Chelsea Finn

@chelsea-finn · 22 positions · On ppll ↗

Assistant professor of computer science at Stanford, and co-founder of the robot foundation-model company Physical Intelligence.

Chelsea Finn did not write this page.

We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.

What Chelsea Finn says most about

robotics 17

Reinforcement learning cannot be scaled for robots the way it was for language models, because every attempt spends real robot hours instead of data-centre compute.
Most state-of-the-art robot foundation models have no memory at all, and that is what stops them doing long multi-step tasks.

See all 17

LLMs 3

Reinforcement learning cannot be scaled for robots the way it was for language models, because every attempt spends real robot hours instead of data-centre compute.
Most future robotics training data will be robots' own autonomous attempts, the way most language-model data is now generated by the model itself.

See all 3

OpenAI 2

Robot models will reach ChatGPT-level capability within the next few years, even without a ChatGPT-style moment.
Robotics will not get a ChatGPT moment, because the distribution channel for a physical model is a physical robot and that cannot spread in five days.

See all 2

reinforcement learning 2

Reinforcement learning cannot be scaled for robots the way it was for language models, because every attempt spends real robot hours instead of data-centre compute.
One pre-trained robot model now matches or beats the specialists that were fine-tuned with reinforcement learning for the very tasks they were built for.

See all 2