Latent Space: The AI Engineer Podcast · Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy - Jason Ma, Dyna Robotics
Audio streams from the publisher’s own feed. Verbatim does not host or copy it.
used their own world-action model to generate synthetic training videos and actions, creating a self-fulfilling data flywheel
“We trained the final model on the SpaceX cluster ... a few weeks of pre-training, then synthetic data NRL and then shipping it”
“You need more pre-training data… give the model a bunch of data and think… transform this data… pre-trained on it would make the model smarter”
“Synthetic data generation techniques… take some amount of tokens and explode it into something much larger and pull learning signal out of that”
“First thing you ask is, are you using my data to train third-party AI models? And it's gotta be a hard no”
The strongest new moments on your topics, in the speakers' own words. Free for three topics.
No card. No password. One-click unsubscribe.
Pro watches up to 25 names or phrases and emails you the morning after they are said.
See Pro, $29 a monthVerbatim quotes a short extract and links to the episode; the episode belongs to its publisher. Remove a show.