Should math students still pursue a PhD as AI advances?
What happened
On October 9, Hacker News featured a guest post by Álvaro Lozano-Robledo. He argues that students who love math should not give up a math PhD just because large language models are advancing, and says his own research output tripled in a few months. He treats LLMs as research collaborators but stresses they still need a mathematician's guidance and intuition, and that math education, understanding and communication keep their value. On whether AI will eventually surpass all human math research ability, he says plainly he cannot know, and proposes a hypothesis that model capability is bounded by the "convex hull" of existing mathematical ideas.
Written by AI from the coverage · updated 2 hours ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pagePickÁlvaro Lozano-Robledo: what to tell math students as AI advances
In a guest post, Álvaro Lozano-Robledo argues that students who love math should not drop plans for a math PhD just because LLMs are improving. His own research output tripled in a few months, and he treats LLMs as research collaborators that still need a mathematician's guidance and intuition. He also says he cannot know whether AI will surpass all human math research ability, and offers the "convex hull" of existing mathematical ideas as a hypothesis for where model limits lie.
Heat over time
Not enough continuous observations to draw a trend yet.