github.com/aardappel/go... It’s like they took every pattern you try to stick to for performant code and just made it a first class language feature. Very cool language and idea.
Seth Stadick
Seth Stadick
@ducktapeprogrammer.bsky.social
Bioinformatics. Rust. Mojo. Also on https://dev-null.sh/@sstadick
817 followers489 following353 posts
dev-null let's you create posts via your ssh keys: dev-null.sh/settings/ter... Behold, an automated posting of the nightly diff for Mojo! dev-null.sh/@sstadick/co...
Anyone else on /dev/null yet? It seems promising, but I don't know anyone there yet: dev-null.sh/@sstadick
Mojo gets support for its equivalent of ‘if let’ and ‘while let’: github.com/modular/modu...
Full nightly changlog, there’s cool stuff in progress: mojolang.org/releases/nig...
GPU kernels in Rust! developer.nvidia.com/blog/introdu...
enums and match proposals for Mojo: github.com/modular/modu...
ls -transubstantiationalist That works. tmpout.sh/5/2.html
One of the most balanced articles I've read on how LLM usage and learning interact: larsfaye.com/articles/ai-... Well worth the 10 minutes to read it.
A good write up on how the new fast floating point arithmetic in rust v1.98.0 can be used pythonspeed.com/articles/fas... releases.rs/docs/1.98.0/ (the `algebraic_*` methods)
rust-lapper v2.0.0-beta.1 is published! This is a large update to mitigate the worst case scenario where there are large intervals that envelope many smaller intervals. It's a lot faster (more details below), and keeps the same api, as well as the BITS implementation. crates.io/crates/rust-...
Perf: This should generally be 1.5x faster than the original rust-lapper, and 76-90x faster on the worst case scenarios. Huge thank you to the bio-polars team that put out this blog post which inspired the second look at rust-lapper biodatageeks.org/polars-bio/b...
rust-lapper v2.0.0-beta.1 is published! This is a large update to mitigate the worst case scenario where there are large intervals that envelope many smaller intervals. It's a lot faster (more details below), and keeps the same api, as well as the BITS implementation. crates.io/crates/rust-...
The main change is adding 32-interval block index that allows for skipping, as well as doing simd comparisons when there are some matches present, while maintaining forward ordered iteration on query. See the PR for details: github.com/sstadick/rus...
github.com/microsoft/sn... There is a rust crate as well. tl;dr; Allocator that is great when allocations on one thread are freed by a different thread
Now that it's open source, I can plug a few helper libraries I've been pulling along to 1.0! But first, the best spot to start is mojolang.org For looking for libraries: github.com/modular/modu...
Mojo is now fully open source! www.modular.com/blog/mojo-op...
mojopt - A Mojo library for parsing CLI args based on the Rust Structopt crate. github.com/sstadick/moj...
Lastly, I'm a contributor to the `pixi-build-mojo` backend, which is what @prefix.dev can use to build Mojo projects: pixi.prefix.dev/latest/build... And allows you to install dependencies directly from GitHub, which is really cool!
Now that it's open source, I can plug a few helper libraries I've been pulling along to 1.0! But first, the best spot to start is mojolang.org For looking for libraries: github.com/modular/modu...
Mojo is now fully open source! www.modular.com/blog/mojo-op...
hat - a thin wrapper for pixi for a more cargo like interface to working with mojo projects github.com/sstadick/hat
extramojo - Things I missed in the Mojo stdlib, such as Rust-like Reader/Writer, BitVec, bstr, etc github.com/ExtraMojo/Ex... Docs site here: extramojo.github.io/ExtraMojo/ex...
Mojo is now fully open source! www.modular.com/blog/mojo-op...
Mojo 1.0 is out! www.modular.com/blog/modular... Compiler open sourcing still looks on track for sometime in the next few months.
Notably, phase 2 should start soon, which has the two things I'm most excited for: concurrency and pattern matching. mojolang.org/docs/roadmap...
Seth Stadick reposted
New preprint🚨(*long* time coming)! We introduce seqproc: an efficient, flexible, and concise tool for describing and transforming sequencing-read geometry. The aim: make complex protocols easy to specify & transform without giving up speed or accuracy. www.biorxiv.org/content/10.6... 🧵
I have some large changes queued up for rust-lapper to address the worst-case in these benchmarks: biodatageeks.org/polars-bio/b... BUT, it's now a fundamentally different data structure. The API is the same for using it. Do I release as a v2 of rust-lapper, or release as a brand new crate?
The speedups roughly 1.5x faster on the general cases and ~76x faster on the worst case in that benchmark post. Lapper as an idea originates from @brent-p.bsky.social, and a large part of its uniqueness is how simple it is.
"Once this is implemented then, in a twist of fate, Zig will become one of the only toolchains capable of producing actually memory safe executables" At 1-6x performance penalty. BUT, still very cool and interesting way to do this. codeberg.org/ziglang/zig/...
To be clear: I think this is cool because it means more money going into biotech related software. The same way companies are suddenly ready to pay to have good data storage practices and slap APIs on everything, they may now pay attention to their lab infrastructure.
> AI changes this equation in the same way machine tools changed manufacturing or cloud computing changed software deployment: by collapsing the marginal cost of experimentation. mattshlosberg.substack.com/p/biotech-is...
purplesyringa.moe/blog/quadrup... I have not seen likely/unlikely used in order to allow the cpu to parallelize loop instructions before. Very cool.
> AI changes this equation in the same way machine tools changed manufacturing or cloud computing changed software deployment: by collapsing the marginal cost of experimentation. mattshlosberg.substack.com/p/biotech-is...
This is pretty much spot-on and it make all these lab-in-the-loop companies look prescient. My experience from doing this sort of work is that it's the boring sample chain-of-custody and metadata tracking stuff that will widen/narrow the bottleneck
Seth Stadick reposted
My secret 4th ALGO paper is out! We show a tight space lower on non-minimal k-perfect hash functions, generalize PtrHash into a non-minimal k-PHF, and then use it to develop a hash set implementation that is up to 1.6x faster than other hash sets! With Stefan {Hermann, Walzer} and Peter Sanders
Ragnar Groot Koerkamp, Stefan Hermann, Peter Sanders, Stefan Walzer: Non-minimal k-perfect hashing: Tight lower bounds and an application to fast static hash tables https://arxiv.org/abs/2607.07257 https://arxiv.org/pdf/2607.07257 https://arxiv.org/html/2607.07257
There are going to be WILD internal code bases in 5 years, given how bad some of the “one person built their own world over 10 years” already is inside companies.
This is the Curse of LISP in action that I've been worried about happening too with LLMs by the way. It's so easy for one person to build a whole project, that they all do, but there's never any need to work together, so nobody does, nothing gets mindshare and eventually everything dies on the vine.