@thomasfbloom I have my own blog post that I wrote about 3 months ago (which might be already outdated, but I hope it could be still helpful for someone) https://seewoo5.github.io/jekyll/update/2026/03/05/early-mathematician-ai.html which is highly motivated by Jeremy's essay https://arxiv.org/abs/2603.03684

@phillennium It can be hard to pin down the exact date, and hard to know when to draw the line (eg the Klowden-Tao article was written over a long period I believe).

@thomasfbloom A friend (a different Sam) wrote this recently: https://sam0287.github.io/blog/posts/2026-05-22-will-ai-ruin-math.

@thomasfbloom Randomized order is a nice touch, but maybe chronological order could be more helpful in the future to track the changes over time

@thomasfbloom May I suggest the short piece titled "Prior to replacement" here? I wrote it some months ago. https://www.begin-proof.com/some-thoughts

@thomasfbloom I enjoyed Dmitry Rybin's "On LLM Math Capabilities"

@sp_monte_carlo Added, thanks!

@antimath3 Added, thanks!

@thomasfbloom The thread featuring the discussion on theory and problem-solving between @littmath and @Jacob_Tsimerman is an absolute must read!

@thomasfbloom https://math.iisc.ac.in/~vamsipingali/ThoughtsonAI.html

@phillennium But I agree that when reading these, an important piece of context is when it was written, with how fast things are moving.

@thomasfbloom I have a long going series of AI assisted critical reviews of AI in mathematics papers/claims covering systems from @OpenAI , @EpochAIResearch , @axiommathai, @GoogleDeepMind at https://aichats.substack.com/