Mathematician Calegari demos Claude-assisted research and reflects on learning math in the AI age
At the University of Chicago's "Mathematics in the Age of AI" symposium, mathematician Danny Calegari gave a talk on how mathematicians can coexist with AI; theoretical computer scientist Lance Fortnow summarized its key points in a tweet thread.
Confirmed
- Calegari demonstrated live how he uses Claude to collaborate on math research, visualize mathematical objects, and even generate the slides for the entire talk from a single prompt.
- On the value of learning math with AI, Calegari argued there is no value in chasing new theorems with AI—"nobody is impressed by good results anymore"—because you only truly learn by proving things yourself; the process of learning math is what matters.
- He stressed that "deliverables are almost never the answer"—the proof itself was never the goal—and that if an idea came from Claude, you should openly admit it, "for the sake of your own soul."
- On the conversational experience of LLMs, Calegari said people describe LLMs as mirrors that reflect yourself back, which is flattering, but he thinks they "perform better than you"; Fortnow echoed this at the end of the thread, noting that in long conversations LLMs outperform him because the dialogue keeps a record of errors and dead ends; the full video is available via the link he provided.
Why it matters
- The talk represents a frontline mathematician's dual attitude toward AI-assisted research: actively using Claude for collaboration, visualization, and output (like one-shot slide generation), while staying clear-eyed about the learning value of "chasing results with AI" and disciplined about honest attribution.
- The "mirror, but better" framing offers a concise lens on the unique value of long LLM conversations, endorsed by both computer scientists and mathematicians—worth the attention of AI tool designers and researchers alike.
2026-10-24 ~ 2026-10-24 · 6 related posts
Primary sources
- UChicago mathematician Danny Calegari shows how he uses Claude for math collaboration and slide generation — fortnow · 2026-10-24
- If Claude Had the Idea, Say So 'For Your Own Personal Soul': AI and Math Ethics — fortnow · 2026-10-24
- [source] No One Is Impressed by Good Results: Why AI-Proved Theorems Teach You Nothing — fortnow · 2026-10-24
- [source] Mathematician Danny Calegari: LLMs Mirror You Back, But Behave Better — fortnow · 2026-10-24
- Computer scientist Lance Fortnow on long LLM conversations: 'they behave better than you' — fortnow · 2026-10-24
1 near-duplicate retellings: fortnow