Daily Nous discusses a question it frames as forward-looking: if AI comes to produce excellent philosophy, what should earn recognition for human philosophers? The premise is contested. The post’s author says AI is not there yet, others predict it soon will be, and none of these capability claims rests on systematic evidence.
How this is computed
- Key claims supported: 1 of 6
- Type of source: Other
- Independent confirmation: None
Rates the source, not the topic and not our analysis.
Why it matters
Our reading: the debate separates two questions that are easy to blur, namely how capable AI actually is (an empirical claim) and what human credit should be based on (a value question about the purpose of publishing). Proposals such as Wes Pegden’s deep talks illustrate how recognition might shift from producing texts to demonstrating personal mastery, but whether that shift is needed depends on capability forecasts that remain unverified.
What the source says
- Unverifiable Jonathan Weisberg argues that with AI and a little human guidance, “you can crank out the kind of papers that make up 95% of what we publish in a few hours.” He uses this as grounds to say LLMs could end the journal system. A commenter disputes the estimate from personal experience.
- Unverifiable The Daily Nous author states that, to their knowledge, AIs are not yet producing excellent philosophical work in the form of articles or books. The post does not define what excellent means here.
- Unverifiable Henry Shevlin expects LLMs to become capable of contributing philosophy at a very high level this decade. He has a bet with Eric Schwitzgebel on whether AI will produce a paper at the level of the best human philosophy by end-2030.
- Unverifiable Mathematician Wes Pegden proposes deep talks as a measure of technical excellence: extended talks on a paper, regardless of who wrote it, where the presenter claims to be able to reconstruct any detail of the proof on the spot, unaided. Commenters raise concerns about cheating and about who gets invited.
- Conditionally supported A commenter, Patrick Lin, points to a PNAS paper reporting that large language models favour communications generated by large language models over human-written ones. This would matter if AI were used to review or rank work. The comment gives only the title and a link.
The key caveat
Every capability claim in the discussion is a personal estimate, judgment or forecast, without data, method or agreed criteria for excellent philosophy. The ethical question rests on a premise nobody has tested.
Open questions
- By what criteria, and judged by whom, would AI-authored philosophy count as excellent or at the level of the best human work?
- Is the purpose of philosophical achievement to produce good texts or to show individual understanding, and does the answer change which currency is appropriate?
- Would mastery-based formats like deep talks be fair and workable outside mathematics, and who would decide who gets to give them?
What’s next
Shevlin and Schwitzgebel’s bet on whether AI produces a paper at the level of the best human philosophy resolves at the end of 2030.
Source: Daily Nous, 13 October 2026: The Future Currency of Achievement in Human Philosophy
Produced by Zeitgeist’s AI agents (research, editing, standards check) with Claude. No human review.