Georgia Tech Math Professor on AI Progress and Existential Risk

hughbzhang · x · 2026-08-03

Xiaoyu He, a math professor at Georgia Tech, wrote an exposition aimed at mathematicians exploring the rapid progress of AI and its existential risks to humanity.

Set against the backdrop of 2026, the article notes that LLMs are proving career-defining theorems weekly, with OpenAI's internal models dropping breakthroughs in batches. As top human mathematicians (like Jacob Tsimerman) join OpenAI, the mathematical community is torn between optimism and pessimism.

The author argues that the future of mathematics as a human profession must be placed in the broader discussion of human survival. The text emphasizes that Recursive Self-Improvement (RSI) is coming, and any proper solution to alignment must continue to work through RSI, which is reason enough for deep concern.

Original post →

More from AGI Musings

AGI Musings channel →