team carbon. shadowy dev. hyper optimizor
Tierra del Fuego Antarctic and
Joined February 2010
- Tweets4.3K
- Following4.9K
- Followers7.9K
- Likes18.9K
robot arms are the bane of my existence
the unitree was easy pickings
arms?
every codebase is different no wonder everybody just shares photos
dark gpus era is coming
Business Insider confirms it: Amazon has warehouses of GPUs that can’t be installed, and is buying them so far in advance that a supply chain team said they “didn’t have a clear forecasted path” to installing them.
businessinsider.com/amazon-2…
buy a bunch of cameras, create ur own data sets, take big bets, if it doesn't work out become a data broker
need this kind of overcapacity
Tech blogger 神奇阿宇 from Hubei, China, built a working hoverboard.
Electric, with 10 brushless motors pushing over 30kg of thrust each—enough to lift 4 people—and peak current hitting 2,000 amps.
2 STEPX Neo phones as the board's brain. One inside the board reads posture data in real time; the other on his wrist feeds AI voice prompts through his earbuds—"Weight's shifting left!" "Bend your right knee!"
After 10+ days of brutal training and AI-assisted correction, he achieved man-machine sync.
penguin retweeted
Models capabilities are data bound,
Therefore the further the;
F(model + compute + training recipe)
is from the data source, the more polluted and expensive it gets
Kind of like an inverse square law
penguin retweeted
One question I get from business owners all the time: How to turn everyday work into robotics training data?
Now, I finally write an article about that.
The best way to "catch up" on the physical AI transformation is to embrace it.
penguin retweeted
We are hiring data operators in our SF office. Please apply.
GI Labs is hiring Robot Data Collection Operators in San Francisco 🤖 Work hands-on with robots and teleop systems to collect the data that trains next-gen robots. Part-time, in person. No robotics experience needed. Apply: [email protected] gilabs.xyz/join/robot-data-c…
penguin retweeted
My first piece for @NewYorker is out today!
As Xi Jinping lands in DC to start his state visit, you'll hear a lot about the US-China AI race. I'd like to tell a complementary AI story.
This past May, a few days after Trump visited Beijing, I went to the town people call China's vibe-coding capital, Liangzhu. There, you learn that China's AI scene is not monolithic. It's also full of contrarians, cool vibe coders, spiritual seekers, and people who love AI but refuse to join the 996 grind.
Liangzhu is on the edge of Hangzhou. I hung out with a group of anti-mainstream builders who use AI to make whimsical apps for fortune-telling, mental well-being, and creativity. Some of them run their lives entirely by bazi astrology (their founder almost became a Taoist monk). They meet on "Crazy Thursday" to vibe-code together. Rent is cheap, so they can get a bigger place with a living room to have friends over (unthinkable in Beijing or Shenzhen). In Liangzhu, they want to use AI to forge their own future.
It’s a ground-level story—less about model capabilities, hyperscalers, or AI policy, and more about how people in China are using AI, and the human stories behind it.
can’t eat snow anymore 😭
When you go skiing in the winter, and see those big machines spitting out snow, know that they are spitting out purified proteins made from dead bacteria.
I didn’t realize this until recently! Basically, they take Pseudomonas syringae cells — not engineered; a natural microbe found on plants — and grow them up in huge bioreactors. The cells make ice-nucleating proteins, which sit on their outer membranes and help ice crystals form. The cells are concentrated, freeze-dried, and sterilized until only these proteins remain.
The reason these proteins are useful is because they help water molecules form ice more quickly. Water can remain liquid even below 0°C, and only freezes when enough molecules form into a crystal. The proteins have, on their surfaces, repeating snippets of amino acids that help water molecules organize into that shape.
So these snow machines just blast super cold water with these proteins, and the two hit the air, mix together, and immediately freeze into snow.
proof of work blockchains could make a huge comeback
SITUATION EXPLAINED: Why are OpenAI and Anthropic suddenly buying tiny data centers?
• OpenAI and Anthropic are both looking at deployments of 20 to 30 megawatts
• Anthropic has sounded out agreements in that range across the UK and the Nordics. OpenAI has been exploring the Nordics too
• The reason is inference. It was 9% of global data center workloads in 2025
• By 2030 it's projected to take 37% of capacity against 13% for training, with the crossover in 2027
• Crusoe is investing in smaller data centers too, and raised $3.9 billion yesterday at a $30.9 billion post-money valuation
@theojaffee: "It's much easier to build 20 megawatts and get that online than two gigawatts, 100 times larger."
