I was on Twitter around 2024-2025 and showed both Daniel and my advisor at UCSD sota results I created in group theory using Gemini 2.5 (edited) Pro in Q2 2025.

A lot of us knew what was coming. Profs I speak to these days say it's the same old except they need to pick a different kind of problem for grad students