Everything in AI/tech this week in 5 minutes

Happy Sunday. Here's what actually mattered in tech this week, broken down so it's worth your five minutes, then a bit on what's been on my mind. Let's get into it.

Nvidia is buying Hugging Face for a reported $13 billion. If you have ever downloaded a model, you probably have used Hugging Face, it is the open-source hub the whole ML world runs on. Nvidia grabbing it pushes them way past chips and into the models and tools layer, and it is a return to the software and cloud side they had pulled back from a year ago. The deal is reported and not signed yet, so it could still fall through. What it means: the hub the whole open-source community leans on would now be owned by the company that sells the chips, and whether it stays neutral is the thing worth watching. Read more

Meta settled child-safety claims with 29 states for roughly $17 billion. The states argued Meta deliberately built Facebook and Instagram to be addictive for kids, and it is now the biggest payout in tech history. Beyond the money it forces real product changes, time limits for under-18 users and stricter age checks. It is the clearest sign yet that the move-fast era of consumer tech is running into real accountability. Read more

China ran a full Olympics for humanoid robots. Over 2,000 robots from 16 countries competed in sprinting, soccer, kickboxing and taichi in Beijing, more than double last year's edition. One of them ran the 100m in 8.64 seconds, faster than Usain Bolt, though there is a padded wall past the finish line because the robots are still much better at sprinting than stopping (i hope you saw the funny videos as well haha). Behind the spectacle it is a real flex, China is going after humanoid manufacturing the same way it took over EVs and solar. Read more

OpenAI says its first custom chip beats Nvidia's best. The chip, codenamed Jalapeño and built with Broadcom in just 16 months, reportedly outperforms Nvidia's Blackwell on AI work per watt. For years every lab has rented Nvidia's chips at Nvidia's prices, so a lab building silicon that beats them is a big deal. What it means: if the labs can make their own chips, compute gets cheaper and less tied to one supplier, and cheaper compute eventually flows down to everyone building on top. Read more

Sometimes you have to burn the bridges. Some of you might have seen that in an unexpected turn of events I got accepted into the university of oxford this week. And when I read that email I was shocked, I just didn’t know how to react. I felt like this was one of the biggest opportunities I ever got and a lot of my friends acted like it was the obvious choice to take it. But here I was in switzerland, just moved, enrolled into ETH, just started a new company. I was locked in. And then I thought back to what made me apply to ETH in the first place. It was the 1.5 years of world class education and the opportunity to meet some of the most cracked people in the world. Oxford would without a doubt be the heavier hitter on my CV and purely optimizing for prestige the choice would be easy, but I didn’t want to do that anymore. I feel happier chasing my passion instead of prestige and at this exact moment life was testing me. Giving me the absolute gem of prestige and academic status right before my eyes and seeing if i would fold.

I didn’t fold and declined the offer 30 minutes after getting it, essentially burning the bridges and leaving me with no choice but to lock in and do everything I said i would in switzerland. Thanks for all the nice messages, you guys keep me going.

🎬 My latest content:

🔨 Tool of the week: Hubspots free marketing certification. Clicking the link supports the Newsletter :)

From ideation to measurement, HubSpot Academy's free Content Marketing Certification teaches you to build a scalable content strategy that converts. Join 200,000+ professionals who have advanced their career with HubSpot Academy. Enroll free and walk away with a strategy built to last.

Get Started Free

Have a great week,

Chris