RSI in mathematics?
The concept of Recursive Self-Improvement (RSI), traditionally a goal in software engineering and AGI development, is now being considered in the context of fundamental mathematics. Inspired by the self-referential nature of mathematics and recent advancements, such as the 700 proofs from OpenAI, there's interest in creating a recursive improvement loop. This would involve feeding new mathematical proofs back into a training corpus to identify further implications, thereby continuously advancing mathematical understanding.
时间与来源
时间显示为 UTC
显示时区：UTC
本地时区尚不可用，暂时显示 UTC。
发布当时偏移：UTC+02026年10月29日 12:11 UTC
收录当时偏移：UTC+02026年10月29日 13:00 UTC
- 发布
- 2026年10月29日 12:11
- 收录
- 2026年10月29日 13:00
- 来源类型
- 开发者社区
- 档位
- 社区
- 信源状态
- 正常
档位是按信源手工设定的编辑判断，不是逐条打分。
recursive self improvement has been a goal for a long time. and I think we have been thinking about it in terms of SWE and building towards AGI. watching the insane progress in math, and how mathematics is built on itsself, I was wo dering about a recursive improvement loop in fundamental math. where the 700 proofs from OpenAI get fed back in to the training corpus and it's told to "find the next implications" and looped.