nafyi.

PUBLIC VIEWPOINTS

John Schulman

Interview participant ·

1 interviews · 3 viewpoints · 1 topics

Content updated:

Translations are for reading; original excerpts remain the evidence.

Viewpoints by topic

Attributed viewpoints, ordered by interview date. A snapshot of these conversations, not a definitive statement of someone’s beliefs.

AI research

View this topic
AI research

Generalization is difficult to predict

John Schulman says the field depends heavily on generalization and that it is difficult to predict when generalization, including out-of-distribution generalization, will occur.

Supporting evidence

Original excerpt

The whole field relies a lot on generalization and it’s very hard to predict when you’re going to get generalization, or when you’re going to get some kind of out-of-distribution generalization

John Schulman·

AI researchers debate how close we are to recursive self-improvement

Additional context is not included in this excerpt. Read the original conversation.

Open original · 9:42 (opens in a new tab)Transcript source (opens in a new tab)

Start time comes from the supplied transcript. Playback alignment is awaiting review.

Open the episode and seek to 9:42.

John Schulman
Share insight
AI research

Research methods may have substantial room to improve

John Schulman suggests that clever small-scale experiments could support theories that generalize to larger experiments. He expects research methods to be far from their ceiling.

Supporting evidence

Original excerpt

If you think hard enough, you probably could have expected some of these things beforehand. There is probably some very clever way to do a small-scale experiment that’ll let you build the theory that then will generalize to the large-scale experiment. So I would expect that we’re nowhere near the ceiling of how well you can do research

John Schulman·

AI researchers debate how close we are to recursive self-improvement

Additional context is not included in this excerpt. Read the original conversation.

Open original · 11:34 (opens in a new tab)Transcript source (opens in a new tab)

Start time comes from the supplied transcript. Playback alignment is awaiting review.

Open the episode and seek to 11:34.

John Schulman
Share insight
AI research

Distillation depends on a realistic prompt distribution

John Schulman says a realistic prompt distribution can help a distilled model match a larger model. He contrasts this with easily verifiable tasks, which may produce strong benchmark results but weaker performance across a broader distribution.

Supporting evidence

Original excerpt

If you have a really good realistic prompt distribution for distillation, you can match the big model really well. But 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

John Schulman·

AI researchers debate how close we are to recursive self-improvement

Additional context is not included in this excerpt. Read the original conversation.

Open original · 26:48 (opens in a new tab)Transcript source (opens in a new tab)

Start time comes from the supplied transcript. Playback alignment is awaiting review.

Open the episode and seek to 26:48.

John Schulman
Share insight

Statements by interview date1

Statements are ordered by the original interview publication date; differences in wording do not establish a change of position.