Profile avatar
samlaki.bsky.social
Emissary at the machine frontier. this site is green baby
79 posts 383 followers 320 following
Regular Contributor
Active Commenter
Posts 24 Comments 26

Delighted to share our new position paper: arxiv.org/abs/2412.16075! The o1/o3 path to math reasoning is based on LLMs and large-scale test-time search. We argue for a different path that uses formal proof assistants for ✅ creating high-quality synthetic data ✅ rigorous test-time feedback. (1/2)

submitted 63 days ago • 2 comments

Mini thread on developer trust for #AMD. Effectively: the default assumption for AMD is that the stack will be broken. The community tried to help, but as Dave Airlie said years ago “throwing source code over the fence doesn’t make you FOSS” #HPC #ROCm

submitted 62 days ago • 2 comments

youtu.be/gRqoXy-0d84

submitted 67 days ago • 6 comments

Splendid stuff; very digestible and usable: arxiv.org/abs/2412.05239 'Conditions for uniform in time convergence: applications to averaging, numerical discretisations, and mean-field systems' - Katharina Schuh, Iain Souttar

submitted 77 days ago • 1 comment

wow TTT is truly incredible.... @pinkddle.bsky.social how were your experiments with it??

submitted 77 days ago • 1 comment

structured matrices save me

submitted 79 days ago • 1 comment

submitted 87 days ago • 0 comments

i need a huge TV as a monitor fr

submitted 88 days ago • 0 comments

i need a menhera

submitted 88 days ago • 2 comments

Just had a new paper hit the arxiv (will appear at NeurIPS '24), joint with Sungjin Im, Thomas Lavastida, Ben Moseley, Aidin Niaparast, and @vsergei.bsky.social: arxiv.org/abs/2411.16030 . I think it's super cool, so a quick thread!

submitted 89 days ago • 3 comments

how it feels to use a fountain pen for the first time in 10 years and write absolute chicken scratch

submitted 90 days ago • 0 comments

arrgh goddamn roommates have a party with a loudass speaker

submitted 92 days ago • 0 comments

submitted 92 days ago • 0 comments

question for ML people out there: is there any model/ optimizer/ algorithm out there (for any kind of ML task) that will make your life much easier but it incredibly godawful slow to be worth considering practically?

submitted 93 days ago • 3 comments

submitted 94 days ago • 0 comments

Super fun Time earlier today at the “Democratizing Al Accelerators for HPC Applications" BoF!

submitted 95 days ago • 1 comment

video uploads take a lil bit longer

submitted 96 days ago • 1 comment

██╗ ██╗ ██████╗ ██████╗ ██╗ ██████╗ ██║ ██║██╔═══██╗██╔══██╗██║ ██╔══██╗ ██║ █╗ ██║██║ ██║██████╔╝██║ ██║ ██║ ██║███╗██║██║ ██║██╔══██╗██║ ██║ ██║ ╚███╔███╔╝╚██████╔╝██║ ██║███████╗██████╔╝ ╚══╝╚══╝ ╚═════╝ ╚═╝ ╚═╝╚══════╝╚═════╝

submitted 96 days ago • 1 comment

submitted 96 days ago • 0 comments

wow this fella supports avif uploads too

submitted 97 days ago • 0 comments

time for a rewatch

submitted 97 days ago • 1 comment

need

submitted 100 days ago • 0 comments

given the recent focus in test time compute we might see renewed interest in online and meta learning optimization. time to undust those OGD and MAML papers

submitted 101 days ago • 1 comment

100 followers 🎉🥳 Won’t let you guys down, math content is coming! (just need 6000 more to catch up)

submitted 101 days ago • 0 comments