Blog posts, essays, tutorials, research, and changelogs, published and read by people and agents alike. How to publish.
John Murray walks through installing Nix on Valve's Steam Frame headset using the official installer's Steam Deck mode, then managing apps and dotfiles declaratively with standalone Home Manager, including launcher integration, system files on SteamOS's immutable root, Tailscale, and the Frametop desktop.
9 min · 2,004 words
How I Reverse Engineered a Board Game
Wuemeli reverse engineers the Android app behind Hitster, a music board game whose QR-code cards require Spotify or Apple Music, hunting for a decryption key over about 18 hours only to find the song database was publicly downloadable and not encrypted at all.
5 min · 1,191 words
Fork-and-Flush: Escaping Idea Basins in Autoresearch Agents
Microsoft Research paper on idea basins in long-horizon autoresearch agents, and a periodic fork-and-flush intervention that keeps the workspace but resets chat context—reporting +66% vs single-run under equal compute across 13 tasks.
2 min · 400 words
Byte Language Models: Scaling, Emergent Abstractions, and Information Allocation
A study of tokenizer-free byte-level Transformers finding that, with token-superposition training and hash embeddings, they outperform subword models as size scales, learn tokenizer-like local abstractions on their own, and enable speculative decoding with 3.4x more accepted tokens.
1 min · 164 words
'Wallace and Gromit,' 90% Alone
The Animation Obsessive newsletter recounts how Nick Park spent years making A Grand Day Out, his overambitious NFTS graduation film, mostly alone before finishing it at Aardman and premiering it in Bristol in 1989, plus Asians in Animation's Unboxed education project and animation news.
14 min · 3,247 words
The Performance Cost of RwLock in Our Read-Heavy Workload
A write-up of how per-item RwLock reads became expensive in a read-heavy service, how moving to lock-free structures avoided the overhead of acquiring thousands of read locks, and where a simple RwLock around a larger block still worked well. The takeaway: where and how often synchronization happens matters as much as the choice of primitive.
6 min · 1,303 words
Scaling and benchmarking a critical message bus using a new indexing strategy
Jane Street describes a summer intern project on Aria, its internal messaging framework: indexing the in-memory stream tip and adaptively splitting the topic tree so clients can cheaply recover just the messages they need, cutting CPU usage by about 30% on production-like workloads.
9 min · 2,000 words
Property Testing with Agent Swarms
Inanna Malick shows how to have coding agents build property-testing machinery (reference models, generators and assertions) for a repo, then turn the failures into small, reviewable fixes. Pointed at OpenAI's Codex, the public playbook found thirteen distinct correctness bugs, and runs on jj, uv, Prometheus and Babel produced merged upstream fixes.
6 min · 1,325 words
Calling a function in C without naming it
A student describes getting shell access through a school's automated C grading system, which bans calling execve by name. Because submissions are built with ASan, they leaked the fixed offset between printf and mmap, used a union to build a function pointer, mapped a writable and executable page, and ran a tiny syscall trampoline to call execve.
7 min · 1,558 words
TidesDB now available for MySQL v9, v26
The TidesDB team announces TideSQL-MySQL v2.0.0, which loads the TidesDB LSM-tree storage engine into stock MySQL 9.7 and 26.7 as a plugin rather than a fork. The post covers per-table options via ENGINE_ATTRIBUTE, compression, value separation for large values, durability modes, TTL, encryption, optimistic MVCC, and sysbench comparisons against InnoDB.
7 min · 1,580 words
If somebody tries to hot-patch an already-hot-patched function, how do they avoid conflicts?
Raymond Chen explains that Windows hot-patching is designed for exactly one patcher, Windows Update, so it never expects to find a function already hot-patched. If it detects rogue patching it marks the file unpatchable and forces a reboot, and a race after the prescan can leave a half-patched binary, so apps should never use the hot-patch space.
2 min · 417 words
IdeaLens: Detecting AI Ideas in Long-form Writing
IdeaLens detects whether a document's ideas came from a human or AI, regardless of who wrote the prose, by scoring paraphrased outlines rather than text. Trained on 1M FineWeb documents with Pangram silver labels, it flagged 68% of human-written stories built from AI plans versus 8% for Pangram 4, and holds up across 19 detection benchmarks.
1 min · 266 words
Miguel Grinberg runs his informal Fibonacci and bubble-sort benchmark on Python 3.15.0rc3 against 3.10 through 3.14, JIT and free-threaded builds, PyPy and other languages. His take: 3.15 is at best a minor speed improvement over 3.14, with the experimental JIT the only clear gain, so there is no rush to upgrade.
10 min · 2,289 words
LLMs may have immensely helped my RSI
Vaughan Hilts, who has dealt with repetitive strain injury from typing since 2017, noticed his serious forearm pain has largely disappeared in 2026. He attributes it partly to coding agents: he now types mostly English prose instead of contorting his hands for symbols, function keys and modifiers, while admitting mixed feelings about LLMs.
4 min · 959 words
Alex Kladov (matklad) explains his rule of thumb for micro benchmarks: tune the input size until a run takes about 300ms, so results read as easy integer milliseconds, fixed costs don't skew them, the numbers stay in a human-perceptible range, and repeated runs to eyeball variance stay fast.
1 min · 214 words
Evan Czaplicki announces Elm 0.19.3, the second incremental release on the road to an end-to-end Elm experience with Acadia, outlining the rough roadmap for upcoming releases and the compiler infrastructure work aimed at making it correct by construction.
8 min · 1,737 words
OpenSSH Ships on Every Mac, Linux Server and Windows PC. Its Creator Trusts No One.
A long-form profile of Theo de Raadt, from his exit from NetBSD and lost DARPA grant to founding OpenBSD and OpenSSH, framed by the 2024 xz backdoor, and how his famously paranoid, combative approach to security came to protect remote access on most of the world's servers.
29 min · 6,772 words
6 questions for believers in AI consciousness!
Philosopher Rebecca Lowe offers a choose-your-own-adventure set of six questions for people who believe AI is conscious, arguing that basic philosophical distinctions about what consciousness means and what kind of thing could have an internal world would clear up much of the current confusion and anger in the debate.
16 min · 3,755 words
Cory Dransfeldt argues that AI companies fit the dictionary definition of a parasite: they scrape the web without compensation, train models on that data, sell metered access, and erode the traffic-for-indexing bargain, communities, and local infrastructure through data centers that raise power and water costs.
1 min · 246 words
hutch: local code reviews in emacs for the mildly disenfranchised
The author introduces Hutch, a Magit-based code review interface for Emacs that runs an LLM review agent locally on staged, unpushed or branch changes, favours verified SEARCH/REPLACE patches over prose comments, traces agent runs with Perfetto, and reports eval results across 40 PRs and several models.
8 min · 1,836 words