OpenAI robotics team called dead, Chris Paxton pushes back
A one-line claim on X, a rumor about an empty conference booth, and a researcher saying hold on.
NIK posted that OpenAI has officially disbanded its robotics team, five months after relaunching it. The post ended with two words, "it's over". No OpenAI statement is cited anywhere in it.
Carlos DP, who was at CoRL, said the claim lines up with what people were saying at the conference. His read was the empty OpenAI booth, which he called conspicuous, plus hallway rumors. But his version is not that the work stopped.
What Carlos DP actually said
His post says robotics is still being worked on at OpenAI, and that it has been reorganized around the core large language model as the reasoning solution. That is a different claim from a team being shut down. It is a team being folded into something else, if it is true at all, and he hedges with "it seems".
Chris Paxton pushed back on both. Replying to Carlos DP, he wrote that there were hard-core, real roboticists at CoRL from OpenAI. In a separate post he said he does not think this means OpenAI is giving up on robotics.
There were hard-core, real roboticists at CoRL from OpenAI
Shown versus claimed
Nothing here is sourced to OpenAI. The disbanding claim is one post with no attribution. The supporting evidence is a booth that was empty at a conference and rumors that circulated there. The pushback is one researcher saying he personally saw OpenAI roboticists at that same conference.
So the strongest thing anyone in these posts can actually point to is who showed up at CoRL, and the two sides disagree about what that means. An empty booth can mean a team was cut. It can also mean nobody booked a booth. People attending without a booth is exactly what Paxton describes.
The reason this spread fast is that OpenAI relaunching robotics was itself news, and a reversal five months later is a clean story. It is also the kind of story that is easy to post and hard to check. If OpenAI has restructured the work rather than killed it, the public signal would look similar from the outside, which is roughly the gap Carlos DP and Paxton are arguing in.
Nobody knows yet. Until OpenAI says something, the state of the record is one unattributed claim, one conference-floor read that partly contradicts it, and one researcher who was in the room saying the people are still doing the work.
I don't think this means OpenAI is giving up on robotics
Sound like you were at the demo
One email a day. Under three minutes. The demos, the deployments and the part the press release left out.
