What a time to be alive

One of my favorite AGI definitions is this: If we take all human knowledge accumulated before 1900 and give it to an AI with one goal — understand the nature of reality and the laws of mathematics, physics, and chemistry… Will that model eventually rediscover the same knowledge we have today? Will AI independently arrive at E=mc² and other fundamental discoveries? To me, this is one of the most beautiful descriptions of AGI. PS: For this, we need world models and high-quality simulations where AI can run experiments. AGI is not only intelligence. It is also the quality of the environment and the tools.

TBH, everything seems pretty obvious to me here. This is classic Trump style — he wants to make it clear who’s in charge. And in this situation, Anthropic doesn’t really have another option.

That number sounds astronomical at first glance. But if you zoom out, it might actually be a massive steal. Cinema is arguably the crown jewel of modern art. Maybe the most culturally significant art form of the 20th and 21st centuries. Today, the Mona Lisa is basically priceless. Over the next centuries, iconic film rights could appreciate in a similar way. A hundred years from now, movies like Forrest Gump may have the same cultural status. Now look at the Warner Bros portfolio: - Harry Potter - DC Comics - The Lord of the Rings - The Hobbit - Game of Thrones - Friends - Scooby-Doo - This is not just IP. These are cultural monuments. At $110B, Paramount may be buying one of the greatest cultural portfolios ever created for pennies on the dollar.

The Voynich manuscript is one of the most fascinating mysteries of our time.

How did Atlas learn football – and why? Go behind the scenes of School of Football with @Hyundai_Global and discover a glimpse of the Next of robotics. Keep reading for more from our team: bosdyn.co/4oaTqGz Next Starts Now #Hyundai #NextStartsNow #Atlas #BostonDynamics #SchoolofFootball #Robotics

New NVIDIA physical AI agent-ready skills are changing how robotics researchers work. 🤖 With NVIDIA robotics skills, researchers can automate the most common development steps from scene preparation to simulation and learning with Omniverse libraries, Isaac frameworks, and physical AI open datasets. Specialized skills like Isaac Mobility extend that even further. Learn more ➡️ nvda.ws/4epz4VR

My deep conviction is that infrastructure companies should not be spending a single dollar on marketing or advertising today. Retail users simply won’t find them anymore. That’s the real madness. People are building inside AI agents. They are building inside cloud environments, inside Codex, inside tools like Claude. Infrastructure companies now need partnerships and contracts with AI companies like OpenAI and Anthropic. They have little chance of surviving if they keep spending money trying to acquire customers directly, because those customers are no longer coming through traditional funnels. The retail funnel is now in the hands of the major AI companies. A person who wants to build a simple tourism website, for example, will do it through tools like Codex, Lovable, or Claude. They won’t know that Supabase or Replit exists. They’ll just build the site, while the infrastructure runs underneath. And 90% of those users will never grow large enough to discover these infrastructure companies in the first place. That’s why when Salesforce announces a strategic partnership with OpenAI, the stock jumps. Investors understand what’s happening. We are entering a world where companies like Salesforce won’t be able to acquire retail customers through marketing alone. The funnel has shifted almost entirely toward the major AI platforms. Ironically, chatbots have already taken control of the retail funnel. Lately, I keep hearing the comparison that chatbots are turning into giant operating systems. And the more I think about it, the more I believe it. Chatbots are becoming the interface, while almost the entire world of services and infrastructure sits behind them. Users won’t even know those services exist. From a branding perspective, that’s almost suicidal. If users never see your brand, AI companies can eventually replace your service with their own. I think we may see major regulatory changes because of this. Chatbots could end up becoming something closer to browsers. Remember the browser wars? The debates around Microsoft, Netscape, and whether browsers should give equal access to competing services instead of promoting their own technologies. It feels like we may be heading toward a similar moment. Eventually, brands will realize that AI companies are absorbing most of the value while the brands themselves become invisible. The brand stops mattering. What matters is having a contract with OpenAI or Anthropic. That’s the crazy part. My guess is that we'll eventually arrive at some kind of democratic solution to this, because the current trajectory concentrates too much power in the hands of a few AI platforms.

The thing nobody tells you about exponential change is that it feels like nothing is happening right up until the moment everything happens at once.

# on shortification of "learning" There are a lot of videos on YouTube/TikTok etc. that give the appearance of education, but if you look closely they are really just entertainment. This is very convenient for everyone involved : the people watching enjoy thinking they are learning (but actually they are just having fun). The people creating this content also enjoy it because fun has a much larger audience, fame and revenue. But as far as learning goes, this is a trap. This content is an epsilon away from watching the Bachelorette. It's like snacking on those "Garden Veggie Straws", which feel like you're eating healthy vegetables until you look at the ingredients. Learning is not supposed to be fun. It doesn't have to be actively not fun either, but the primary feeling should be that of effort. It should look a lot less like that "10 minute full body" workout from your local digital media creator and a lot more like a serious session at the gym. You want the mental equivalent of sweating. It's not that the quickie doesn't do anything, it's just that it is wildly suboptimal if you actually care to learn. I find it helpful to explicitly declare your intent up front as a sharp, binary variable in your mind. If you are consuming content: are you trying to be entertained or are you trying to learn? And if you are creating content: are you trying to entertain or are you trying to teach? You'll go down a different path in each case. Attempts to seek the stuff in between actually clamp to zero. So for those who actually want to learn. Unless you are trying to learn something narrow and specific, close those tabs with quick blog posts. Close those tabs of "Learn XYZ in 10 minutes". Consider the opportunity cost of snacking and seek the meal - the textbooks, docs, papers, manuals, longform. Allocate a 4 hour window. Don't just read, take notes, re-read, re-phrase, process, manipulate, learn. And for those actually trying to educate, please consider writing/recording longform, designed for someone to get "sweaty", especially in today's era of quantity over quality. Give someone a real workout. This is what I aspire to in my own educational work too. My audience will decrease. The ones that remain might not even like it. But at least we'll learn something.

The peace this cat radiates is the best thing I’ve seen on X lately 😊

What amazes me most is how unlikely this timing feels. Sometimes I think the real anomaly is not AI itself, but our timing. The universe is 13.8 billion years old, and it may last for trillions of years more. Yet we seem to have appeared almost at the very beginning of the beginning — right when intelligence is starting to leave biology and accelerate itself through AI. This level of anomaly makes me wonder: maybe we are not here, now, completely by accident. We are living extremely early in the possible lifetime of the universe, yet we happen to witness the moment when intelligence begins to accelerate beyond biology. Statistically, that feels almost absurd. Maybe “here and now” is not accidental.

Exactly 💯💯💯 People are uncomfortable when someone stronger exists next to them. We almost wiped out lions and most dangerous predators, and the ones that remain are preserved mostly for nature. The same applies to states. When one country becomes much stronger than another, its neighbors almost always start building up their own power. This is a fundamental part of human nature. That’s why I think fears about AI and robots completely taking over humanity are greatly exaggerated. Humans will control the entire process and most likely won’t allow AI to gain more power and influence than we have ourselves. The human role in the future will probably be largely about controlling robots. We won’t fully trust autonomous systems. Even if factories are filled with hundreds or thousands of robots, there will still be humans monitoring them through cameras, interfaces, or VR control systems. I think the number of robots in the future could reach trillions. They won’t work only on Earth, but across other planets too, while humans control them remotely. The profession of robot operator, observer, or supervisor may become one of the most common and important jobs of the future — maybe even humanity’s main activity. 🤖

Does anyone even know what's going on at @ssi?

Time to build major base on the Moon!

O M G 🤯 On the other hand, it’s not strange at all. @karpathy always loved research, and he’s more of a scientist than a sev/engineer. I think Anthropic is the best place for him.

Ebola in the DRC: According to the Congolese Ministry of Health, 131 deaths have been recorded out of 513 suspected cases. I built a live Ebola monitor for Africa: ida.am/ebola/

Those who come after us are better than us.

I made the Hantavirus Monitor AI-powered 🦠 It scrapes X every 15 minutes, then Gemini grades each post for source authority, quality, and factuality. Only credible signals surface — grouped by country, deduped, and live on the map. No human curation. ida.am/hanta/

It’s genuinely impressive when you realize that on X, most media pages like The Washington Post or The New York Times get almost no engagement.