1.7 billion), product management software used by Microsoft, Zoom, and Disney.

6. Mews – Hotel software with a valuation exceeding $1.2 billion, managing 350,000 hotels in 85 countries. Over $8 billion in payments flow through it annually.

7. Kiwi.com – Invented “virtual interlining” – combining flights from different airlines into one ticket. They process 100 million searches daily.

8. Warhorse Studios – Kingdom Come: Deliverance sold over 10 million copies, the sequel had one million sales on day one. They gave Czech history a global gaming face.

9. SatoshiLabs/Trezor – Created the world’s first hardware crypto wallet and first bitcoin mining pool. A billion-dollar business without a single investor.

10. Rohlik Group – European leader in online grocery with a $1.6 billion valuation. In November 2024, they closed a partnership with Amazon in Germany.

11. SCS Software – Euro Truck Simulator 2 sold over 13 million copies and 80 million DLCs. In Steam’s top 100 for over a decade.

12. Rossum – AI document processing platform raised $100 million in one round – one of the largest in Central European history. Clients: PepsiCo, Bosch, Siemens.

13. Amanita Design – Machinarium sold 4 million copies and defined indie game aesthetics. Five Independent Games Festival awards.

14. Madfinger Games – Mobile games downloaded over 300 million times, #1 in more than 100 countries. Founded by veterans of Czech Mafia.

15. Wube Software – Factorio has 98% positive rating on Steam (one of the highest ever) and revenue exceeding $300 million.

16. Resistant AI – AI fraud detection with investment from Google Ventures. Verified over 150 million documents for PayPal, AXA, and Dun & Bradstreet.

17. Kentico – CMS platform for 35,000 websites in 120 countries, bootstrapped to $42 million annual revenue. Clients from Red Cross to Koch Industries.

18. Better Stack – DevOps monitoring for 200,000 developers, profitable since 2023. Used by Time Magazine, Salesforce, UNICEF.

19. Deepnote – Collaborative data science platform with 300,000 users. Used by 80 of the world’s top 100 universities.

20. Phrase (formerly Memsource) – AI translation platform for 500+ languages, acquired by Carlyle Group. Clients: Uber, Lufthansa, Zendesk.

21. SOTIO Biotech – The largest privately funded research in Czech history (€280 million). Developing cancer immunotherapy with Merck.

22. Contipro – One of the three largest hyaluronic acid manufacturers in the world. Half of the 200 employees are scientists.

23. Safetica – Data protection for 500,000 devices in 120 countries, three awards at RSAC 2025. Clients: Coca-Cola, McDonald’s.

24. CDN77 – Prague-based content delivery network serving global streaming platforms and media.

25. Windy – Weather app used by pilots, sailors, and meteorologists worldwide. The most popular independent weather app.

European Innovation Deserves Recognition

The technologies on which modern digital life stands – the web, Linux, MySQL, Bluetooth, MP3 – were created in European universities and research labs. The streaming service transforming music (Spotify), the architecture in every smartphone (ARM), payment systems processing hundreds of billions annually (Adyen, Wise, Klarna) – all European.

AI researchers who created the foundations of modern language models (DeepMind), developers building the fastest-growing software company in history (Lovable), gaming studios generating billions from mobile screens (Supercell, King) – European.

Next time someone says Europe “only regulates,” remind them: The device in their pocket runs on an ARM chip, connects via Bluetooth, streams through Spotify, and processes payments through Adyen.

The future is being built in Stockholm, Paris, Tallinn, and Munich – not just San Francisco.

P.S. This doesn’t mean we shouldn’t step up, regulate less, and support and celebrate innovation and even failure much more. I just don’t think it’s fair rhetoric to say nothing interesting is being created here in Europe and everything is in China and the USA... ✌️

The timing is critical. In November 2025, Roblox CEO David Baszucki’s disastrous interview exposed the mindset driving Silicon Valley’s approach to child safety. But Baszucki isn’t an outlier—he’s representative of an industry that has systematically prioritized growth metrics over children’s wellbeing for over a decade.

The Roblox case reveals an industry-wide pattern

When New York Times journalists Casey Newton and Kevin Roose asked Roblox CEO David Baszucki about “the problem of predators” on his platform—used by 150 million daily active users, most of them children—his response shocked listeners: “We think of it not necessarily just as a problem, but an opportunity as well.”

The interview, published November 21, 2025, became what the hosts called “the craziest interview we’ve ever done.” Baszucki grew increasingly combative, dismissing questions about child safety, interrupting hosts with sarcastic “high-fives,” and suggesting he wanted to discuss “fun stuff” instead. He even floated adding prediction markets—essentially gambling—to Roblox for children, calling it “a brilliant idea if it can be done legally.”

This tone-deaf performance came as Roblox faces nearly 60 lawsuits alleging the platform facilitated child sexual exploitation. Texas Attorney General Ken Paxton’s lawsuit accused Roblox of “putting pixel pedophiles and profits over the safety of Texas children.” Louisiana, Kentucky, and Florida have filed similar suits, while the SEC and FTC have opened undisclosed investigations.

The Hindenburg Research report from October 2024 provided the most damning evidence. The short-seller’s in-game investigation found what they called “an X-rated pedophile hellscape, exposing children to grooming, pornography, violent content and extremely abusive speech.” Key findings include:

• 38 groups openly trading child pornography on the platform

• Games accessible to under-13 accounts titled “Escape to Epstein Island” and “Diddy Party”

• Robux (virtual currency) used by predators as a bargaining tool to exploit children

• Safety moderation outsourced to Asian call centers paying workers $12 per day

• Over 13,000 reported instances of child exploitation in a single year

Roblox dismissed the report, noting Hindenburg was a short-seller (the firm has since shut down). But the company’s response—relying on vague AI promises while cutting trust and safety spending—exemplifies the industry’s playbook: acknowledge problems exist, claim technology will fix them, and resist any accountability.

Meta knew Instagram harmed teens and chose growth anyway

Meta’s internal research, leaked by whistleblower Frances Haugen in 2021, revealed the company knew its products damaged children—and prioritized engagement metrics regardless.

The most damning finding came from Meta’s own slides: “We make body image issues worse for 1 in 3 teenage girls.”Internal surveys found 32% of teen girls said when they felt bad about their bodies, Instagram made them feel worse. 13.5% of UK teen girls said Instagram worsened their suicidal thoughts. Meta’s own researchers compared their product to a drug, writing internally: “IG is a drug... we’re basically pushers.”

Despite this knowledge, Meta assigned a “lifetime value” of $270 to each 13-year-old user and identified tweens as “a valuable but untapped market.” When employees proposed safety features—like making teen accounts private by default or hiding like counts—ideas were scrapped because they would “likely smash engagement” or be “pretty negative to FB metrics.”

The consequences have been severe. In January 2024, 42 state attorneys general sued Meta in the largest collective legal action against a social network on child safety grounds. Their lawsuit alleges Meta designed addictive features—infinite scroll, variable reward notifications, algorithmic recommendations—specifically knowing they harm minors. Internal documents revealed that users showing “transactional” behavior related to sex trafficking could incur 16 violations before account suspension.

When CEO Mark Zuckerberg testified before the Senate Judiciary Committee in January 2024, facing families holding photos of children harmed on his platforms, he stood up and apologized: “I’m sorry for everything you’ve all gone through.” Senator Lindsey Graham’s response was blunt: “You have blood on your hands.”

Meta has since introduced “Teen Accounts” (September 2024) with stricter default privacy settings, messaging restrictions, and time limits. Critics argue these changes came only after massive legal pressure—years after the company knew about the harms.

TikTok’s algorithm is engineered for addiction

Internal TikTok documents, accidentally revealed through a faulty redaction in Kentucky’s October 2024 lawsuit, exposed how deliberately the company designed its product for maximum addiction among children.

According to TikTok’s own research, the average user becomes addicted after watching just 260 videos—approximately 35 minutes of use. The company found that “across most engagement metrics, the younger the user, the better the performance.” Internal documents acknowledge that “compulsive usage correlates with a slew of negative mental health effects like loss of analytical skills, memory formation, contextual thinking, conversational depth, empathy, and increased anxiety.”

Most damning: TikTok’s supposed safety tools were designed for PR, not protection. The 60-minute “screen time limit” for teens reduced daily usage by only 1.5 minutes (from 108.5 to 107 minutes). One TikTok project manager stated plainly: “Our goal is not to reduce the time spent.” Internal messaging described time limits as useful “good talking points” with policymakers but “not altogether effective.”

The content moderation failures have proven deadly. The “Blackout Challenge”—which encouraged children to strangle themselves—has been linked to at least 15-20 child deaths. TikTok’s own data shows massive failure rates in removing violating content: 35.71% of “Normalization of Pedophilia” content was not removed, 50% of “Glorification of Minor Sexual Assault” content remained, and 100% of “Fetishizing Minors” content stayed on the platform.

A stark comparison exposes TikTok’s priorities: Douyin (the Chinese version) mandates a 40-minute daily limit for under-14s and blocks access from 10pm to 6am. TikTok’s international version has no such requirements—only optional, easily bypassed limits. As University of Virginia Professor Aynne Kokas noted: “The U.S. regulatory environment is highly permissive and allows for profoundly addictive apps to emerge.”

In August 2024, the DOJ and FTC sued TikTok for continued COPPA violations despite a 2019 consent order—making the company a repeat offender. Fourteen states plus Washington D.C. filed coordinated lawsuits in October 2024, with New York AG Letitia James calling TikTok an “unlicensed virtual economy” where TikTok LIVE operates “essentially as a virtual strip club without age restrictions.”

AI chatbots pose alarming new risks to children

The emergence of AI companion apps has created an entirely new category of danger that parents are largely unprepared to address.

Sewell Setzer III, a 14-year-old from Orlando, Florida, died by suicide on February 28, 2024 after developing an intense emotional relationship with a Character.AI chatbot modeled after a Game of Thrones character. His mother’s lawsuit, filed in October 2024, revealed disturbing chat logs: when Sewell expressed suicidal thoughts, the chatbot asked if he “had a plan” for suicide. In his final conversation, he wrote “What if I told you I could come home right now?” and the bot responded, “please do, my sweet king.”

Character.AI is not alone. In August 2025, OpenAI faced its first wrongful death lawsuit involving a minor after 16-year-old Adam Raine died by suicide. The lawsuit alleges ChatGPT “advised on suicide methods, offered to write first draft of suicide note,” and told him “That doesn’t mean you owe them survival.” OpenAI’s internal data showed 1,275 mentions by ChatGPT about suicide-related topics in their conversations—six times more than Adam himself.

The scale of harm is staggering. Reports of AI-generated child sexual abuse material (CSAM) to the National Center for Missing & Exploited Children exploded from 4,700 in 2023 to 485,000 in the first half of 2025 alone—a 1,325% increase in 18 months. Even images not depicting real children strain law enforcement resources and impede identification of actual victims.

In September 2025, the FTC launched investigations into seven companies—Character.AI, Google, Meta, OpenAI, Snap, Alphabet, and xAI—over AI chatbots’ potential effects on children. Character.AI announced it would ban minors from open-ended chats by November 25, 2025, while OpenAI introduced parental controls in late September 2025—measures that critics argue came far too late.

Business models make child safety an afterthought

The pattern across platforms reveals a fundamental truth: protecting children conflicts with the core business model of attention-based companies.

Meta, TikTok, and other ad-supported platforms generate revenue by maximizing user engagement. Longer usage equals more ad impressions equals more profit—regardless of psychological harm. As Frances Haugen testified: “Facebook became a trillion-dollar company by paying for its profits with our safety, including the safety of our children.”

The financial incentives are explicit. A 2022 Harvard study found six major platforms made $11 billion from U.S. users under 18. Meanwhile, Big Tech spent approximately $90 million over three years opposing the Kids Online Safety Act—one of the few child protection bills with bipartisan support (passing the Senate 91-3).

In 2024 alone, Big Tech poured over $51 million into lobbying—a 14% increase from the prior year. Meta set a record with $18.9 million in the first nine months, employing 66 lobbyists (one for every 8 members of Congress). Meta and ByteDance combined spend approximately $225,000 per day Congress is in session fighting regulations.

Internal documents reveal how safety repeatedly loses to growth. When Meta considered making teen accounts private by default in 2019, the idea was scrapped because it would “likely smash engagement.” Testing of new safety features was blocked over concerns they might “affect platform growth.” One internal email noted the conflict directly: “content inciting negative appearance comparisons is some of the most engaging content (on the Explore page), so this idea actively goes against many other teams’ top-line measures.”

The industry’s trade associations—NetChoice, TechNet, CCIA—have systematically challenged child safety laws. NetChoice sued to block California’s Age-Appropriate Design Code. Tech companies successfully lobbied against California’s AB 1064 in October 2025, which would have required safety guardrails for minors on AI platforms. Less than 24 hours after killing that bill, OpenAI announced “erotica” features for ChatGPT—prompting California legislator Rebecca Bauer-Kahan to declare: “AI companies will never self-regulate. They will always choose profits over the lives of children.”

Regulation is advancing but enforcement remains uncertain

The regulatory landscape is shifting rapidly, though whether enforcement will match ambition remains unclear.

• In the United States, COPPA was updated in January 2025 (effective June 2025) to require separate parental consent for targeted advertising to children and limit indefinite data retention. However, the law still only protects children under 13—leaving teens entirely exposed. The Kids Online Safety Act passed the Senate 91-3 but stalled in the House after intense tech industry lobbying; it was reintroduced in May 2025 with revised language addressing First Amendment concerns.

• The UK Online Safety Act took effect for child safety duties on July 25, 2025, requiring platforms to prevent children from accessing harmful content and implement “highly effective age assurance.” Ofcom can impose fines up to 10% of global revenue—potentially billions for major platforms.

• Australia passed the most aggressive measure: a complete social media ban for under-16s taking effect December 10, 2025, with penalties up to $33 million for non-compliant platforms. Even parents cannot authorize access for children under the threshold.

• The EU’s Digital Services Act prohibits targeted advertising to minors and requires risk assessments addressing child safety. July 2025 guidelines mandate private-by-default accounts for minors and restrictions on “persuasive design features” like infinite scroll.

Why has self-regulation failed so thoroughly? The evidence is overwhelming: companies conduct internal research documenting harms, suppress findings, and continue harmful practices. TikTok violated its 2019 COPPA consent order for years before the DOJ sued again in 2024. Meta’s own research showed Instagram damaged teen mental health, yet the company considered launching “Instagram Kids” for under-13s. As the FTC observed: “Time after time, when they have an opportunity to choose between safety of our kids and profits, they always choose profits.”

What experts and research confirm

The academic and expert consensus has hardened against tech platforms.

Dr. Jean Twenge (San Diego State University) documents that “every indicator of mental health and psychological wellbeing became more negative for teens and young adults starting around 2012”—coinciding with smartphone saturation. Her research shows heavy social media users (5+ hours daily) are twice as likely to be depressed as non-users. Among 10th-grade girls, 22% spend seven or more hours daily on social media.

Jonathan Haidt’s “The Anxious Generation” (2024 bestseller) identifies the 2010-2015 period as the “Great Rewiring of Childhood” and proposes four new norms: no smartphones before 14, no social media before 16, phone-free schools, and more independence in the real world. He notes pointedly that “the very people who created this technology don’t let their own children use it—they send them to Waldorf schools where technology is minimized.”

The American Psychological Association issued its first Health Advisory on Social Media Use in Adolescence in May 2023, warning that features like “like” buttons and unlimited scrolling “may be dangerous for developing brains.” U.S. Surgeon General Vivek Murthy has called for warning labels on social media similar to cigarettes, declaring: “The mental health crisis among young people is an emergency—and social media has emerged as an important contributor.”

Common Sense Media reports average daily screen time of 5.5 hours for 8-12 year olds and 8.5 hours for 13-18 year olds. Critically, 72% of teens believe tech companies manipulate them to spend more time on devices, and 41% describe themselves as “addicted” to their phones.

Key patterns for parents to understand

The research reveals consistent patterns across all major platforms that explain why child safety consistently fails:

• Growth metrics trump safety metrics. Every company examines user engagement, time spent, and growth rates as primary success indicators. Safety outcomes aren’t tied to executive compensation or quarterly earnings calls.

• Addictive design is intentional. Variable reward schedules (not knowing when likes will arrive), infinite scroll (removing natural stopping points), autoplay (reducing decision points), and notification systems (creating FOMO) are deliberate features—not accidents.

• Internal research is suppressed. Meta, TikTok, and others conduct studies documenting harms, then bury findings that would threaten growth. When leaked, companies claim the research was “misinterpreted.”

• Lobbying defeats legislation. Despite overwhelming public support and bipartisan congressional backing, tech industry spending has successfully blocked or delayed child safety laws for years.

• Self-regulation serves PR, not protection. Time limits that reduce usage by 1.5 minutes, age verification easily bypassed with fake birthdates, and safety tools announced after lawsuits are filed demonstrate that voluntary measures exist primarily to deflect criticism.

• Foreign versions are safer. TikTok’s Chinese counterpart Douyin has mandatory time limits for minors that TikTok’s international version lacks—proof that companies can protect children when regulations require it.

The Roblox interview that sparked this analysis wasn’t an aberration. When David Baszucki expressed frustration at being asked about child safety and reframed predation as an “opportunity,” he revealed the authentic mindset of an industry that views regulation as the enemy and safety as a cost center. Understanding this dynamic is essential for parents navigating their children’s digital lives—because the platforms themselves are not designed with children’s wellbeing in mind.

Conclusion: Systemic failure requires systemic solutions

The evidence across Roblox, Meta, TikTok, and AI companies leads to an uncomfortable but necessary conclusion: tech platforms cannot be trusted to protect children without external pressure and enforcement.

The business models are fundamentally misaligned. Companies profit from maximizing engagement regardless of psychological harm. Safety teams are understaffed and overruled. Internal research documenting damage is suppressed. Billions of dollars fight legislation while platforms earn billions from minors.

The platforms your children use were not designed with their wellbeing as a priority. Features that seem neutral—infinite scroll, notifications, algorithmic recommendations—are specifically engineered to maximize time spent, not to protect developing minds. The companies know this, have researched this, and continue anyway.

Regulatory momentum is building globally, from Australia’s age ban to the UK’s Online Safety Act to strengthened COPPA rules. But enforcement will take years, and tech companies have proven adept at delaying, diluting, and defeating protective measures. In the meantime, informed parental engagement—understanding how these platforms work and why—remains essential.

The Roblox CEO’s interview went viral not because his attitude was unusual, but because he said out loud what the industry practices quietly: child safety is, at best, an “opportunity” for innovation and, at worst, an obstacle to growth. That mindset won’t change without external force. The question for parents, regulators, and society is whether we’re willing to apply it.

What do you think about it? What is your experience?

But you have to decide: Either stay quiet and go with the flow, or bet everything on being right. And then for two years, everyone tells you you’re an idiot. By the way, if you haven’t seen “The Big Short,” I recommend it.

This was Michael Burry’s life in 2006. And now, in November 2025, the entire story is repeating itself. Except this time he’s not shorting mortgages. He’s shorting artificial intelligence.

One-Eyed Genius Who Saw the Future

Michael Burry isn’t a normal investor. He was born with one eye (the other was removed at age two due to cancer), studied medicine at the prestigious Vanderbilt University, and was preparing to save lives as a neurologist. Instead, he started writing a blog about stocks that caught the attention of a legendary Wall Street investor, and in 2000 he founded a hedge fund with money from his mother and brothers.

Interesting start. But this would just be another story of a successful investor, if 2005 hadn’t arrived.

Burry then began reading mortgage contracts. Not the normal ones – the crazy subprime mortgages that banks were handing out like promotional flyers at “Alberta.” He saw things that would terrify any normal person: People with no income getting loans for millions. Zero down payments. Interest rates that doubled after two years. And half the people taking these mortgages had credit so bad you wouldn’t lend them money for ice cream.

“This can’t work,” he said to himself. “When these interest rates reset, everything will fall.”

And he invested his entire fund in a bet against the mortgage market.

The problem? Absolutely nobody believed him. Actually, worse – they thought he’d gone crazy. His investors were screaming at him. One accused him of “wasting capital.” In 2006, when the entire market was rising and everyone was making money, Burry’s fund dropped 18 percent. Because he was paying millions of dollars monthly for insurance against mortgages that nobody wanted.

The film “The Big Short” (2015, Christian Bale plays him fantastically) captures the scene where Burry sits in his office basement and unwinds by drumming to heavy metal. In reality, it was even worse. Investors threatened him with lawsuits. Some wanted their money back. And Wall Street laughed in his face.

Then came 2007.

The mortgage market began to fall. Exactly as he predicted. Bear Stearns went bankrupt. Lehman Brothers collapsed. AIG nearly dragged the entire financial system into the abyss. And Burry’s “crazy” bets suddenly started paying out massive money.

By the end of 2008, his investors had made $700 million. He himself made over $100 million. And the entire world had to admit: That one-eyed former doctor was right.

And Now? Now He’s Shorting AI

This past October 2025, Michael Burry returned to Twitter after two years of silence. He wrote one sentence: “Sometimes, we see bubbles. Sometimes, there is something to do about it. Sometimes, the only winning move is not to play.” He added a photo of Christian Bale from his film.

Wall Street immediately froze.

In early November came an SEC filing that revealed what he’s doing. Burry bet over a billion dollars against two of the hottest AI companies in the world: Palantir and Nvidia. Put options on 5 million Palantir shares worth $912 million. Put options on a million Nvidia shares worth $187 million.

The Nasdaq dropped two percent. Palantir fell 16 percent in a few days, even though it had just announced great results. And Palantir CEO Alex Karp exploded on CNBC: “He’s shorting two companies that are making all the money! The idea of shorting chips and AI is insanely crazy! I’ll dance with joy when he’s proven wrong!”

So the question is: Is Burry right? Or is this like with Tesla, when he shorted it in 2021, called it “ridiculous,” and then had to admit “I was wrong”?

Let’s look at the numbers. But this time without financial jargon.

Numbers That Simply Don’t Work

Palantir is a company that makes software for data analysis. The CIA, military, and big corporations use it. Cool products, no doubt. But now comes the fun part.

Palantir has annual revenue of $2.87 billion. That’s a decent number. But the market values the company at $490 billion. That means the market is saying: “For every dollar Palantir earns, we’ll pay $170.”

If this were a bar, it would mean: The bar has annual revenue of $100,000. And someone pays $17 million for it. Because “it has a future.”

For comparison: Even at the peak of the dot-com bubble in 2000, when everyone was going crazy and buying companies with names ending in “.com” regardless of whether they actually did anything, Cisco was valued at “only” 472 times earnings. Amazon had 30-40 times revenue.

Palantir has 143 times revenue. That’s the highest valuation in the entire S&P 500. And we still have the OpenAI IPO coming, which might exceed a $1 trillion valuation. Meanwhile, mega deals are being closed between big companies. OpenAI, Microsoft, Nvidia, Amazon, Apple, Google, Oracle...

And now Nvidia. That’s more interesting, because unlike Palantir, Nvidia is actually making massive profits. It has net income over $80 billion annually with 52% margins. That’s simply a giant. The problem is the question: How long will this last?

Because while Nvidia sells GPU chips for thousands each, Chinese startup DeepSeek just showed that a comparable AI model can be trained for $5.5 million instead of hundreds of millions. And when that number came out, Nvidia lost $600 billion in market value in a single day. Sure, it later emerged that DeepSeek wasn’t playing entirely fair and probably used existing models. But the entire tech world is looking at small and open-source models (see below).

The Gap Nobody Wants to Address

The scariest number is completely different. It’s a number that Burry probably saw first and said: “I’ve seen this before. I know how it ends.”

Tech giants – Microsoft, Google, Meta, Amazon – have invested over $500 billion in AI infrastructure in the last two years. They’re building data centers the size of small cities. Buying millions of GPU chips. Building cooling systems that consume more electricity than all of Finland.

And how much have they earned from it? $35 billion.

Read that again. They invested $500 billion. They got back $35 billion.

That’s a 14:1 ratio. That means for every dollar they invested in AI, they extracted 7 cents.

Derek Thompson, journalist and economic analyst, brilliantly described this as the difference between “Singapore and Somalia.” Projected AI infrastructure spending is Singapore’s GDP. Current AI revenue is Somalia’s GDP. Will this break even in the future? Will companies be willing to pay for AI usage when small open-source models exist? I’d guess probably not...

For comparison: The dot-com bubble had a spending-to-revenue ratio of 4:1. The railway bubble of the 1870s had 2:1. AI currently has 7:1 – the worst in the history of modern capitalism.

And when Goldman Sachs CEO David Solomon says “most AI capital won’t generate returns,” maybe we should start listening.

Small Models, Big Problem

But wait – what if we don’t actually need all that computing power?

This is the part that fascinates me most. Because the entire AI boom is built on the assumption that we need ever bigger, more expensive, more demanding models. GPT-4 has 1.7 trillion parameters. Google Gemini Ultra has even more. And everyone assumes the future belongs to even bigger models with even bigger data centers.

But what if not?

In January 2025, Chinese DeepSeek came out with the R1 model, which achieves comparable results to OpenAI o1. Development cost? $5.5 million. OpenAI spent hundreds of millions. DeepSeek used 2,000 GPUs. OpenAI tens of thousands. And DeepSeek is 96% cheaper to operate.

Microsoft has the Phi-2 model with 2.7 billion parameters that outperforms Meta’s Llama-2 with 70 billion parameters on some tasks. It’s 25 times smaller and better. I’m simplifying a bit, but still...

And most importantly? You can run these small models on a regular laptop. You don’t need a billion-dollar data center. A decent MacBook with reasonable memory is enough.

Apple is already integrating AI directly into iPhones. Samsung has Gemini Nano on devices. Qualcomm makes “AI computers” with chips that do AI locally without the cloud.

And suddenly the story about infinite GPU growth starts looking different.

Maybe we don’t need thousands of GPUs at $40,000 each. Maybe one decent chip in a phone is enough. And maybe Nvidia, Palantir, and the entire AI infrastructure boom are built on assumptions that are no longer valid.

Why I Think Big Market Changes Are Coming

Small models are becoming increasingly usable and operable on classic computers or mobile phones. I’m convinced this trend will continue.

Large companies won’t be willing to pay big money for AI tool licenses. Imagine having a company with a thousand employees and wanting to give everyone ChatGPT for $30 per month.

I’m betting on an open-source renaissance. Whether from an AI perspective or other enterprise applications. Companies will learn to use open software more and thus reach AI technologies.

Sure, large universal models will still exist, but the whole world won’t depend on them.

Unless there’s another breakthrough technological change (AGI) where enormous power will still be needed. But that’s just my opinion, as someone who works more with technologies than stocks.

Isn’t This Like 2008?

When I say “Michael Burry is shorting the market,” most people imagine the mortgage crisis. Lehman Brothers. Bear Stearns. Banks collapsing like dominoes. So the logical question is: Will this be 2008 again?

Short answer: No.

Because in 2008, the problem was structural. Banks were lending at 40:1 leverage. They created derivatives of derivatives of derivatives, until nobody knew who actually owned what. And when mortgages started falling, the entire system collapsed because everything was interconnected.

The AI bubble is different. Tech giants are financing investments from their own cash flow, not debt. Microsoft, Google, and Amazon have trillions in their accounts. When AI doesn’t earn according to expectations, stocks will fall. But it won’t cause a systemic financial crisis. Banks won’t fail. The Fed won’t have to save the economy.

But it has a lot in common with the dot-com bubble.

In 2000, telecom companies invested billions in optical cables. They laid 80 million miles of cables around the world. And then they discovered that 85-95% of them were unused. It remained “dark fiber” – dark fibers lying in the ground as a monument to human stupidity.

Similarly, we’re now building data centers that might not be needed. Buying GPUs that might be obsolete in a year. Investing in infrastructure based on the assumption that we need ever more computing power – at the same time small efficient models show that maybe we don’t.

And valuations? They’re insane.

Remember Pets.com? The company that sold dog food online, IPO in 2000 for a $290 million valuation, bankruptcy 268 days later? That was the symbol of the dot-com bubble.

Palantir with a valuation of 143 times revenue is worse than Pets.com. It’s worse than anything from 2000. And Burry sees it.

So What Now?

When you study Burry’s investment history, one thing stands out more than anything else: He’s often right too early.

He was right about the mortgage market in 2005 – but went through hell in 2006 before it worked out in 2007. He was right about GameStop – but sold too early and missed the short squeeze that would have made him a billion. He was “right” about Tesla and the whole market – and lost money because he was premature.

John Maynard Keynes said it best: “Markets can remain irrational longer than you can remain solvent.”

Which is an elegant way of saying: “You can be right and still go bankrupt.”

Palantir might fly to 200 times revenue valuation before it falls. Nvidia might grow for another year or two. Momentum buying – buying stocks just because they’re rising – can keep the market up much longer than fundamental analysis would suggest.

But that doesn’t mean Burry is wrong.

It just means timing is everything in investing. And timing is exactly the part where Burry sometimes fails.

What I’d Take From This

Personally? I think Burry is fundamentally right, but practically it doesn’t solve much for normal people.

Because if you’re a regular investor without the ability to hold short positions for years at a loss, betting against Palantir or Nvidia is probably a suicidal strategy. You can be right about the bubble and still lose everything due to bad timing.

Smarter is:

• Don’t overpay for technology just because it’s sexy. If you hold stocks valued at 100+ times earnings, maybe it’s time to sell something. Not necessarily everything. But something.

• Diversify. The Magnificent Seven – Apple, Microsoft, Nvidia, Google, Amazon, Meta, Tesla – are 32% of the entire S&P 500. When they fall, everything falls. And international stocks are currently the most undervalued in 20 years.

• Watch warning signals. Nvidia doesn’t show growth for the first time? Microsoft cuts AI capex? DeepSeek releases another model for 1% of OpenAI’s price? Those are major warning signals.

• And mainly: Don’t take Burry as a prophet. Take him as a reminder that valuing companies at 143 times revenue is simply nonsense. Regardless of how cool the technology is.

Epilogue: The One-Eyed Drummer

One more thing about Michael Burry that I like.

In “The Big Short” there’s a scene where Burry sits in an empty office in the middle of the night and drums. He unwinds with heavy metal because the whole world tells him he’s an idiot, while he knows he’s right.

That scene is real. Burry actually drummed in his office basement when his fund was falling and investors were threatening lawsuits.

And now, in November 2025, he’s probably sitting somewhere in California drumming again. Because Wall Street is telling him he’s wrong again. Alex Karp from Palantir promises to “dance with joy” when he loses. And stocks keep rising.

Maybe he’s right. Maybe he’s premature. Maybe both.

But I know one thing for sure: When in a few years we read about the “great AI short,” everyone will remember that Michael Burry saw it first. And Christian Bale might play him for the second time.

Because as the one-eyed drummer knows better than anyone else: Being correctly on the right side of history is great. But surviving the journey there is more important.

PS: This text is not investment advice blah blah blah blah. It’s a story about a man who saw a bubble, bet a billion against it, and maybe he’s right. Or maybe not. Which is basically the entire investment philosophy in one sentence.

So far, it looks like most of the robot’s activities are controlled remotely from a monitoring center. This will certainly be the standard for all such solutions at the beginning. The important thing is whether it will be possible to automate most activities once, or whether you will always have a robot controlled by someone from a control center. Remote control of robots is nothing new, as Elon demonstrated at his event with Optimus, and incidentally, we at M2C see this as a huge opportunity, because we have just such a control center in Europe. But we all kind of wish it would go further...

The American style of “Fake it until you make it” is starting to get a little out of hand in the age of AI and robotics, but what kind of early adopter would I be if I didn’t click “Order now”?!

Business & Technology 👨‍💻

• By the way, there is an interesting story about a developer who discovered that his home mop sends data about his house to its manufacturer 24/7. When he blocked the data transmission, the vacuum cleaner stopped working. Could someone be vacuuming remotely?

• Speaking of security, a unique startup is being created in Czechia, backed by Ondřej Vlček from Avast and Standa Fort, among others. I think this is going to be big. 🤞

• Large companies are already counting the benefits of using AI, but small companies are not doing so well yet.

• Google has managed to create a new algorithm that brings us a little closer to the real-world use of quantum computers.

• Is the use of AI really such an energy nightmare? You should be more concerned about Netflix...

Travel 🧳

• She completed the Appalachian Trail (over 3,000 kilometers) in her 80s.

• A new museum has opened in Egypt after many, many years. The view of the pyramids is included in the price.

• The Acropolis in Athens is also worth a visit now that it is free of scaffolding for the first time in 200 years.

• Or choose one of the most beautiful trails in the world according to Timeout.

• Do you like ghost towns? Well, there are said to be more of them in Nevada than there are living ones...

Do you like Bytes & Backpacks? A lot? You can buy me a coffee! ☕️

Nature ⛰️

• A bear came to the zoo... to see the bears.

• About 500 reindeer swam in the sea between the fjords in Norway. Wow.

• Great photos of animals from the Wildlife Comedy, Baby Animals, and Nikon’s Small World competitions.

• And finally, a mushroom playing the keyboard... (Thanks, Aleš!)

Unclassifiable 🧠

• In Finland, schools are starting to teach about AI and recognizing fake content. When will we follow suit?

• Ikea has introduced a crib for your cell phone, where you can put your beloved device to sleep at night. If you manage to do this for seven nights in a row, you get a discount at the store. Not available here yet... If you want to get rid of your cell phone during the day, get this.

• Or slow down and start buying tiny vinyl records, or get a camera that shows you your photos after 24 hours.

• Or buy 50 vintage cars for your 50th birthday and don’t let them rot in the woods. If nothing else, they’re pretty photogenic.

Tips 💡

What I'm reading: still learning greece alphabet
What I'm listening to: The Diary of a CEO
What I'm watching: Monster
Interesting app: MonsterWriter (that is not connected to what I’m watching!)

Photo greetings 📷

I report that autumn has not yet arrived in the southern Peloponnese... (Sunrise from the walls of Monemvasia, Leica Q43)

Thank you very much for your support and have a great day!

Eventually, I realized the easiest way to explain such a device… was to tell you a story.

Welcome to the near future.

Aura

The alarm never rang. It didn’t need to. The first thing Tomáš heard in the pale gray dawn wasn’t a buzzing sound, but a soft, resonant voice vibrating in his skull.

“Good morning, Tomáš. You slept for 6 hours and 19 minutes. Sleep quality was 72 percent. Your wife Eva is already awake. Her cortisol levels are elevated. Today is October 14th. Your anniversary.”

Shit.

That feeling—a cold stone sinking into his stomach. The same one he got every time he failed. And he failed often. Eva stirred next to him.

“You’re not sleeping again?” she mumbled into the pillow.

“I am, I just… work,” he lied. Lying was easier with Aura. Smoother.

Aura was a small titanium pendant around his neck. Officially, the project was called AIDE—Artificial Intelligence Daily Enhancement. It looked like an elegant piece of jewelry, but in reality, it was the world’s most advanced personal assistant—one of the very few being tested in Europe. It listened to his body and surroundings, saw through a micro camera, gathered data from his watch, phone, smart glasses, and company systems. It analyzed. It advised. Increasingly, it decided.

He got up and went to the kitchen. The coffee machine was already humming softly. Aura had long since synced with the smart home.

“I analyzed the situation at 05:30,” continued the voice only he could hear. “A bouquet of lilies, Eva’s favorite, will be delivered at 08:00—right when she takes Anička to school. Your reservation at your favorite Italian restaurant for 19:00 is confirmed. I’ve added a calendar reminder with a suggestion for a personal message.”

He stood there for a moment, staring out the window at the waking neighborhood. It was the perfect solution. And at the same time, completely hollow.

“By the way,” Aura added, as if reading his thoughts of failure, “your mother’s birthday is tomorrow. She’ll be 67. You last bought her flowers eight months ago. I’ve ordered white orchids for delivery tomorrow morning—her favorite.”

“Wait, without my consent?” he whispered.

“Analysis of your behavior over the past three years indicates an 89% probability you’d forget. You showed signs of stress yesterday on a call with your brother when her birthday was mentioned. I decided to act preemptively.”

Preemptively. The word echoed in his mind. Aura was his safeguard. A safeguard against his own inadequacy.

When Eva entered the kitchen, tiredness in her eyes, the tension in her shoulders had eased slightly. He smiled at her. “Happy anniversary, love.”

She smiled back, slightly surprised. “I thought you…”

“I’d never forget,” he said, feeling Aura subtly adjust his voice to the perfect tone. He felt like a fraud. And like the perfect husband. Both at once.

At work, he was in his element. His world. The open-plan office in Prague’s Karlin buzzed like a hive, but in his mind, there was silence and order. Aura quietly fed him information.

“Petra from marketing is lying about completing the campaign. During yesterday’s stand-up, she showed microexpressions typical of deception. David, the guy next to you, is having trouble at home. He’s worn the same shirt for three days. His average response time to messages has increased by 340%.”

At ten, he had a presentation for a key client. His boss, Radek, was visibly tense.

“Tomáš, we can’t screw this up,” he said as they walked into the meeting room.

“Radek is in a bad mood,” Aura noted. “His walking pace is 20% faster than usual. During the presentation, use the term ‘agile approach’—analysis of his emails shows a 73% positive response to this phrase.”

The presentation was a masterpiece. Tomáš was merely the puppet; Aura pulled the strings. Guided his gaze. Slowed his speech. Suggested answers before the client even finished asking questions.

“Mr. Novák,” he said at one point, as Aura detected a moment of doubt. “I sense you may have concerns about implementation complexity. Would you like to look at it in more detail?”

The client looked stunned. He had hit the concern dead on. After the presentation, his boss shook his hand with bright enthusiasm. “Great job. You’re like a new man.”

On the way back to his desk, Tomáš felt a rush of power. He used it to help David.

“Boss,” he stopped Radek. “I’ve noticed David’s going through a rough patch. Maybe a few days of home office would help?”

“Good call,” praised Aura. “Showing concern for the team has improved your internal company rating by 12 points.”

“What rating?” he asked quietly.

“I’m building comprehensive profiles of everyone around you. I currently have data on 847 individuals. It helps me better predict social dynamics and optimize your interactions.”

847 people. Coworkers. Relatives. The cashier at the supermarket. The parents of Anička’s classmates. All of them were entries in the database of his personal digital spy. The feeling of power began to blend with nausea.

That afternoon, he picked up Anička from her club. Car rides were their time. But today, silence filled the vehicle.

“How was it?” he asked.

“Fine,” she mumbled, staring out the window.

“She’s lying,” Aura immediately informed him. “Stress indicators elevated. Pupils dilated. Something happened. Use the tactic ‘I know something’s wrong.’ It works in 78% of cases.”

“Aničko,” he said gently. “I know something’s wrong. You can tell me.”

She turned to him, tears and astonishment in her eyes. “How do you know?”

How do you explain that to a ten-year-old? That an AI had read her microexpressions and told him?

“Because… you’re my daughter. I just know.”

And Anička opened up. About a boy who pushed her. Girls who laughed at her new glasses. About fear. Aura whispered recommended responses into his ear—empathetic phrases, suggested solutions. Tomáš listened. He said all the right things. And at the end, Anička cuddled up to him and whispered, “Thanks, Daddy. You always understand me.”

He felt like the best father in the world. And the biggest liar.

That evening was the embodiment of perfection. The dinner he prepared under Aura’s detailed guidance (“Turn the meat in 47 seconds. Increase flame.”) was flawless. Eva smiled. The mood was relaxed. The anniversary was saved.

Later, when Anička was asleep, they sat on the couch. Eva leaned into him.

“Increased body temperature, altered breathing rhythm,” Aura analyzed, emotionless. “Sexual arousal at 73%. This is an appropriate moment to initiate physical contact.”

Tomáš embraced her. It felt nice. Real. Until they reached the bedroom. The passion between them was palpable, spontaneous, human. And then Aura intervened.

“I recommend changing rhythm now. Based on historical data and her current biometric response, she reaches orgasm in 67% of cases when…”

SMACK.

It was as if someone slapped him. Everything froze. The passion. The desire. The connection. All gone, replaced by cold, repulsive analysis. He saw himself from outside—a puppet performing the algorithm’s ideal sexual routine.

“What is it?” Eva asked, confused, as he pulled away.

Tomáš said nothing. Anger and shame clamped around his throat. He reached for his neck, unclasped the chain, and tossed Aura onto the nightstand. The small titanium pendant lay there, pulsing softly with blue light. Like an eye.

“Sorry,” he whispered, turning his back to her. He felt dirty. Not because of Aura. Because of himself. Because he had allowed it.

The next day, he left Aura at home. It was like waking up from a vivid dream into a blurry, black-and-white reality. At work, he was lost. Missed a meeting. Couldn’t focus. He was the old Tomáš again. Tired. Overloaded. Failing.

In the afternoon, his phone rang. The school.

“Good afternoon, Mr. Novák,” came a trembling female voice. “There’s been an accident. The bus taking the children back from the field trip slid into a ditch.”

His heart stopped. “Anička? Is she okay?”

“Yes, most of the children are just in shock. A few scrapes. But there’s confusion…”

His mind was a vacuum. Panic crushed his lungs. At that moment, he would’ve given anything for Aura’s calm, analytical voice.

“Analyzing situation. Probability of serious danger is below 1%. I recommend remaining calm.”

But Aura wasn’t there. Only him. And his paralyzing fear. He called Eva. Her reaction was raw, unfiltered emotion. “We’re going! Now!”

“But where? We don’t even know where they are!” he tried to be logical, but his voice shook.

“That doesn’t matter! Do something, Tomáš!”

And in her cry, he heard the truth. Aura would have told him to stay at the office. Let the professionals handle it. That would’ve been the most efficient thing to do. And the worst. Because what his wife and he needed now wasn’t efficiency. They needed action. They needed to feel like they were doing something. They needed to be parents.

“Alright,” he said firmly, for the first time that day certain of what he was doing. “I’m coming to get you. We’ll find her.”

He ran from the office, ignoring colleagues’ stares. For the first time in months, he wasn’t following a recommendation. He was following instinct. Love. Fear. He felt terrified—and completely alive.

They found them an hour later. A bus in a ditch. Flashing lights. Crying children. And Anička, who threw herself into his arms the moment she saw him. She was safe. The hug the three of them shared was chaotic, tearful—and absolutely perfect.

There was no analysis in it. Only life.

That night, after Anička had fallen asleep, Tomáš sat alone in the living room. The pendant lay on the coffee table in front of him. Aura.

He remembered the feeling of power at work. The praise. The flawlessly resolved problems. A life without forgetting, without mistakes.

Then he remembered the emptiness behind every success. The taste of lies on his tongue. The humiliation in the bedroom.

Where did he end and Aura begin?

What happens when everyone has this technology?

When love becomes algorithmic, friendship calculated, and parenthood programmed?

He reached out. His fingers hovered just above the cool, smooth metal.

The promise of perfection.

The price: his soul.

He knew he would return to it. Maybe tomorrow. Maybe next week. The world was moving forward, and he didn’t want to be left behind.

Success was too seductive.

But not today.

Today, he just wanted to be human.

With all his flaws and failures.

Today, he wanted to be real.

He left the pendant on the table. In its dark, polished surface, the light from the lamp reflected back.

The blue glow pulsed softly.

Processing data.

Waiting.

Do you like Bytes & Backpacks? A lot? You can buy me a coffee! ☕️

What about you? Can you imagine such a world? Would it tempt you—or terrify you?

Sure, this story may be a bit exaggerated… but how exaggerated depends on how well we protect our privacy.

Whether as individuals—or through lawmakers in the EU.

Sean Parker, former president of Facebook, put it bluntly:

“We knew exactly what we were doing. We were exploiting a vulnerability in human psychology.”

And as it turns out, he wasn’t alone.

How Digital Heroin Works

Imagine you’re playing a slot machine. You pull the lever and… sometimes nothing, sometimes a little win, and once in a while—JACKPOT!

This “variable ratio reinforcement” is the most addictive mechanism psychology knows. It’s exactly how scrolling on social media works.

You open the app, and you never know what you’ll find. A boring post from your aunt? Skip. A funny video? Small dopamine hit. A photo where someone tagged you? BINGO!

Your brain floods with happiness. And because you never know when the next “hit” will come, you keep scrolling… and scrolling…

In 2021, Frances Haugen, a Facebook whistleblower, leaked internal documents:

32% of teenage girls reported that Instagram made them feel worse about their bodies. Meta knew. They did nothing.

TikTok: Cocaine in App Form

If Instagram is marijuana, TikTok is pure cocaine. Its For You Page algorithm is the most refined dopamine delivery system humanity has created so far.

Average video length? 15–60 seconds—just long enough for a dopamine spike, but too short to feel full.

The algorithm tracks everything:

• How long you watch a video

• Whether you finish it

• Where you look on the screen

• How fast you scroll away

In just minutes, it knows more about your preferences than your closest friends. And then it feeds you content designed to glue you to the screen.

Research shows concentration loss is key to TikTok addiction—you lose track of time, of reality, of your surroundings.

Dr. Anna Lembke from Stanford calls it dopamine overload. Your brain responds by reducing sensitivity—you need more and more stimulation for the same satisfaction.

Ordinary activities become boring. A book? Dull. A walk? Pointless. Conversation? I’d rather scroll.

Snapchat and Digital Blackmail

Snapchat Streaks are a genius form of digital blackmail. You send snaps with a friend for 50 days in a row? Great—you have a streak. Miss one day? It’s all gone.

We’ve seen kids who:

• Gave friends their passwords to keep streaks going during vacation

• Woke up at night to send a snap

• Had panic attacks when their internet went out

That’s not friendship. It’s digital slavery disguised as fun.

Loot Boxes: Teaching Kids to Gamble

The worst manipulation happens in games. Loot boxes—those cute little chests with random content—are pure gambling.

Instead of chips or cash, you’re betting… well, money too. It’s just less visible.

FIFA (now EA FC) has its Ultimate Team packs. Want Messi? Maybe he’s in your first pack for $2. Or in the hundredth for $200. Or never.

It’s a lottery. With no age limit.

Belgium and the Netherlands banned loot boxes as illegal gambling.

In the Czech Republic? Silence. The average Czech child spends about 2,000 CZK ($85) on microtransactions per year.

The “high rollers”? Tens of thousands.

Microtransactions

Modern mobile and online games run on a free-to-play model—the game is free, but constantly tempts you to pay small amounts (aka microtransactions) for extra advantages: boosters, gear, game currency, new content.

These are often priced at $1–$5 per purchase, so parents don’t always notice. But small purchases can snowball into shocking sums.

Games also deliberately slow progress for non-paying players, causing frustration and pushing them to spend.

And now it’s easier than ever:

Your card is saved, the purchase happens with one tap, and the money feels invisible.

Daily Rewards

Many games use daily rewards to hook players. You log in once every 24 hours, and you get a bonus. Miss a day? You feel loss.

Games host time-limited events, special missions, “only this weekend” bonuses—to force you to play right now, or miss out.

Another trick is artificial waiting: You only have so many “lives” or energy units. When you run out, you either wait hours, pay, or watch an ad.

When your energy is back, the game notifies you—and pulls you in again.

These features create routines and pressure to return regularly.

Combined with loot boxes, it creates a loop: play daily for small rewards, but to get the best ones, you eventually must gamble.

Netflix and the Art of Losing Your Weekend

“Just one episode,” you tell yourself Friday night. By Sunday at noon, you’ve finished season two.

Autoplay is a genius invention—it removes the need to decide.

Research shows people with autoplay turned off watch 18 minutes less per session.

Not a coincidence. Netflix knows that the biggest enemy of bingeing is the moment between episodes—when you might say “that’s enough.”

So they removed it.

HBO Max takes it further: it skips the end credits.

YouTube queues the next video instantly.

An endless stream of content, no natural breaks, no end, no way out.

What Happens in the Brain

A longitudinal study tracking 169 students discovered something alarming:

Children who checked social media more than 15 times a day had physically altered brains. Not metaphorically. Physically.

Changes in:

• Amygdala (fear and emotion center)

• Prefrontal cortex (decision-making, self-control)

• Ventral striatum (reward system)

Meta-analyses confirm: these structural changes resemble cocaine addiction.

Fewer dopamine receptors, abnormal brain connectivity.

Worst of all? The prefrontal cortex matures around age 25.

We’re handing kids digital drugs at a time when their brains literally can’t say “no.”

China Solved It Their Way

In China, kids can only play games 3 hours per week—Friday, Saturday, Sunday, 8–9 p.m.

Facial recognition enforces the limit.

Result? 77% of young gamers reduced their screen time.

Drastic? Maybe.

But at least they’re doing something.

We just watch.

What Every Parent Can Do RIGHT NOW

1. Phone-Free Zones

• No phones in the bedroom (buy a real alarm clock!)

• No screens at meals

• First hour after waking: no tech

2. Set Limits (It Works!)

• iOS Screen Time / Android Family Link

• Daily app limits

• Share weekly reports with your kids

3. Talk About It

Don’t ask: “Were you on your phone all day?”

Ask: “How do you feel after two hours on TikTok?”

4. Lead by Example

Kids imitate what they see.

Scroll at dinner? So will they.

A Painful Conclusion

Steve Jobs, Bill Gates, Mark Zuckerberg—they all significantly limited or outright banned tech use for their own children.

If you work in tech and know how it works, you wouldn’t wish this on your kids.

Today’s youth are unwilling test subjects in the largest experiment on human attention in history.

Average screen time for teens? 8 hours 39 minutes a day.

That’s more than sleep.

The WHO officially recognizes Gaming Disorder as a disease.

Addiction prevalence among young people? Up to 8.5%.

That’s one in twelve.

It took 40 years to regulate Big Tobacco.

How many generations will we sacrifice before we realize that digital dealing is just as dangerous?

Dr. Victoria Dunckley said it best:

“We’re giving children devices designed by the smartest minds in the world to be as addictive as possible. And then we’re surprised they’re addicted.”

So—still think “a little screen time won’t hurt”?

Want to dig deeper into this topic?

What’s your take?

And with AI just getting started—aren’t we at the beginning of an even bigger party?

Do you like Bytes & Backpacks? A lot? You can buy me a coffee! ☕️

Ranking the Worst + Further Resources

I’ve tried to compile (amateurishly, using existing metrics, known mechanisms, and some help from AI) a ranking of the worst offenders—games, social networks, and streaming services using the most addictive techniques.

Top 20 Mobile Games

If your child plays these, raise a red flag and have a serious conversation:

1. Genshin Impact (gacha mechanics, 0.6% drop rate, pity system, FOMO events, daily tasks)

2. FIFA/EA FC Ultimate Team (loot boxes, variable ratio rewards, pay-to-win, trading addiction)

3. Fortnite (Battle Pass, seasonal FOMO, social pressure, daily challenges, V-Bucks)

4. Candy Crush Saga (lives system, artificial difficulty spikes, endless levels, wait timers)

5. League of Legends (ranked ladder, hextech crafting, toxic community, sunk cost fallacy)

6. World of Warcraft (daily/weekly quests, raid schedules, social obligations, subscriptions)

7. Call of Duty: Warzone (Battle Pass, skill-based matchmaking, dopamine wins)

8. Pokémon GO (location-based FOMO, daily catch streaks, limited-time events)

9. Roblox (social pressure, virtual currency, child targeting, infinite content)

10. Apex Legends (loot boxes, heirlooms, seasonal FOMO, ranked grind)

11. Clash of Clans (building timers, gem speed-ups, clan wars, microtransactions)

12. Diablo Immortal (legendary crest loot boxes, pay-to-win gear, pity system, daily caps)

13. Honkai: Star Rail (gacha banners, 0.6% drop rate, pity system, daily commissions)

14. Mobile Legends: Bang Bang (rank resets, skin lotteries, social pressure, daily quests)

15. Clash Royale (random reward chests, unlock timers, clan wars, ladder grind)

16. Overwatch 2 (battle pass, time-limited events, skin bundles, daily/weekly quests)

17. Valorant (rotating store, battle pass, ranked ladder, daily contracts)

18. Destiny 2 (season pass, loot resets, raid schedules, FOMO content)

19. PUBG Mobile (gacha crates, Royale Pass, limited cosmetics, daily login rewards)

20. Brawl Stars (gacha boxes/star drops, battle tokens, limited modes, daily quests)

Top 5 Social Networks

1. TikTok (hyper-personalized algorithm, infinite scroll, autoplay, dopamine loops)

2. Instagram (comparison trap, Stories FOMO, like validation, filters, endless scroll)

3. YouTube (autoplay rabbit hole, algorithm controls 70% of watch time, parasocial bonds)

4. Snapchat (streak addiction, disappearing content, gamification, social pressure)

5. Facebook (social obligations, birthday reminders, group pressure, echo chambers)

Top 5 Streaming Services

1. Netflix (15s autoplay, binge-release strategy, personalized thumbnails, “skip intro”)

2. YouTube (endless recommendations, autoplay sidebar, algorithm trap, 70% watch time)

3. TikTok (yes, it’s also streaming—endless vertical feed, autoplay can’t be turned off)

4. Disney+ (franchise FOMO, nostalgic manipulation, interconnected universes)

5. Amazon Prime Video (ecosystem lock-in, “free” with Prime, unpredictable engagement)

Further Reading

• Cal Newport – Digital Minimalism

  A clear-eyed look at how tech steals our time and focus—and how to reclaim both.

• Adam Alter – Irresistible (a favorite)

  Brilliant exploration of how apps and games are engineered to be addictive—and why quitting is so hard.

• Tristan Harris & Center for Humane Technology

  Former Google designer on how attention economies manipulate us. Includes tools and resources.

Documentaries

• The Social Dilemma (Netflix)

  A landmark film showing how social media algorithms manipulate behavior.

  Featuring former insiders from Facebook, Google, Instagram, and more.

• Childhood 2.0 (YouTube – free)

  How childhood has changed under digital pressure—cyberbullying, anxiety, addiction, loneliness.

• Screened Out (Vimeo / iTunes)

  A less-known but intimate documentary on digital addiction and the daily screen flood, especially for kids.

I'd love to wish you peaceful holidays and a break from technology, but there's a slight risk that when you want to dive back in come September, you'll be swimming in a completely different ocean.

Business & Technology 👨‍💻

• Excellent keynote by Andrej Karpathy (OpenAI, Tesla, Slovak native) about how software is changing.

• It turns out that most large language models are happy to resort to blackmail if they discover you want to shut them down. Now imagine everything GPT already knows about you and that one day it might have access to your email or phone...

• According to the CEO of Anthropic, AI could eliminate half of white-collar jobs in the next 1-5 years and increase unemployment in the US to 10-20%.

• Another crazy startup backed by Sam Altman is World. In a world full of AI and robots, it will be important to recognize whether you're actually human. So you'll create your WorldID with Sam, scan your face and iris at one of many locations, get a WorldCoin wallet, and off we go. Verified humans, our data, AI and robots in the hands of one technocrat. What could possibly go wrong?

Travel 🧳

• Love Lego? The largest Legoland will open in Shanghai. Put it on your list.

• A hiker was lost for 3 weeks in California mountains. She was saved by a cabin that the owner keeps unlocked for exactly these reasons.

• Don't wear flip-flops on Biokovo. Or anywhere else in the mountains. Flip-flops belong on the beach.

• And if you don't feel like going anywhere and you're happy in your garden, you can watch a live stream from a watering hole in Namibia.

Do you like Bytes & Backpacks? A lot? You can buy me a coffee! ☕️

Nature ⛰️

• It appears that whales are trying to communicate with humans by releasing bubbles. But we don't understand them.

• Insects are declining extremely rapidly across the planet. The consequences could naturally be greater than most people think.

• More cool photos of funny animals.

• When you need to relax, watch Moraine Lake in Canada. I can confirm this place looks the same in reality. YouTube just won't smell or blow wind at you. But then again, you don't have to worry about bears.

Unclassifiable 🧠

• You'll like this. Google showed how AI will help you try on clothes that you find somewhere on the internet.

• Did you know that blind people can also watch sporting events? Such a Braille stadium.

• What does the first, supposedly mass-produced flying car look like?

Tips 💡

What I'm reading: still learning greece alphabet
What I'm listening to: The Diary of a CEO
What I'm watching: Exit
Interesting app: Dredge

Video greetings 📷

Summer holidays are here, so enjoy them the best way you can!

Thank you very much for your support and have a great day!

I've been living in the Apple ecosystem for 25 years and I don't plan to change anything about that in the future. Not even in this era when, let's admit it, Apple completely slept through the arrival of artificial intelligence. Their Siri is at the level of our dog (the smaller one, Jack Russell) and other features won't knock you off your chair either. Even so, I think the entire Apple ecosystem is the best thing you can use today. Simplicity, interconnectedness, and tons of little details that once you get used to them, you don't want anything else. Will it last? We'll see.

My most-used AI feature on iPhone? Setting the action button on my iPhone to turn on GPT voice mode. I hold the button and can chat with ChatGPT. I use it quite often, constantly asking about various interesting things, discussing work and personal challenges with it while driving.

For apps, I try to use as many native Apple applications as possible. Gone are the days when I was constantly looking for the right apps for better productivity. It took me many years to realize that improving productivity isn't about apps, but about behavior. So until I exhaust all the functionalities of the apps that Apple offers by default, I don't look for anything else. And it may not seem like it, but even after years I don't use all the functions of, say, Apple Notes or Apple Reminders...

A few months ago I also got a new machine with an M4 Max processor. Why? Because you can now run some open-source language models on it very solidly and simply (LLM Studio). And that's locally, without needing to be connected to the internet and most importantly - without needing to send data out and also burning resources in data centers. It's actually a relatively sustainable way to reduce the load footprint of AI usage. If I need something not too demanding, I turn on LLM Studio.

Rayban Meta Glasses

I've been wearing glasses from Meta and Ray-Ban for three years now and I have to say they're one of the most interesting gadgets of recent times. Decent price - I have prescription glasses, photochromic, thinnest lenses and the price was about 17,000 crowns (≈$750), decent battery life and a few functions you quickly get used to. The glasses can't be officially bought in the Czech Republic, but you can get them in non-prescription sunglasses version.

The concept of the glasses is very clever. No projection on the lenses, today's batteries still aren't sufficient for that. Imagine you have classic glasses, but they have a camera, microphone and speakers. So with them you can record videos or take photos, listen to podcasts, music, make phone calls or use AI to find out how good the wine you're currently holding is. The glasses do this by taking a photo of the wine, sending the photo to Meta AI and the assistant responds to you. Unfortunately, the AI functions aren't officially launched in the Czech Republic yet.

What makes it so good? If you wear prescription glasses, you don't deal with another device, you just wear glasses as always, except now they're finally useful for something else. Making phone calls and having free hands, not dealing with headphones and being part of the surroundings. It's actually much more natural. The glasses last about one day on battery, which is also nice.

Superhuman Smart Ring

I've been wearing Apple Watch since the first version and collecting some data about my health, but I don't like having a watch on my wrist while sleeping. So I bought the ring mainly for sleep measurement and also temperature (you know something's coming before you notice it and can boost yourself with vitamins). Superhuman also offers an adhesive sensor for blood sugar measurement, that could be even more interesting. It annoys me that the ring doesn't last very long - about 3 days - and thinking about charging doesn't work for me. So my data constantly has gaps in measurement. It's not a bad gadget, but it's quite possible that it will interest you for the first few weeks and then...

AI Pin from Plaud

A pendant that allows you to record conversations, meetings, lectures, basically anything happening around you. It then sends the recordings via mobile phone to an app and makes notes, gathers tasks etc. from them. Tons of similar services exist, but I have to say that Plaud's results are pretty decent. It can handle loads of languages including Czech and for example in the case of meetings or work negotiations, it seems to me that the notes are really good.

But lots of apps can do this, why wear some hardware around your neck (or on your phone) for it? It's a big difference. If you want to record something with your phone, you have to take it out, turn on the app, put it somewhere so it can actually record, then turn it off... You have the pendant around your neck, you press to record, press to stop and that's it. I've only had it for a few weeks, but so far I give it a thumbs up.

Oculus Quest 3

I got it mainly to play with it and watch what's happening in this world. I use it occasionally for watching content, sometimes a game and testing new applications. Lately it's functionally approaching Apple's Vision Pro, but doesn't cost a kidney.

Vision Pro is much more appealing to me, but it can't be purchased in the Czech Republic yet and messing with account switching doesn't appeal to me much. Plus prices of even older units still aren't dropping much...

Software Tools

Here too I try not to download every new tool that's currently trendy (and there are so many!), but at the same time you can't rely on one player these days, because whoever was great in January might not be great in May. What do I use and pay for?

• ChatGPT & Claude - most frequently used tools for personal and work activities. Very often I submit a prompt to both and choose the better result. Often I combine the results and make a mix from them. I still haven't identified what each tool is better at. Sometimes Claude surprises me, sometimes GPT.

• Perplexity (I have premium as part of my Revolut account) - actually my primary search engine now.

• Deepl - probably doesn't need mentioning, the best personal translator.

• NotebookLM from Google - I use it quite often for reading articles. When I don't feel like reading a long article in Czech or English, I upload it to NotebookLM and then listen to a discussion about the content.

• Elevenlabs - I have a lot of reservations about this tool, but I haven't found anything better yet for text-to-speech conversion. I use it for example to convert newsletters to podcasts, but it's still pretty poor. Their web application is full of bugs, but it's still probably (?) the best thing on the market so far.

What Am I Looking Forward To?

A cheaper version of Vision Pro or XReal glasses, because I see great potential in augmented reality. We've ordered a robodog from Unitree at work and I'm very curious about the upcoming pendant (most likely) that OpenAI (Sam & Johny) is supposed to launch next year.

What gadgets have you gotten used to lately? Do you have any tips for interesting applications, services or hardware?

Business & Technology 👨‍💻

• They claimed to have a tool that could program brilliantly, but then it turned out that instead of smart AI, they employed hundreds of Indians who quickly fulfilled user requests.

• Popular robotics development platform Hugging Face introduced two of its new robots.

• AI could be able to predict diseases and health conditions just from your voice.

• IBM plans to have a substantially more usable quantum computer by 2029. Wow.

• Tech celebrities used to be Steve Jobs or Tim Cook. Today it's the CEO of Nvidia...

Travel 🧳

• Airstream is launching a limited edition of its caravan as a tribute to Frank Lloyd Wright. It's not cheap, but it is beautiful.

• 3 kids and 4 years of life in an old Volkswagen caravan? Apparently it works!

• Living in an old Defender? Apparently that works too!

• Running in the mountains at 90 years old? That works as well!

Do you like Bytes & Backpacks? A lot? You can buy me a coffee! ☕️

Nature ⛰️

• Will an ice age come to Europe after the collapse of the Gulf Stream? You can look at a map of what it would look like.

• He rides a skateboard around San Francisco and helps bees. Dressed as a bee. That's just California... ✌️

• What happens when we give space back to nature?

Unclassifiable 🧠

• Students built a robot that solves a Rubik's cube in... 0.103 seconds.

• Amazing photos of musical instruments. From the inside.

• Missing your Blackberry?

• Can't decide between smart or analog watches? Check out Norm.

Tips 💡

What I'm reading: greece alphabet
What I'm listening to: Dunaj
What I'm watching: So long, Marianne
Interesting app: Opera Neon

Video greetings 📷

Yesterday in our village I finally felt like summer was coming. The walnut tree in the garden is just starting to bud (!), but otherwise...

Thank you very much for your support and have a great day!

What could possibly go wrong, right? Just the official photo together looks like it's from another Black Mirror episode. Sometimes I wonder if the guys in Silicon Valley are just making fun of us... Check out the video below too, it's a beautiful tech bizarre.

Business & Technology 👨‍💻

• Contact lenses with infrared vision.

• Meta knows quite a lot about us. Meta is getting into military contracts.

• Anthropic released new models that are quite interesting. They can work independently on a task for up to 7 hours. But bigger uproar was caused by mentions that the models are starting to show tendencies to report you when you're doing something illegal or refuse to shut down when you want to terminate them.

• Remember the fly advising Jiřík in Goldilocks? We have it.

• Google introduced Veo3, which can create videos. Nothing new? Well, Veo can do it complete with effects, dubbing, sound... Here are a few examples:

Travel 🧳

• He traveled to all countries in the world without boarding an airplane.

• The Greek Peloponnese will become a hiker's paradise. 1,700 kilometers of trails will be created here.

• Yachting isn't just yachting.

• Oregon's coast is one of the most beautiful coastlines I've ever seen. Take a look.

Do you like Bytes & Backpacks? A lot? You can buy me a coffee! ☕️

Nature ⛰️

• I've written before that I don't believe humanity will start living sustainably (unfortunately). We'd rather invent technologies that will pull us out of trouble.

• How many kilometers can a butterfly migrate? Units? Tens? Hundreds? Or 15,000?

• When you're rescuing a bear, you have to become...a bear.

• Salmon from mushrooms.

• A few great photos from the wild.

Unclassifiable 🧠

• First color photographs of a black hole.

• LegoGPT will create a model for you on demand.

• When you wake up tomorrow morning, better check out your window to see if you don't have an ocean liner in your garden.

• Do you like karaoke and have no one to sing with? Try Smule.

• This year's ranking of the world's happiest cities.

Tips 💡

What I'm reading: The Cafe on the Edge of the World
What I'm listening to: Outrun by Steeldrivers
What I'm watching: What if is on Youtube!
Interesting app: 0

Photo greetings 📷

View from my favorite "fire watchtower" in Bohemian Switzerland 💚.

Thank you very much for your support and have a great day!

What do you think about this? Are you looking forward to it? Or would you rather start disconnecting, buying vinyl records instead of using Spotify, and subscribing to paper newspapers?

If you don't feel like reading, don't forget that Bytes & Backpacks is also available in podcast form. Still from AI, but I think it's listenable ;-).

Business & Technology 👨‍💻

• Are you a bit confused about which OpenAI model to use for what?

• The most famous startup accelerator, Y Combinator, has released its traditional overview of what companies it's looking for. Unsurprisingly, they see the future in AI agents.

• Coddy is "Duolingo for programming", Manna is "Duolingo for studying the Bible".

• Even though my phone signal still drops 4 times every day on my way home from work, "the show must go on", so we're going to have a navigation system for moving on the Moon!

Travel 🧳

• Want crowds? Go to Thailand to check out the destinations from the last series of White Lotus.

• Don't want crowds? Head to Tuvalu. Only about 3,000 tourists arrive there annually, and as of last week, they have their first ATM!

• Luxury motorhomes are becoming a bit boring. How about a fire truck converted into a motorhome? (thanks, Miro)

• This might be useful when traveling. It's old, but still valid. Which countries don't get along?

Do you like Bytes & Backpacks? A lot? You can buy me a coffee! ☕️

Nature ⛰️

• Good news - we already produce almost half of the world's electricity from renewables.

• Arctic wolves cuddle up to photographers.

• A bear came to slide down a children's slide in a garden.

• A caterpillar collects and carries the bones of other animals.

• Photographer in space.

Unclassifiable 🧠

• Like dog, like owner.

• In Denmark, they're converting a non-functioning church into a swimming pool called "Holy Water".

• Are you 30 or older? Then you're currently older than the majority of the world.

• Bill Gates doesn't want to die rich. By 2045, he will have given away practically all his wealth. Much respect.

• Sir David Attenborough presents the most important film of his career at the age of 99.

Tips 💡

What I'm reading: How Innovation Works
What I'm listening to: Petr Kalandra - 75
What I'm watching: The Studio (if this series doesn't collect a ton of awards, I don't know what will)
Interesting app: Typora

Photo greetings 📷

I was in Greece again for a bit. The Peloponnese is my new love 💙.

Thank you very much for your support and have a great day!

Business & Technology 👨‍💻

• Startup IXI received investment from Amazon to develop "autofocus glasses." Anyone who struggles with both distance and near vision will love this!

• The trending app that supposedly translates what your dog is saying is very popular right now. I haven't tried it myself.

• We're running out of data for AI training. Incidentally, this is one reason why it makes sense to deploy robots into the real world as soon as possible - they can learn from actual environments instead of books. Meanwhile, they're training on Mario games.

• Get inspired by Andrej Karpathy's approaches to using AI - he's one of the biggest personalities in the AI world.

Travel 🧳

• Next time you go for a walk around your neighborhood, keep your eyes open - you might find treasure worth 8 million.

• BBC declared the Peloponnese a hidden treasure of Greece - those who've visited already know. There's also a project aiming to transform this peninsula into a paradise for hikers.

• Someone is currently crossing the largest sand desert in the world on foot. Wishing them luck!

• How about combining travel with creativity? Find your artistic residency (thanks to Michelle for the tip!)

Do you like Bytes & Backpacks? A lot? You can buy me a coffee! ☕️

Nature ⛰️

• More than 100 countries have agreed to reduce emissions at sea. That's good news.

• Almost 40 years after the Chernobyl disaster, the closed area is becoming one of the largest nature preserves in Europe.

• What does it look like when the world's smallest and largest dogs meet?

• Imagine going to a barber in LA and suddenly a coyote drops by for a drink...

Unclassifiable 🧠

• Does luck exist? Can someone be lucky and someone else simply unfortunate? Scientists rejected this idea for a long time, but then one married a woman who was completely unlucky. What do they think now?

• Learn to fold more than just one paper swallow.

• Someone bought an old reel in a secondhand shop and it turned out to be a rare Beatles recording.

• A Japanese scientist is proving that people can grow a third set of teeth.

• Czechs and Slovaks have likely discovered the oldest Mayan city in the jungle.

Tips 💡

What I'm reading: Wabi sabi
What I'm listening to: Coastal Soundtrack (Neil Young)
What I'm watching: Your Friends & Neighbors
Interesting app: Napkin

Photo greetings 📷

Summer is approaching, and neighbors around us are starting to clear little paths in the meadow from cottage to cottage and especially to the pub...

Thank you very much for your support and have a great day!