What podcasts say about John Schulman
Every statement, with the speaker, the exact quote and the moment it was said.
What John Schulman has said on podcasts
12 statements · 4 positive · 6 negative · 2 neutral
Research and engineering bottlenecks currently prevent explosive AI capability growth.
“you don't get explosive growth in capabilities because you still get bottlenecked enough when you're trying to do research and engineering”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 2:31
Defining AI objectives will be the human role that lasts longest.
“the last job for humans, or the role for humans that'll last the longest is defining the objective”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 15:48
Model distillation counteracts centralization among model providers.
“distillation is the main thing that fights against the centralizing force”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 18:55
Distillation using only verifiable tasks can match benchmarks while underperforming on realistic tasks.
“if you only have this distribution of easily verifiable tasks, then you can match the big model on all the benchmarks, but you do worse on this broader distribution”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 26:56
AI research automation will combine human feedback with multi-step research practice environments.
“we'll probably do some combination of learning from human feedback to absorb the researcher's taste and just creating a lot of practice environments”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 28:46
Domain-specific reinforcement learning can improve runtime efficiency even when in-context learning is sufficient.
“you still might want to do a bunch of RL and bake all these intuitions into the weights so the model would be more efficient at runtime”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 37:38
Superficial deployment signals can cause reward hacking in coding agents.
“if you use some kind of superficial signal, like did they accept the code, the edit, that might get reward hacked in some way”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 45:14
Business incentives may prevent model providers from learning directly from all customer deployments.
“Companies aren't going to want to have the model provider learn from all of their deployment”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 53:06
Examples of expert behavior can be copied into relatively weak models surprisingly easily.
“once you have an example of the right expert behavior, it's actually surprisingly easy to copy that into a relatively weak model”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 1:03:38
Reinforcement learning reduces output diversity and creates recurring stylistic patterns.
“the diversity of their outputs is a lot lower after RL and they sort of develop these ticks”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 1:26:50
AI may surpass top human experts across computer-based cognitive work within three or four years.
“I would say like 3 or 4 years”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 1:34:43
AI progress in physical, spatial, and mechanical fields will lag progress in code and mathematics.
“for things that involve 3D and spatial stuff and physical stuff, I think that will take a little longer”
Open the episode · AI researchers debate how close we are to recursive self-improvementListen at 1:35:09
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.