✳flâneur — a map of the web's best reading

Thoughts - Albert Qiaochu Jiang

albertqjiang.github.io · 123 words · saved by 1 readers

The general atmosphere is contagious: the echo chamber of AI is indeed comfortable and exciting. When so many people get together to have low-latency conversations that plastically shape their views, an electrifying air ensues. I have been enlightened countless times during a mere week. However, it’s worthwhile to remember that it is still an echo chamber and we should ground ourselves firmly in reality >95% of the time to not be intoxicated and completely detached. Two main themes dominate the conference for me: AI for maths and language models. Here I want to discuss them, first individually and then how they relate to each other. More people than ever are getting interested in AI for maths. The tutorial on machine learning for theorem proving captivated a room of ~300 people; For the first time I heard “theorem proving” being mentioned in the halls of a machine learning conference venue by random people. Behind the popularity are two seemingly contradictory views: one is that AI is

Thoughts - Albert Qiaochu Jiang You are using an outdated browser. Please upgrade your browser to improve your experience. Thoughts The Agricultural Revolution of Mathematics Published: March 30, 2026 Why am I doing formal maths/coding now. The weight of poems Published: August 02, 2025 I find myself constantly wandering in bookshops looking for new poems and poets, often without any success. I want to achieve mathematical superintelligence, so why am I doing LLMs now? Published: December 10, 2024 I’ve recently defended my PhD, which is on AI for formal mathematics. Now I’m a full-time researc

Explore this link on the map →

related reading