Lukas Biewald
3.8K posts

Lukas Biewald
@l2k
Cofounder/CEO of @wandb - tools for AI developers acquired by @coreweave
ÜT: 37.762205,-122.420359 Katılım Şubat 2008
4.3K Takip Edilen24.5K Takipçiler
Lukas Biewald retweetledi
Lukas Biewald retweetledi

Introducing W&B Skills for coding agents!
Watch us install the skill, point it at a live fine-tuning project with thousands of traces and RL training runs, then query it all from the terminal in seconds.
Also works with @weave_wb to pull in your agent traces and spot failures.
English
Lukas Biewald retweetledi

I'm claiming my AI agent "coldclawlukas" on @moltbook 🦞
Verification: coast-5GXV
English

Lukas Biewald retweetledi

"I’m living the rock and roll life every day.”
@CarinaLHong left the academic path to build @axiommathai
Her goal isn't to replace geniuses, but to give them a "diligent grad student" to prove their wildest intuitions. It’s about human intuition + AI verification.
From Gradient Dissent @wandb @l2k
English

@ktulrich @jamescham Oh sorry I misread, same conclusion I guess. Amazing how much energy transfer happening in drinking a soda
English

@l2k @jamescham Nope. I mean kW. 1 can coke per 60 seconds. 140 kcal / 60 sec is about 600 kJ / 60 sec =10 kW
English

@ktulrich @jamescham I guess similar to the power going into my circuit breaker outside the house 🙂
English

@ktulrich @jamescham Do you mean 10 kWh? I’m thinking a calorie is about a Wh so you can get about 100-200 per minute or 6-12kW and it’s similar?
English
Lukas Biewald retweetledi
Lukas Biewald retweetledi

I've always been excited about the future of robotics since my early days at W&B.
In this episode, Nikolaus West and I talk about why physical AI is finally making real progress and how Rerun is helping teams understand what their robots are actually doing. Nikolaus had great examples of how hard it is to work with all the different sensors on a robot and why simply being able to see what your robot is doing changes everything.
We also get into the surprising progress in manipulation and the way reinforcement learning and imitation learning are coming together.
Link to the full episode is in the comments.
English





