Cosmopolitan Third Edition(justine.lol)
1100 points by jart | 220 comments
Justine Tunney's announcement of Cosmopolitan libc version 3.0.1, featuring a new linker for building fat binaries that run on multiple OSes and architectures.
First screen, captured .
Looks similar
Weird Lexical Syntax(justine.lol)
407 points by jart | 214 comments
Justine Tunney's blog post surveying surprising lexical syntax quirks across 42 programming languages learned while building a syntax highlighter for llamafile.
LLaMA now goes faster on CPUs(justine.lol)
1286 points by lawrencechen | 413 comments
Justine Tunney's blog post describing 84 new matrix multiplication kernels that make llamafile prompt evaluation 30-500% faster on CPUs than llama.cpp.
Modern Perfect Hashing(blog.sesse.net)
103 points by bariumbitmap | 17 comments
Bash one-liners for LLMs(justine.lol)
265 points by severine | 54 comments
Restartable Sequences(justine.lol)
236 points by grappler | 59 comments
Related stories
Benchmark in Milliseconds(matklad.github.io)
143 points by surprisetalk | 47 comments
A blog post by matklad arguing that micro benchmarks should be sized to run around 300 milliseconds, explaining the reasons for this rule of thumb.
Zlib-rs is faster than C(trifectatech.org)
329 points by dochtman | 437 comments
Trifecta Tech Foundation blog post announcing zlib-rs 0.4.2 and benchmarking its decompression and compression speed against C implementations zlib-ng and zlib-chromium.
Htmx Is the Future(quii.dev)
741 points by quii | 553 comments
Chris James' blog post arguing that HTMX offers a simpler alternative to SPA frameworks like React and Angular for building smooth web applications.
I built a programming language that LLMs can't use(blog.picheta.me)
244 points by dom96 | 179 comments
Blog post by Dominik Picheta introducing KillSwitch, a programming language he designed so that large language models cannot use it.
Beating NumPy matrix multiplication in 150 lines of C(salykova.github.io)
372 points by p1esk | 69 comments
Blog tutorial implementing high-performance CPU matrix multiplication in about 150 lines of C, following the BLIS design and outperforming NumPy/OpenBLAS on a Ryzen 7700.
I wasted weeks hand optimizing assembly because I benchmarked on random data(vidarholen.net)
376 points by thunderbong | 148 comments
Blog post recounting how the author spent weeks hand-optimizing assembly for a Java platform, only to find the gains came from benchmarking on unrealistic random data.