The future is software writing its own software. Which is why I'm so in love with Pi: a coding agent that can extend itself :) lucumr.pocoo.org/2026/1/31/pi/

"However, we wish to stay close to programming practice and define patterns as follows" moscova.inria.fr/~maranget/pa...

my 2 days of GPT-5.2 xhigh so far

stream underscores

LLMs automate typing, not thinking. Mikayla Maki shares her mental model for working with agents, and how she puts it into practice: zed.dev/blog/on-prog...

I get the feeling that <system-reminder> is what makes Claude Code feel different (and sometimes awful) from other coding agents I've tried

Testing out progressive disclosure patterns for compiler testing tools with @ampcode.bsky.social

purescript-analyzer's playground now supports loading PureScript packages from the package set, check it out: purefunctor.github.io/purescript-a...

Weekend thoughts on Gas Town, Beads, slop AI browsers, and AI-generated PRs flooding overwhelmed maintainers. I don't think we're ready for our new powers we're wielding. lucumr.pocoo.org/2026/1/18/ag...

It just remembered a warning it should give you with AI driven coding, it is mentally taxing in ways you will not notice. You must madate breaks even if you dont feel like you need them.

what do we have here 👀

I'm still bullishly against AI image/video generation even if i'm neutral on LLMs now. genuinely what is the point

wrote some words about Opus 4.5

The Stanford University class CS99: Functional Programming and Theorem Proving in Lean 4, taught by Leni Aniva (@aniva.bsky.social), features Cirno and Rumia as teaching tooltips throughout the slides in the course about the proof assistant Lean. web.stanford.edu/class/cs99/

Putting Signals in your compiler? Turns out this is already happening. marvinh.dev/blog/signals...

A while back, irked by the sheer misrepresentation of the performance of recursion, I wrote a soft introduction to continuation-passing style and tail-call optimisation. Never really shared it anywhere but might as well:

🔥🔥🔥 music.apple.com/ph/album/tid...

i think it's mostly not actually about syntax, it's just that the people saying that don't know how to express their actual concerns. e.g. with rust it's usually not about the syntax as much as its about the complexity of the language even with things that are much simpler in other languages

Every so often I see comments about "weird" syntax (typically towards Rust/Haskell/etc.) being a reason for people to hate a language and that just feels like a massive skill issue

The PureScript Analyzer now uses verifying step traces for its build system, the same tracing mechanism used by Shake! As a result, caching for downstream phases in the compiler is much more aggressive and should reduce redundant computations from insignificant edits github.com/purefunctor/...

I love when abstractions fall right into place

Now I gotta figure out how to make the build system parallel

TIL that Option<NonZero<u32>> has the same size as NonZero<u32> and I'm wondering if it's a worthwhile optimisation to make for flattened ASTs? At 1 billion nodes yeah it'll save 12 GB but at a million 12 MB doesn't matter as much for modern hardware.

The purescript-analyzer currently hovers around 1.0~1.3 GiB of memory usage for 871k (!!!) lines of PureScript. A good chunk of the memory being allocated comes from the AST, specifically the mechanism it currently uses for flattening the representation.

Metroid Prime Remastered is one of the prettiest games I've played on the Switch. The 60 fps is honestly pretty jarring for me given most games I've played on the console and it's amazing how much fidelity they squeezed out of it

I wonder how useful it'd be to allow source files to annotate how much memory should be pre-allocated for certain algorithms to avoid repeated allocations as structures grow

The Banko Sentral ng Pilipinas knows that SMS is an insecure mode of communication, so why not force banks to get rid of SMS OTPs? business.inquirer.net/497198/bsp-w...

been a hot minute since I cooked this good

Switch 2 single-handedly deleted whatever need I have left of a gaming PC

Type holes are infinitely more useful than LLMs

UNDERTALE is on sale for $1. If you're a bluesky user who hasn't played this game - now is the best time to join the other 99% of people on this website and pick it up. store.steampowered.com/app/391540/U...

they're writing TypeScript in what now devblogs.microsoft.com/typescript/t...

erm

I think it'd be cool if the purescript analyzer could reduce memory allocations in non-interactive mode by turning off certain information from being gathered during compilation

the album mix of Mimic is so darn good youtu.be/Mygq0AjrWLQ

Instance heads are annoying to lower

type checker that auto-plays youtu.be/L8XbI9aJOXk on trivial unification errors

Zen has a learning curve (at least coming from base Firefox) but it definitely delivers on the base promise. I also finally have a reason to use containers more seriously with its workspaces.

The only thing that's stopping me from switching back to Emacs is that flash.nvim has a better implementation than avy

How bad could I shoot myself in the foot by using de Bruijn indices for higher-ranked polymorphism

What do you call the phenomenon of being bad at a bad language because of good languages

I quit using LLMs to write code that I would have written in the first place purefunctor.me/blog/i-quit-...

Note to self: make sure the binary you're benchmarking bails if it doesn't find the files it's supposed to be using

I think the honeymoon phase with Neovim is over and now I find myself wanting to switch back to... Emacs

I was just thinking about how skolemisation would fit in a constraint-generating type checker and I was wondering if it makes sense to use implication constraints for them.