Pinned
2025 year in review is up 🪩 cog.spin.systems/2025-in-review On dewarping, schema inference, and embeddings
Louis Maddox
@permutans.bsky.social
Combinatorially curious https://spin.systems
248 followers135 following4.4K posts
Pinned
2025 year in review is up 🪩 cog.spin.systems/2025-in-review On dewarping, schema inference, and embeddings
mibz2: a high-performance bz2 decoding library built around memory-level parallelism lmmx.github.io/mibz2/ Speeds up bzip2 decoding by interleaving independent inverse-BWT traversals, keeping multiple memory accesses in flight, plus SIMD and block-level parallelism boosters 💫
Counterintuitive approach to perf hunting: look for where the CPU is waiting, not (just) where it is working Claude came up with a trick to create a bz2 decoder 1.76× as fast as lbzip2 via memory-level parallelism (had identified "memory-latency bound" & then rolled its own PMU measurements)
If you ask an agent to explain their goals during a session it gives an account pretty fit for docs, akin to the program’s “ideal user” story It produces accounts structured in chained “what for?” Q&A, to “establish the final reason guiding choice of means” up to “one final why”
On agent /goal setting with Anscombe’s builder (§45) Claude notes an agent can be a foreman as much as a worker 📜 Anscombean SKILL to elicit an intention (speculating that this can effectively structure a user’s expression of their intention) github.com/lmmx/anscomb...
For fun: Claude (O. 5.5) précis of Anscombe's Intention in 'progressive disclosure' style github.com/lmmx/anscomb...
Feels vulgar to state that Anscombe has pretty direct lines of application to agents
On agent /goal setting with Anscombe’s builder (§45) Claude notes an agent can be a foreman as much as a worker 📜 Anscombean SKILL to elicit an intention (speculating that this can effectively structure a user’s expression of their intention) github.com/lmmx/anscomb...
Similar approach to 📄 Kärkkäinen et al. (2012) Slashing the Time for BWT Inversion www.cs.helsinki.fi/u/tpkarkka/p... — this paper also describes a way to halve the number of cache misses during BWT inversion [which the memory-level parallelism amortises by 2x] giving a 4x overall speedup
Louis Maddox reposted
Replying to @lu.is
TIL: someone has ported English to Rust github.com/Aaryan-Kapoo...
Louis Maddox reposted
Louis Maddox reposted
Skimming this new PDF, I feel some similarities to the sub-n*lg(n) FFT writeup by David Harvey. It's all like breaking the n^3 matmul bound: instead of doing many independent mat-vec-mul units of work, do them all at once, exploiting structure to save a few operations. bsky.app/profile/curi...
Here's a direct link to the PDF claiming a faster 3SAT algorithm: github.com/openai/math/...
Louis Maddox reposted
Benchmark tracker 📉 lmmx.github.io/wikidata-pq/... ⏱️ Pipeline wall time 15s → 3s and CPU 41s → 6s 🧠 Peak RSS per worker 1.8 GB → 242 MB 🚀 Expected release ETA 60h → 7.5h Big wins: building tables mid-download (−16% ETA), overlapping stages (−8%) & merging col chunks (−4%)
Wikidata GH runner ingest speedrun any % lfg
Louis Maddox reposted
Replying to @hugovk.dev
🐍🚀 Out now: security releases of #Python 3.11–3.15 🔒 They block the compromised #eMudhra intermediate CA keys in the ssl module as a last resort. Normal pip use, including upgrading pip itself, is now safe from attacks using those keys, even before pip is updated discuss.python.org/t/python-3-1...
Erol Alkan on bloghouse: "Many people don’t realise that around that time, we were all on iChat" www.loudandquiet.com/interview/wi...
✍️ nihilunbounded, Who Watches the Watchmen? (2026) nihilunbounded.substack.com/p/who-watche... On Fields medallist Jacob Tsimerman (OpenAI)’s “appeal to engineering epistemology in mathematics” in this interview (“exclusive first podcast” 7th August 2026: curtjaimungal.substack.com/p/he-won-mat...)
Louis Maddox reposted
is this a secure private sigil?
I made base-1353 encoding, so I can fit 256-bit values, like “96cca2a0c8a3263c74ef36274abf55f410d992034d16e8285811a9ad3fc59c13”, into short 26 character strings, like “ᓒᔳᣏዧᐎᑙᖜዌᙔᐸƫȶƃᒨዊጶᙵኻѫyፐᎉᑫᖫᘪᒩ”. (Also, I think these sigils are beautiful.) johnservil.github.io/sigil/play.h...