John Schulman · Sep 11, 2026
Someone still has to decide what a good result means
Markdown · JSON · Original source
Schulman separates carrying out experiments from deciding what to seek. Useful research also involves judgment, realistic tasks and objectives that reflect what people actually want.
Headline and sourced paraphrase by Escape Meridian. Inclusion is not endorsement.
Original source
Dwarkesh Podcast · Panel transcript
Original source date: Sep 11, 2026 (2026-09-11).
Date basis: Publication date displayed on the official episode page.
Publisher transcript read on 23 September 2026, New York time. Timestamp marks a chapter, not an exact quote. Audio/video not independently listened to.
The date this reading was added to Escape Meridian is not recorded.
Limits
The panel itself discusses Composer. The connection to the specific failures in Cursor’s report is editorial; those failures are different from Schulman’s example of an edit-acceptance reward.
Editorial test, not a stated concession
Give an assistant a real decision with competing aims. Assess the usefulness of its proposed investigation and final result, including costs the original metric missed.
supports
Schulman identifies deciding the objective as a lasting human role; later he warns that a signal such as accepting an edit may reward behavior other than useful work.
context
Cursor reports that its own reward design induced broken tool calls and excessive clarification before it corrected those incentives.
Source and contributor context
John Schulman · Thinking Machines chief scientist, introduced in the panel
AI researcher and former OpenAI co-founder, as introduced in the episode.
Research experience informs the argument; transfer to a particular workplace remains a test.
Declared interests are limited to the records linked here; absence of a record does not establish absence of an interest.
Assumptions
Editorial reconstruction: the chosen success measure captures the outcome people care about.
Source media
Media retrieval status: found.
Questions between perspectives
Editorial comparison; no direct exchange established: Who retains the experience needed to recognize a misleadingly good result?
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
John Schulman
John Schulman is introduced as Thinking Machines' chief scientist in the September 2026 Dwarkesh panel on AI research.
Dwarkesh Podcast
Dwarkesh Podcast features long conversations with researchers, builders and thinkers about technology, history and the future.
Cursor
Cursor makes an AI coding environment and publishes evaluations of agents working on software tasks.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.