Max W. Shen

@kchonyc arxiv.org/abs/2512.02303

Check out our poster and oral presentation at NeurIPS AI4Science, and I'd be happy to chat! I am a researcher in Frontiers Research in Prescient Design, Genentech with @kchonyc and we may be hiring soon!

Are equivariant architectures worth it for guaranteeing 3D-rotational symmetry when modeling proteins? We study this in *Training Dynamics of Learning 3D-Rotational Equivariance* (Oral at NeurIPS AI4Science '25, accepted at TMLR) 🧵👇

@rishabh16_ arxiv.org/abs/2512.02303

@rishabh16_ Try googling for it :)

argmax.blog/posts/deep-ret… The deep retrieval + remix hypothesis

argmax.blog/posts/data-two… What is data? The classical and postmodern views

@SsbmGinger @stocmoney I was PR #2 in San Diego in 2013 btw so others can attest than I am also a real person

@SsbmGinger @stocmoney I can attest fro is a real person lol. At MIT around 2016 I was better than fro ybot and unlucky. Fro got insanely better after slippi came out and last I heard he's in NYC

Alright has anyone ever met Fro116? Is this a real person? Spare me with the “he’s a really busy guy who goes to MIT so he has no time for in person tournaments” My guy has 500 games on ranked in a week LOL

In capitalism, narrow AI may be the greatest threat to AGI. argmax.ghost.io/on-creating-ag…

@tzenszen Thanks for the kind words!

@maxwshen I enjoyed this blog post and catching up on the previous ones. Thanks for putting them together!

New blog post - argmax.ghost.io/treating-high-…. A unifying and simple concept - iteratively refining high dimensional output seems to work surprisingly well in practice

@StewartSlocum1 Thanks!

@maxwshen Insightful post!