Talking to normies about AI, constantly flipping between these 2 moods

Not only does ChatGPT know Farsi (you can both query in Farsi, and ask for answers in Farsi even when you type English), but it also understands Finglish!! i.e. typing Farsi using latin characters!! 🤯🤯

a short teaser of what we’ve been up to lately ⁦@stanford⁩: here is an end-to-end policy running on low-cost arms!

It's surprisingly easy to undervalue engineering contributions in ML research. During conference reviews one of the criticisms of the Imagen paper was that "it's just diffusion scaled up". Things often break in interesting ways at larger scales.

🎉Last week, amid the ICML rush, we were informed that our paper (w/ Sonali Parbhoo and Marzyeh): "Risk Sensitive Dead-end Identification in Safety-Critical Offline Reinforcement Learning" was accepted to Transactions on Machine Learning Research! 🎉 Short thread introduction 👇 🧵 (1/7)

Can text-to-video generation help decision making? Introducing UniPi, which acts by synthesizing a video of what it will do: universal-policy.github.io UniPi can generate diverse videos/actions across many environments (and combinatorially generalize!): (1/6)

Eric Jang

Tried to hit Ctrl+Z when I made a mistake while writing on the whiteboard... What have I become... 😱😳

Really cool work from Danny et al.!

Computer vision has been solved.

Was having an audio conversation with ChatGPT today. At some point its voice got phlegmy and then it coughed… EXCUSE ME WHAT?!

Led by Google DeepMind, we present ALOHA 2 🤙: An Enhanced Low-Cost Hardware for Bimanual Teleoperation. ALOHA 2 🤙 significantly improves the durability of the original ALOHA 🏖️, enabling fleet-scale data collection on more complex tasks. As usual, everything is open-sourced!

Introducing Ideogram 1.0: the most advanced text-to-image model, now available on ideogram.ai! This offers state-of-the-art text rendering, unprecedented photorealism, exceptional prompt adherence, and a new feature called Magic Prompt to help with prompting.

Huge congrats to Sherry Yang for the beautiful results in this work!! Amazing work by the team Yilun Du,Jonathan Tompson, Leslie Kaelbling, Dale Schuurmans, Pieter Abbeel with a cameo from yours truly.

I’m excited to introduce RoboCasa, a large-scale simulation framework for everyday tasks. Scaling is the key driving force to unlocking generalist robots, and RoboCasa leverages simulation to take scaling to a whole new level. A short 🧵