I didn't sign the Fields medallists' letter
2 hours ago
- The author recounts an early childhood attempt at proving Fermat's Last Theorem, which failed but spurred a lasting interest in mathematics and difference sequences.
- They emphasize that thinking about problems, regardless of solving them, is a key learning experience, as highlighted by a letter from 25 Fields medalists.
- The author declined to sign the letter despite agreeing with its core points, preferring to offer a nuanced personal perspective via blog post.
- They clarify their stance: they are not opposed to AI, having worked on automatic theorem proving, but acknowledge the 'bitter lesson' that AI can now match human problem-solving without deep insights.
- The primary worry is that AI could discourage new generations from pursuing mathematics, potentially eroding the human mathematical tradition.
- They explore the impact of AI on individual and collective mathematical understanding, noting both potential benefits (e.g., more results to digest) and risks (e.g., loss of deep understanding).
- They suggest that the crisis may be temporary, with AI potentially helping digest results, but fear the collapse of social structures that support mathematical learning.
- They call for imaginative supervision of PhD students and better communication of mathematics' value to policymakers.
- They argue that the problem of credit systems is transitory, and that AI may not be entirely negative if it leads to better understanding of results.