frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Fine-tuning an LLM to write docs like it's 1995

https://passo.uno/fine-tuning-docs-llm/
25•taubek•1h ago

Comments

mock-possum•26m ago
> we’re not there yet, in part because of how much more powerful connected frontier models are

Is that why though? You need a beast of a machine to run a functional local model in my experience.

I think the big part is there’s significant sticker shock to buying capable hardware.

That said,

> weekend. I chose to try fine-tuning on two models, Llama 3.1 8B Instruct and Qwen 2.5 7B Instruct. At their size (around 8B) they run comfortably on a MacBook Air

Perhaps I spoke too soon?

Anyway

> I chose the Microsoft collection as the source of training materials. The collection contains out-of-print docs published between 1977 and 2005: more than 37 million words, covering old systems and SDKs

this strikes me as a very specific brand of 1995’s prose, spanning about 30 years. It’s a cool article though, so maybe that’s a forgivably clickbaity title.

mschild•14m ago
Running models locally is surprisingly easy and possible even on older hardware.

Obviously not the largest, up-to-date models but for what I expect most people use them for, even on hn, there are some shockingly good models that dont require €4k machines.

I have a desktop with an AMD 6900XT and 5600 with 32GB ram. Obviously no slouch but its several years old at this point. I can comfortably run qwen 3.5 9b and get a speedy 60 token/sec output with decent results.

mock-possum•11m ago
idk I can barely field a 14b on my desktop, and it’s rough trying to replicate the agentic pair programming experience I’m accustomed to with Claude. And I don’t mean it doesn’t work as well, I mean it doesn’t work.

Is there some secret I’m missing? I’ve tried rolling my own harness, and tried a few of the ones the cool kids use - I think pi was the most recent. Not quite my tempo, I’m afraid.

OJFord•11m ago
> this strikes me as a very specific brand of 1995’s prose, spanning about 30 years.

It's probably a fair approach to say the significant influence (training dataset) on writing at a particular time is the preceeding 30 years' material? It's certainly not only what's already written that year (nor anything since).

vintagedave•24m ago
I love old-school docs, and this was a fantastic read. But, I couldn't see the three generated doc pages linked anywhere. Did I miss something?

I'd really like to see the Win2K-style docs on REST, for example.

C++: The Documentary Released Today

https://herbsutter.com/2026/06/04/c-the-documentary-released-today/
67•ingve•2h ago•6 comments

Meta enables ADB on deprecated Portal devices [video]

https://fb.watch/HxPu0fSyeH/
199•jenders•6h ago•61 comments

Fine-tuning an LLM to write docs like it's 1995

https://passo.uno/fine-tuning-docs-llm/
25•taubek•1h ago•5 comments

Anthropic's open-source framework for AI-powered vulnerability discovery

https://github.com/anthropics/defending-code-reference-harness
387•binyu•11h ago•112 comments

Open Code Review – An AI-powered code review CLI tool

https://github.com/alibaba/open-code-review
154•geoffbp•7h ago•31 comments

Do transformers need three projections? Systematic study of QKV variants

https://arxiv.org/abs/2606.04032
151•Anon84•8h ago•31 comments

Azure Linux 4.0 is Microsoft's first general-purpose Linux

https://www.boxofcables.dev/azure-linux-4-0-is-microsofts-first-general-purpose-linux/
98•haydenbarnes•4h ago•65 comments

WiFi Time

https://mitxela.com/projects/wifi_time
49•surprisetalk•2d ago•2 comments

The IsUpMap lets you check the status of over 100 major sites at once

https://isupmap.com/
18•mikelgan•2h ago•6 comments

I'm skeptical about efforts to revolutionize schooling

https://www.scotthyoung.com/blog/2026/05/27/revolutionize-schooling/
149•andrewstuart•2d ago•227 comments

Watching a Z80 from an RP2350

https://emalliab.wordpress.com/2026/05/26/watching-a-z80-from-an-rp2350/
9•ibobev•2d ago•0 comments

Branchless Quicksort faster than std:sort and pdqsort with C and C++ API

https://tiki.li/blog/blqsort
162•birdculture•2d ago•35 comments

Delacroix's Entry of the Crusaders into Constantinople Restored

https://www.louvre.fr/en/explore/life-at-the-museum/delacroix-s-entry-of-the-crusaders-into-const...
17•rawgabbit•4h ago•4 comments

Magenta RealTime 2: Open and Local Live Music Models

https://magenta.withgoogle.com/magenta-realtime-2
20•selvan•3h ago•4 comments

SpaceX, Other Mega IPOs Denied Fast Index Entry by S&P

https://www.bloomberg.com/news/articles/2026-06-04/s-p-dow-jones-keeps-megacap-ipo-rules-as-is-af...
478•tristanj•8h ago•219 comments

Linear Cosine Palettes(2025)

https://blog.djnavarro.net/posts/2025-09-14_cosine-palettes/
12•num42•3h ago•0 comments

Go Experiments Explained

https://www.alexedwards.net/blog/go-experiments-explained
21•ingve•3d ago•8 comments

Reverse-Engineered Userspace Driver for Asus ZenVision Lid OLED on Linux"

https://github.com/tarpediem/zenvision-linux
47•berlianta•2d ago•10 comments

The Pentagon is running an AI propaganda mill targeting Latin America

https://theintercept.com/2026/06/02/la-tilde-propaganda-latin-america-pentagon/
54•Fricken•2h ago•33 comments

Samurai City

https://worksinprogress.co/issue/samurai-city/
151•zdw•3d ago•29 comments

When AI Builds Itself: Our progress toward recursive self-improvement

https://www.anthropic.com/institute/recursive-self-improvement
435•meetpateltech•15h ago•574 comments

The Causes of Long Covid

https://www.science.org/content/blog-post/causes-long-covid
103•maxall4•4h ago•56 comments

KVarN: Native vLLM backend for KV-cache quantization by Huawei

https://github.com/huawei-csl/KVarN
130•theanonymousone•16h ago•13 comments

VoidZero Is Joining Cloudflare

https://blog.cloudflare.com/voidzero-joins-cloudflare/
625•coloneltcb•18h ago•271 comments

Queen bees emerge from special wax chambers

https://cen.acs.org/materials/biobased-materials/queen-bees-special-wax/104/web/2026/06
70•gmays•9h ago•10 comments

Retro-Tech Parenting

https://havenweb.org/2026/05/28/retro-tech.html
283•mawise•15h ago•196 comments

JLink JTAG Access on the Pinecil

https://danielmangum.com/posts/jlink-jtag-pinecil/
55•hasheddan•2d ago•10 comments

WSL 2 is getting faster Windows file system access

https://www.boxofcables.dev/wsl2-per-device-swiotlb-pools-for-virtiofs-and-virtioproxy/
121•haydenbarnes•12h ago•91 comments

Castor: CERN Advanced STORage Manager

https://castor.web.cern.ch/content/home.html
54•naves•11h ago•23 comments

Making Debian or Fedora persistent live images

https://sigwait.org/~alex/blog/2026/05/28/smdBC8.html
71•henry_flower•3d ago•8 comments