frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Accidentally Turing-Complete

https://beza1e1.tuxen.de/articles/accidentally_turing_complete.html
25•bschne•1y ago

Comments

panstromek•1y ago
Nice list. Some of those are arguably not accidental, TypeScript type system seems kinda obvious to be turing complete when it tries to describe dynamically typed langauage.
WalterGR•1y ago
x86 MOV instruction: “The mov-only DOOM [game] renders approximately one frame every 7 hours, so playing this version requires somewhat increased patience.”
a_cardboard_box•1y ago
Rule 110 is only Turing-Complete if you have an infinitely large array of cells, and are able to initialize it with an infinite repeating pattern. If I'm not mistaken, HTML+CSS can only do a fixed-sized array.

With a Turing-Complete language, if a program runs out of memory on one machine, you can run the same code on a bigger machine without modifying it, and it can use the additional memory. With fixed-length rule 110, you need to modify the code if you want to use more memory.

256_•1y ago
This is addressed in the second paragraph of TFA:

"Stuff which is somehow limited (stack overflows, arbitrary configuration, etc) is still considered Turing complete, since all "physical" Turing machines are resource limited."

In my opinion, worrying about infinite memory, in regards to Turing completeness, makes the task of implementing computation much less interesting.

Also, I'm pretty sure CSS only does one generation (or a finite number of them) before stopping anyway.

256_•1y ago
Logic in Doom is particularly interesting to me. Apparently you can fit ~64k logic gates in a map (using the method described). From [1]:

"As the DOOM engine was not designed to be an interpreter, there are some constraints on our programs written against it. The biggest one is how large our programs can be. Since each gate uses at least one tag, we can use this as a metric to derive an upper-bound on the size of a program. As the DOOM engine uses 16-bit tags, this means we can have, at most, 65535 gates. This is not a particularly large number. We may be able to implement a very small CPU but this limit will be hit pretty quickly I believe."

The z80 had ~8,500 transistors. The 8086 had ~29,000 (checking Wikipedia). You could get far fewer if you use a 1-bit microarchitecture, I'm sure. I think there was a DEC (PDP?) computer that used that trick to have a really low transistor count, but I don't remember what it was called.

The real problem is RAM; for this you may as well cheat and modify Doom's code to add a RAM chip, and I/O while you're at it.

You could create a CPU in Doom implementing an architecture for which a C compiler exists, capable of compiling Doom, and run it in the CPU in Doom. For "reasonable" speed you'd have to do more than one simulation step per frame render (in the host Doom). If you ran it for long enough maybe you could get a full frame of Doom in Doom.

[1]: https://calabi-yau.space/blog/doom.html

karmakaze•1y ago
Doom running in TypeScript static type checker[0].

> half trillion lines of types totaling 177 terabytes ran through the type checker around the clock for 12 days to get the first frame

[0] https://news.ycombinator.com/item?id=43184291

karmakaze•1y ago
My favorite one is Conway's Game of Life. It's perhaps the least surprising one, but it's also the most visually appealing. Really like this video that leads up to making the Game of Life in itself[0]. It's something you can show a non-technical person and they can get a sense of how crazy it is that something so simple can do anything.

[0] https://www.youtube.com/watch?v=Kk2MH9O4pXY

Alphabet in talks to acquire assistant startup Howie AI

https://www.axios.com/pro/merger-deals/2026/11/02/google-alphabet-howie-ai-assistant
1•simonpure•2m ago•0 comments

h1: A Taste of the Bitter Lesson in Automated Theorem Proving

https://arxiv.org/abs/2610.35962
1•matt_d•8m ago•0 comments

The Hitchhiker's Guide to Training on Compute Islands – Core Automation

https://www.coreauto.com/hitchhikers-guide-to-training-on-compute-islands
2•mkmk•8m ago•0 comments

Anthropic's growth just flattened (really)

https://www.regenerator1.com/p/anthropics-growth-just-flattened
3•cdrnsf•10m ago•0 comments

Shotwell-Bench: An annotation benchmark that accepts valid alternatives

https://shotwell.ai/blog/shotwell-bench/
1•aliabd•11m ago•0 comments

Turning elementary arithmetic into an enterprise-wide operational capability

https://fizzbuzz.biz/about/
1•cobbzilla•12m ago•0 comments

The SEC Should Block Anthropic's IPO to Protect the Public

https://www.techpolicy.press/the-sec-should-block-anthropics-ipo-to-protect-the-public/
5•cdrnsf•12m ago•0 comments

V100 SXM2 pinout and NVLink reference

https://github.com/pigeon-labs-stack/v100-sxm2-reference
1•airhangerf15•13m ago•0 comments

Internals of FP8 MMA on Hopper

https://simonveitner.substack.com/p/internals-of-fp8-mma-on-hopper
1•matt_d•14m ago•0 comments

The AI bubble could expose your deepest secrets

https://www.bbc.com/future/article/20261029-the-ai-bubble-could-expose-your-deepest-secrets
3•mrkn1•16m ago•0 comments

LLMs are not superhuman at reading and writing neural network weights

https://www.seangoedecke.com/llms-are-not-superhuman-at-reading-and-writing-neural-network-weights/
4•agrm•16m ago•0 comments

Most of our measures of a key cognitive process don't agree

https://arstechnica.com/science/2026/11/most-of-our-measures-of-a-key-cognitive-process-dont-agree/
2•rbanffy•17m ago•0 comments

The latency cost of hosting an API in India

https://latencyradar.com/learn/hosting-in-india-api-response-times/
1•marcascode•17m ago•0 comments

I discovered a sound that leaks your exact location (and it's everywhere) [video]

https://www.youtube.com/watch?v=UytYrhYLOKI&list=TLPQMDIxMTIwMjaw69sXf7oe4A&index=10
4•fortran77•22m ago•1 comments

Web host Squarespace lays off 19% of its employees

https://techcrunch.com/2026/11/02/web-host-squarespace-lays-off-19-of-its-employees/
3•bobbiechen•23m ago•3 comments

Making CUDA Checkpointing 15x Faster on a B300: A Worklog

https://maknee.github.io/blog/2026/Cuda-Checkpointing/
1•matt_d•24m ago•0 comments

Show HN: HackerProducts – HackerTube, but for products mentioned on HN

https://hackerproducts.pages.dev
1•bin01•26m ago•0 comments

Trump vows to protect Musk, floats pre-emptive staff pardons

https://www.forbes.com/sites/saradorn/2026/11/02/trump-says-he-might-pardon-musk-if-democrats-win/
4•TheAlchemist•26m ago•1 comments

Three Dots

https://randsinrepose.com/archives/three-dots/
1•mooreds•29m ago•0 comments

Apple reports fourth quarter results

https://www.apple.com/newsroom/2026/11/apple-reports-fourth-quarter-results/
7•mfiguiere•29m ago•0 comments

I'm Turning into a Project Manager

https://www.310networks.com/posts/im-turning-into-a-project-manager/
1•kookster310•30m ago•0 comments

Tariff talk on S&P 500 earnings calls fell 80%, and refunds now drive most of it

https://equibles.com/research/are-companies-still-talking-about-tariffs-evidence-from-4-952-s-p-5...
1•daniel3303•33m ago•0 comments

A benchmark of data-analysis traps for coding agents

https://github.com/boranfurkan/trustworthy-data-analysis
1•boranfurkan•34m ago•0 comments

How Open Models Win Against OpenAI and Anthropic [video]

https://www.youtube.com/watch?v=zsO5RktNB7k
1•pretext•35m ago•0 comments

On Learning and LLMs

https://julienvincent.io/posts/on-learning-and-llms
3•julienvincent•35m ago•0 comments

China publicly discloses plan for a lunar landing within three years

https://arstechnica.com/space/2026/11/china-releases-timeline-for-lunar-landing-on-the-moon-targe...
1•motionlessveloc•35m ago•1 comments

Why dogs catch yawns from their owners but not other dogs

https://www.economist.com/science-and-technology/2026/10/28/why-dogs-catch-yawns-from-their-owner...
1•vinni2•37m ago•0 comments

Who Powers the Token Economy?

https://hxu296.github.io/ai-compute-landscape/
1•hxu296•37m ago•0 comments

OAuth Client ID Metadata Document

https://oauth.net/2/client-id-metadata-document/
1•mooreds•38m ago•0 comments

Stop Hogging RAM You Monster

https://ramhogger.monster/
1•cooldevcoding•38m ago•1 comments