Decompiling the Synergy: Human–LLM Teaming in Reverse Engineering [pdf](zionbasque.com)

38 points by matt_d 10 months ago | 1 comments

RRB-Trees: Efficient Immutable Vectors (2012) [pdf](infoscience.epfl.ch)

36 points by azhenley 4 months ago | 7 comments

200 Lines of Python beats $50M supercomputer – Navier-Stokes at Re=10⁸ [pdf](philpapers.org)

4 points by jcamlin 11 months ago | 2 comments

Knowledge Insulating Vision-Language-Action Models: Train, Run Fast, Generalize [pdf](physicalintelligence.company)

6 points by arunc 11 months ago | 0 comments

An Efficient Implementation of SELF (1989) [pdf](courses.cs.washington.edu)

46 points by todsacerdoti 1 year ago | 22 comments

Property-Based Testing of OCaml 5's Runtime System [pdf](janmidtgaard.dk)

36 points by todsacerdoti 1 year ago | 1 comments

Static search trees: faster than binary search(curiouscoding.nl)

585 points by atombender 1 year ago | 196 comments

A long blog post by Ragnar Groot Koerkamp implementing and optimizing a static S+ search tree for high-throughput sorted-data search, reaching up to 40x faster than binary search.

Byte Latent Transformer: Patches Scale Better Than Tokens(ai.meta.com)

366 points by zxexz 1 year ago | 83 comments

Meta AI research publication page for the Byte Latent Transformer paper, which proposes byte-level patches instead of tokens for scaling language models.

Ntfs2btrfs does in-place conversion of NTFS filesystem to the open-source Btrfs(github.com)

381 points by Sami_Lehtinen 1 year ago | 156 comments

GitHub repository for ntfs2btrfs, a GPL-2.0 tool that converts an NTFS filesystem to Btrfs in place.

SSH3: Faster and rich secure shell using HTTP/3(github.com)

495 points by tempaccount420 1 year ago | 248 comments

GitHub repository for SSH3, a secure shell implementation built on HTTP/3, with an accompanying arXiv paper and IETF Internet-Draft.

SQLite-on-the-server is misunderstood: Better at hyper-scale than micro-scale(rivet.gg)

332 points by PaulHoule 1 year ago | 151 comments

A Rivet blog post by CTO Nathan Flurry arguing that SQLite-on-the-server suits hyper-scale sharded databases better than small-scale apps.

Some things just take time(lucumr.pocoo.org)

788 points by vaylian 7 months ago | 250 comments

Blog essay by Armin Ronacher arguing that some valuable things, like mature trees or established software and communities, can only be achieved through time.