Nelson Elhage
Software engineer who writes the blog Made of Bugs about performance, debugging and understanding computer systems. Previously worked at Anthropic on interpretability, at Stripe on Sorbet, and at Ksplice.
Nelson Elhage did not write this page. ppll compiled it from their own words, each linked to the source. Is this you? · Report a mistake
Latest GitHub
What they believe
-
A composable concurrency paradigm requires a general-purpose cancellation mechanism.
But I struggle to envision a general, composable, concurrent paradigm without one.
blog.nelhage.comSaid 23 Mar 2026This on korrents
-
Augmenting a general-purpose solver with domain-specific heuristics commonly gives it a meaningful performance boost.
My sense is that this hybrid is fairly common in practice; solvers aren't magical and if you can deduce additional structure using domain-specific analysis, it will often give the solver an important boost.
blog.nelhage.comSaid 21 Oct 2025This on korrents
The bold line is korrents’ wording; the quote is theirs. These 2: what they wrote before what they said aloud, one per source, newest first.
Feed
As its own page →Hiding
21 October
19 October
16 October
15 October
- New reposyncflowy
13 October
23 September
26 August
14 July
3 July
26 June
17 June
15 June
3 May
16 April
23 March
From one piece From error-handling to structured concurrency 2 beliefs · blog.nelhage.com
-
Their words
writing programs in this style makes it much easier to write correct and safe concurrent code
From error-handling to structured concurrencyblog.nelhage.com
-
Their words
But I struggle to envision a general, composable, concurrent paradigm without one.
From error-handling to structured concurrencyblog.nelhage.com
11 December 2025
21 October 2025
From one piece Solving regex crosswords with Z3 2 beliefs · blog.nelhage.com
-
Their words
I suspect the pattern generalizes: If Z3 has first-class support for your problem domain, it's worth starting there!
-
korrents.com
Augmenting a general-purpose solver with domain-specific heuristics commonly gives it a meaningful performance boost.Their words
My sense is that this hybrid is fairly common in practice; solvers aren't magical and if you can deduce additional structure using domain-specific analysis, it will often give the solver an important boost.
4 July 2025
From one piece The ITTAGE indirect branch predictor 3 beliefs · blog.nelhage.com
-
korrents.com
Using a neural network is likely to become easier than designing a custom-tuned prediction algorithm for many tasks.Their words
in 2025 it might end up being easier to just throw a neural net at the problem, instead of a custom-designed-and-tuned prediction algorithm!
-
korrents.com
Most conditional branches in software are strongly biased toward one outcome rather than close to fifty-fifty.Their words
most branches are heavily "biased" one way or another - e.g. a "branch taken" rate of 10% or 90% is more common than 50%
-
korrents.com
Modern CPUs can now predict the indirect dispatch jump in a bytecode interpreter loop with high accuracy.Their words
Modern CPUs mostly no longer struggle to predict the bytecode-dispatch indirect jump inside a "conventional" bytecode interpreter loop.
9 March 2025
From one piece Performance of the Python 3.14 tail-call interpreter 3 beliefs · blog.nelhage.com
-
Their words
I'm hopeful this framework will turn out to be a much more robust style for writing performance-sensitive code, especially over time and as compilers evolve.
-
Their words
However, with so many different software projects out there, each moving so rapidly and depending on and being used by so many other projects, it becomes practically-inevitable that some regressions "like that one" happen, almost constantly.
-
Their words
It's easy to get impressive-looking results if you're comparing against a poorly-tuned baseline, and that observation turns out to explain a surprising fraction of supposed improvements.
5 March 2025
1 March 2025
27 January 2025
From one piece Building personal software with Claude 3 beliefs · blog.nelhage.com
-
Their words
Also, current models, while impressive, still fall well short of expert human performance.
-
korrents.com
AI tooling and interface design lags behind the underlying models' actual capabilities.Their words
Overall, though, this experience reinforced my belief that the tooling and interface design around LLMs is lagging way behind the actual capabilities, and is an area of active experimentation and development, even aside from any future model improvements.
-
korrents.com
As AI makes code cheaper to produce, understanding and architectural design become more valuable.Their words
Thus, code is cheaper than ever, but I suspect that insight and good architectural design and understanding, at least for now, will become more valuable than ever.
3 July 2024
-
Their words
I suspect - but haven't verified - that in many corpora, we will encounter fairly bimodal Jaccard values - two clusters, near 1 and 0.
21 May 2024
-
korrents.com
Reliable, stable developer tooling matters more than which specific technical choices a team makesTheir words
The best tooling in the world will struggle to make up for "regularly losing a day debugging your environment," and so getting that aspect right can easily outweigh almost any other decisions you make.
7 May 2024
5 April 2024
4 April 2024
28 March 2024
26 January 2024
7 January 2024
18 December 2023
11 December 2023
8 December 2023
8 November 2023
Nothing matches.
Beliefs
Korrents What they believe 19 beliefs — each backed by an exact quote.
Each is a — compiled by korrents.com, not by them: the one-line wordings are korrents', the quotes are theirs.
Recent
A composable concurrency paradigm requires a general-purpose cancellation mechanism.
But I struggle to envision a general, composable, concurrent paradigm without one.
From error-handling to structured concurrency Said 23 Mar 2026
Structured concurrency makes it easier to write correct, safe concurrent code.
writing programs in this style makes it much easier to write correct and safe concurrent code
From error-handling to structured concurrency Said 23 Mar 2026
Augmenting a general-purpose solver with domain-specific heuristics commonly gives it a meaningful performance boost.
My sense is that this hybrid is fairly common in practice; solvers aren't magical and if you can deduce additional structure using domain-specific analysis, it will often give the solver an important boost.
Solving regex crosswords with Z3 Said 21 Oct 2025
Show 16 more
When a tool has first-class, native support for a problem domain, that is the best place to start before building a custom encoding.
I suspect the pattern generalizes: If Z3 has first-class support for your problem domain, it's worth starting there!
Solving regex crosswords with Z3 Said 21 Oct 2025
Modern CPUs can now predict the indirect dispatch jump in a bytecode interpreter loop with high accuracy.
Modern CPUs mostly no longer struggle to predict the bytecode-dispatch indirect jump inside a "conventional" bytecode interpreter loop.
The ITTAGE indirect branch predictor Said 4 Jul 2025
Most conditional branches in software are strongly biased toward one outcome rather than close to fifty-fifty.
most branches are heavily "biased" one way or another - e.g. a "branch taken" rate of 10% or 90% is more common than 50%
The ITTAGE indirect branch predictor Said 4 Jul 2025
Using a neural network is likely to become easier than designing a custom-tuned prediction algorithm for many tasks.
in 2025 it might end up being easier to just throw a neural net at the problem, instead of a custom-designed-and-tuned prediction algorithm!
The ITTAGE indirect branch predictor Said 4 Jul 2025
Impressive-looking benchmark improvements often result from comparing against a poorly-tuned baseline rather than a genuine advance.
It's easy to get impressive-looking results if you're comparing against a poorly-tuned baseline, and that observation turns out to explain a surprising fraction of supposed improvements.
Performance of the Python 3.14 tail-call interpreter Said 9 Mar 2025
As software projects grow more numerous, fast-moving, and interdependent, cross-project regressions become practically inevitable.
However, with so many different software projects out there, each moving so rapidly and depending on and being used by so many other projects, it becomes practically-inevitable that some regressions "like that one" happen, almost constantly.
Performance of the Python 3.14 tail-call interpreter Said 9 Mar 2025
Compiler mechanisms that force a required optimization to occur, rather than relying on incidental compiler behavior, are a more robust way to write performance-sensitive code.
I'm hopeful this framework will turn out to be a much more robust style for writing performance-sensitive code, especially over time and as compilers evolve.
Performance of the Python 3.14 tail-call interpreter Said 9 Mar 2025
As AI makes code cheaper to produce, understanding and architectural design become more valuable.
Thus, code is cheaper than ever, but I suspect that insight and good architectural design and understanding, at least for now, will become more valuable than ever.
Building personal software with Claude Said 27 Jan 2025
AI tooling and interface design lags behind the underlying models' actual capabilities.
Overall, though, this experience reinforced my belief that the tooling and interface design around LLMs is lagging way behind the actual capabilities, and is an area of active experimentation and development, even aside from any future model improvements.
Building personal software with Claude Said 27 Jan 2025
Current AI models still fall short of expert human performance.
Also, current models, while impressive, still fall well short of expert human performance.
Building personal software with Claude Said 27 Jan 2025
Document similarity scores across a corpus tend to cluster bimodally near identical and near unrelated, rather than spreading evenly.
I suspect - but haven't verified - that in many corpora, we will encounter fairly bimodal Jaccard values - two clusters, near 1 and 0.
Finding near-duplicates with Jaccard similarity and MinHash Said 3 Jul 2024
Reliable, stable developer tooling matters more than which specific technical choices a team makes
The best tooling in the world will struggle to make up for "regularly losing a day debugging your environment," and so getting that aspect right can easily outweigh almost any other decisions you make.
Stripe's monorepo developer environment Said 21 May 2024
Rapidly growing organizations need a developer environment that works well out of the box for new engineers
In a rapidly growing organization, it's much more important to have a developer environment that "just works" out of the gate, with minimal setup pain for new engineers.
Stripe's monorepo developer environment Said 21 May 2024
Per-engineer productivity inevitably declines somewhat as an organization and its codebase grow
It is almost inevitable that per-engineer productivity drops to some extent as an organization and codebase grows, even though it's also nearly-impossible to quantify that effect.
Stripe's monorepo developer environment Said 21 May 2024
Profilers show where time is spent but not which of that time can feasibly be reduced or is worth optimizing.
Profilers can only show you "what is there"; skillful performance engineering requires understanding the information that shapes and informs the profile but is "not there" in the profile itself.
Performance engineering, profilers, and seeing the invisible Said 18 Dec 2023
A profiler measures time already spent in an operation, not how much time would be saved by optimizing it.
a profile tells you how much time a program spent in a given operation; but not how much time it would spend if you optimized that operation.
Performance engineering, profilers, and seeing the invisible Said 18 Dec 2023
What is a korrent?
A korrent is a belief a person has stated in their own words: one sentence stating the claim, backed by a quote and a source, kept at korrents.com.
Under a name here, the quoted block is what they actually said. The korrent beneath it is the claim those words support, in korrents' wording — tap it to see the record, its source, and who else holds it.
Nobody here wrote their own korrents. They are compiled from public statements, and a person can change their mind, which is recorded too.
Recommends rcmnd
rcmnd What they recommend 13 things — each backed by an exact quote.
Compiled by rcmnd.app, not by them: the quotes are theirs, and each links to the post it was copied from.
Recent
Mixed on Digital Apollo
I honestly don't recall that much from this book. I read it fairly early on in my Apollo exploration. I'd recommend Eyles' Sunburst and Luminary, above, as a better starting point if you're interested in the computers that flew Apollo.
blog.nelhage.com 8 Apr 2019
Mixed on First Man
I enjoyed it, and it got at more of Armstrong's personal story than I have seen elsewhere, but it was not especially memorable for me.
blog.nelhage.com 8 Apr 2019
Liked The Last Man on the Moon
This documentary was a solid telling of the story of that mission, and of the later-in-Apollo scientific goals and geological exploration that was achieved, as well as a wistful look back at the story of the last humans, to date, to set foot on the moon.
blog.nelhage.com 8 Apr 2019
Show 10 more
Recommends Apollo 8: The Thrilling Story of the First Mission to the Moon
This book tells the story of that decision, and of the work and the people who made that mission happen.
blog.nelhage.com 8 Apr 2019
What others say about Apollo 8: The Thrilling Story of the First Mission to the Moon →
Recommends Hidden Figures
The book and movie only tangentially touch on Apollo itself (although it does appear!), but they make the list because they are well worth the read / watch, and they tell an important story of yet another facet of the immense collaboration that was the space program, and the incredible diversity of people and backgrounds and stories and kinds of work that went into making it all work.
blog.nelhage.com 8 Apr 2019
Mixed on Apollo 11
It's well worth a watch, to see what it looked like at the time, but I also found it almost anticlimactic to watch: With the benefit of history there's no real uncertainty or tension about what will happen, and the format gives the viewer disappointingly little "behind-the-scenes" insight into just what it took to make the mission go off
blog.nelhage.com 8 Apr 2019
Recommends Mission Control: The Unsung Heroes of Apollo
I don't actually remember this one in much detail. However, as I mentioned above, Mission Control was the beating heart of Apollo and one of the most fascinating and critical components of the program, so I recommend it anyways.
blog.nelhage.com 8 Apr 2019
What others say about Mission Control: The Unsung Heroes of Apollo →
Recommends Thirteen: The Apollo Flight That Failed
This book covers the mission and the details of the recovery efforts and the incredible work by Mission Control and by the astronauts to get the crew home safely after a catastrophic failure no one had even begun to plan for.
blog.nelhage.com 8 Apr 2019
What others say about Thirteen: The Apollo Flight That Failed →
Loved How Apollo flew to the Moon
This is far and away the best general-audience technical documentation of the Apollo missions I've read.
blog.nelhage.com 8 Apr 2019
Loved Sunburst and Luminary
The only reason it isn't higher on this list is that the content is a bit specialized and narrowly focused on the landing and the software and adjacent topics; it's definitely one of my favorite books here.
blog.nelhage.com 8 Apr 2019
Recommends From the Earth to the Moon
This is a 12-episode HBO miniseries largely based on the previous book. It's also an excellent work, and I'd highly recommend it.
blog.nelhage.com 8 Apr 2019
Recommends A Man on the Moon: The Voyages of the Apollo Astronauts
It's a solid, well-sourced ("based on hundreds of hours of in-depth interviews with each of the twenty-four moon voyagers, as well as those who contributed their brain power, training and teamwork on Earth"), general overview of the heart of the Apollo program. Also an excellent entry point.
blog.nelhage.com 8 Apr 2019
What others say about A Man on the Moon: The Voyages of the Apollo Astronauts →
Recommends Failure Is Not An Option
This was the first book I read about the early space program, and I'd recommend it as a good entry point.
blog.nelhage.com 8 Apr 2019
Where they publish
Site nelhage.com His homepage: work history, projects and an index of where he writes.
Link verified 5 Sept 2026.
Blog Made of Bugs His blog on debugging, performance and understanding systems.
Recent
- From error-handling to structured concurrency 23 Mar 2026 How should we think about error-handling in concurrent programs? In single-threaded programs, we’ve mostly converged on a standard pattern, with a diverse zoo of implementations and concrete patterns. When an error occu…
- Solving regex crosswords with Z3 21 Oct 2025 For a while now, I’ve been fascinated by Z3 and by SMT solving more broadly. While on pat leave recently, I was reminded of the existence of regular-expression crossword puzzles, and allowed myself to get nerdsnip…
- The ITTAGE indirect branch predictor 4 Jul 2025 While investigating the performance of the new Python 3.14 tail-calling interpreter, I learned (via this very informative comment from Sam Gross) new (to me) piece of performance trivia: Modern CPUs mostly no longer str…
Show 17 more
- Performance of the Python 3.14 tail-call interpreter 9 Mar 2025 About a month ago, the CPython project merged a new implementation strategy for their bytecode interpreter. The initial headline results were very impressive, showing a 10-15% performance improvement on average across a…
- Building personal software with Claude 27 Jan 2025 Earlier this month, I used Claude to port (parts of) an Emacs package into Rust, shrinking the execution time by a factor of 1000 or more (in one concrete case: from 90s to about 15ms). This is a variety of yak-shave th…
- Finding near-duplicates with Jaccard similarity and MinHash 3 Jul 2024 Suppose we have a large collection of documents, and we wish you identify which documents are approximately the same as each other. For instance, we may have crawled the web over some period of time, and expect to have…
- Stripe's monorepo developer environment 21 May 2024 I worked at Stripe for about seven years, from 2012 to 2019. Over that time, I used and contributed to many generations of Stripe’s developer environment – the tools that engineers used daily to write and te…
- Performance engineering, profilers, and seeing the invisible 18 Dec 2023 I was recently introduced to the paper “Seeing the Invisible: Perceptual-Cognitive Aspects of Expertise” by Gary Klein and Robert Hoffman. It’s excellent and I recommend you read it when you have a chance. K…
- Advent of Code in C++ Template Metaprogramming 8 Dec 2023 This December, the imp of the perverse struck me, and I decided to see how many days of Advent of Code I could do purely in compile-time C++ metaprogramming. As of this writing, I’ve done two days, and I’m not sure I’ll…
- What's with ML software and pickles? 8 Nov 2023 I have spent many years as an software engineer who was a total outsider to machine-learning, but with some curiosity and occasional peripheral interactions with it. During this time, a recurring theme for me was horror…
- Graceful behavior at capacity 7 Aug 2023 Suppose we’ve got a service. We’ll gloss over the details for now, but let’s stipulate that it accepts requests from the outside world, and takes some action in response. Maybe those requests are HTTP requests, or…
- Efficiency trades off against resiliency 15 Apr 2023 What’s the “right” level of CPU utilization for a server? If you look at a monitoring dashboard from a well-designed and well-run service, what CPU utilization should we hope to see, averaged over a da…
- Transformers for software engineers 1 Apr 2022 Ever since its introduction in the 2017 paper, Attention is All You Need, the Transformer model architecture has taken the deep-learning world by storm. Initially introduced for machine translation, it has become the to…
- A Cursed Bug 23 Feb 2022 In my day job at Anthropic, we run relatively large distributed systems to train large language models. One of the joys of using a lot of computing resources, especially on somewhat niche software stacks, is that you sp…
- Distributed cloud builds for everyone 31 May 2021 CPU cycles are cheaper than they have ever been, and cloud computing has never been more ubiquitous. All the major cloud providers offer generous free tiers, and services like GitHub Actions offer free compute resources…
- Building LLVM in 90 seconds using Amazon Lambda 21 May 2021 Last week, Frederic Cambus wrote about building LLVM quickly on some very large machines, culminating in a 2m37s build on a 160-core ARM machine. I don’t have a giant ARM behemoth, but I have been working on a too…
- Some opinionated thoughts on SQL databases 30 Mar 2021 People who work with me tend to realize that I have Opinions about databases, and SQL databases in particular. Last week, I wrote about a Postgres debugging story and tweeted about AWS’ policy ban on internal use of SQL…
- Towards solving Ultimate Tic Tac Toe 15 Jul 2020 Summary: Read about my efforts to solve the game of Ultimate Tic Tac Toe. It’s been a fun journey into interesting algorithms and high-performance parallel programming in Rust. Backstory Starting around the beginn…
- Write testable code by writing generic code 12 Mar 2020 Alex Gaynor recently asked this question in an IRC channel I hang out in (a channel which contains several software engineers nearly as obsessed with software testing as I am): uhh, so I’m writing some code to han…
- Test suites as classifiers 1 Mar 2020 Suppose we have some codebase we’re considering applying some patch to, and which has a robust and maintained test suite. Considering the patch, we may ask, is this patch acceptable to apply and deploy. By this we mean…
Link verified 5 Sept 2026. Recent items update automatically from the channel.
Newsletter buttondown.email/nelhage His newsletter, described on his site as 'more experimental or in-progress thought'.
Link verified 5 Sept 2026.
Mastodon @nelhage@mastodon.social His Mastodon account.
Recent
- A while ago, @regehr was tooting about optimizing brainfuck, and I speculated that peeling off the first iteration of a loop might make the loop body easier to optimize: https://mastodon.social/@nelhage/113241955294210125 Today I learned that RPython/PyPy use a very similar idea to support loop optimizations in their trace optimizer: https://dl.acm.org/doi/epdf/10.1145/2480360.2384586 By peeling off a single loop iteration, the optimizer can "see" that the body proper is dominated by the first iteration, and specialize much harder. 26 Jun 2026
- What do you call it when someone writes a dramatization of the trials and tribulations of a bold emacs session heroically handling a keypress which results in writing a single character into the current buffer? `self-insert-command` fanfiction 11 Dec 2025
- After my earlier adventures benchmarking CPython, I ended up going deep down a rabbithole on CPU branch prediction and learning a bunch of interesting things, which I have attempted to write up and share: 4 Jul 2025 blog.nelhage.com
Show 17 more
- https://bughunters.google.com/blog/5424842357473280/zen-and-the-art-of-microcode-hacking Shot: "This opens up the potential for hardware attacks (e.g., reading the key from ROM with a scanning electron microscope), side-channel attacks (e.g., using Correlation Power Analysis to leak the key during validation), or other [...] attacks" Chaser: "We noticed that the key from an old Zen 1 CPU was the example key of the NIST SP 800-38B publication and was reused until at least Zen 4 CPUs." 5 Mar 2025
- @kees Actually you're probably the right person to ask: For unrelated reasons, I've noticed that -fzero-call-used-regs triggers before indirect tail calls, but *not* computed gotos, despite both compiling into similar indirect `jmp`s. Should those be made consistent? Do we worry about indirect `jmp`s being useful ROP gadgets? It seems clear they *could* be, but maybe in practice they're not actually helpful? Example: 1 Mar 2025 godbolt.org
- Compiler folks: anyone know how widely used -fzero-call-used-regs is? Looks like NixOS enables it by default(?) and I see references in kbuild, but anyone know if the more mainstream distros use it widely or at all? 1 Mar 2025
- Claude impressed me recently! A yak shave I was expecting to code by hand, and Claude did essentially all of it, instead. An experience report, and some reflections: 27 Jan 2025 blog.nelhage.com
- I wrote up Jaccard similarity and the MinHash trick for locality-sensitive hashing. I only learned about this algorithm recently and thought it was neat! 3 Jul 2024 blog.nelhage.com
- Things that are not the same: M4 (macros processor) M4 (macOS processor) 7 May 2024
- Seeing a lot of folks commenting on the "Tic-Tac-Toe in `printf`" post that's going around. If you like that one, I also enjoyed this writeup of a CTF challenge structured around a bytecode VM implemented as a single `sprintf`: 5 Apr 2024 ctftime.org
- (De)compression is conceptually quite straightforward: The decompression algorithm specifies a weird domain-specific VM, and the compressor is simply an automated code golfer for that architecture. 4 Apr 2024
- Help me out, cyclists of Mastodon: I want a rear basket for my bike (RadCity 5) that will also let me hang my panniers on either side of the rack. This seems to be a shockingly hard thing to find? Am I missing something? 28 Mar 2024
- I am slowly trying to figure out Nix/NixOS. I just realized something I'm very confused by: Is it really the case “pinning’ in NixOS prototypically means “pinning a single version of the entire package repository, and then pulling in whatever versions it contains”? I mean, "it's all just code" and it seems to be *possible* to mix and match via `fetchTarball`. But is pinning the universe as-a-unit really the normal pattern? 26 Jan 2024
- Is there really no way to pull the raw data storage from an app on an unrooted Android phone, for my own usage on my laptop? The internet really wants to suggest `adb backup` but also it seems like that hasn't worked for a while. 7 Jan 2024
- Two positions I believe, both pretty strongly: 1. LLMs -- deep learning in general -- is novel and weird and powerful and important. The world is becoming *much weirder* as a result, and we do not yet know how deep the rabbit hole goes; and 2. LLMs, as of today, are still quite limited and fall short of "human intelligence" in many ways; they are also relatively much better at "appearing intelligent" than "being intelligent," which makes them easy to overestimate 11 Dec 2023
- I got nerdsniped much harder than intended by Advent of Code and so here's a writeup of how to process a 20,000 element list in C++ template metaprogramming without blowing your compiler's stack, and also about manually compiling a regular expression into a compile-time state machine. Sorry but also not sorry: 8 Dec 2023 blog.nelhage.com
- Help me out, lazymastodon: Kinda random: if I wanted to get a dataset of growth rates of [a sizeable subset of] "tech companies" going all the way back to, like the '50s or '60s does anyone know have idea offhand where I might find such a thing or go about assembling such a thing? Willing to spend some amount of money or "stitching together sources" on it. 9 Oct 2023
- New blog post! This one, on (software) systems operating at or beyond their capacity, what goes wrong, and how to degrade gracefully. 7 Aug 2023 blog.nelhage.com
- https://www.intel.com/content/www/us/en/developer/articles/technical/advanced-performance-extensions-apx.html RISC architecture is going to change everything! (Link via @adrian ) 24 Jul 2023
- 45A/41D: absolute crime against crosswords. I had absolutely no way to ever get it. Horrific. I was enjoying the theme but now I've completely forgotten about it and am just steaming. #nytxw 20 Jul 2023
Link verified 5 Sept 2026. Recent items update automatically from the channel.
GitHub github.com/nelhage His code: livegrep, reptyr, llama, taktician.
Recent
- rules_boost — bazel build rules to use boost in bazel projects 21 Oct 2026
- ultimattt — A Rust Implementation of Ultimate Tic Tac Toe 19 Oct 2026
- reptyr — Reparent a running program to a new terminal 19 Oct 2026
Show 12 more
- crossme — A collaborative crossword-puzzle solver 16 Oct 2026
- New reposyncflowy 15 Oct 2026
- nelhage-sandbox — Sandbox repository for experimentation and small toy projects 13 Oct 2026
- nix-config — I guess I'm trying out Nix 13 Oct 2026
- elisp — nelhage's emacs configuration 23 Sept 2026
- coop — A Coq playground 26 Aug 2026
- llama 14 Jul 2026
- luawa — Experimental toy Lua implementation 3 Jul 2026
- nelhage.com — nelhage.com 17 Jun 2026
- dsync 15 Jun 2026
- arduinoproject — aggregator of random arduino projects 3 May 2026
- agemonia-reference — Some quick reference pages for Agemonia 16 Apr 2026
Link verified 5 Sept 2026. Recent items update automatically from the channel.
About the English under a post
Some people here publish in a language other than English. Where they do, this site shows a machine translation beneath the post, in this typeface — the site's own, not theirs.
The post itself is never changed, moved or hidden: what is set in the serif above is exactly what the person published, and it is what to quote them on. A translation can be wrong in ways that matter, especially about tone.
Only the post's own words are translated. A quoted post, a linked article and a belief on korrents.com are left in their original language.