Hacker News: manx

New comment by manx in "Show HN: Pi-Infinite-Context"

manx — Tue, 13 Oct 2026 15:55:33 +0000

Which evals would you be interested in specifically? And what are you missing in the readme?

New comment by manx in "Show HN: Pi-Infinite-Context"

manx — Tue, 13 Oct 2026 15:15:11 +0000

valid question. I thought a lot about this. The prefix before every fold stays intact (cache hit). But for the messages after the fold there is always a cache miss. Folds don't happen too often, so in practice this turned out to be worth it, because the cache miss comes with a big token reduction in general. I don't have a solid cost analysis for this yet, though.

Show HN: Pi-Infinite-Context

manx — Tue, 13 Oct 2026 15:05:42 +0000

Hi HN! Since I created this extension and heavily refined its core design, I'm not worrying about context size anymore. It just works so well. So I wanted to share. It is stress-tested with OpenAI and Anthropic models.

Since HN is sensitive to good readmes, I wanted to do it well here. I started writing it by hand, but noticed that it was not great. then I decided to go back to LLM help, but iterate on it until it was short, simple and readable. I hope you like it.

In this process I had an insight about AI generated content. Slop is not only the fact that AI was used, but that the creator did not care about a high quality result. I'm interested to hear what you think about this take.

Comments URL: https://news.ycombinator.com/item?id=50072881

Points: 2

# Comments: 6

New comment by manx in "GPT-6 Sol and Luna"

manx — Wed, 23 Sep 2026 12:20:31 +0000

Yes, you need an extension that uses the claude code credentials from the file system, like this: https://github.com/fdietze/pi-claude-auth

Works pretty well for me, even with latest opus-5-5

Mental Models for LLMs

manx — Mon, 14 Sep 2026 13:02:04 +0000

Article URL: https://felx.me/posts/mental-models-for-llms/

Comments URL: https://news.ycombinator.com/item?id=49696145

Points: 1

# Comments: 0

New comment by manx in "Show HN: OSS Agent I built topped the TerminalBench on Gemini-3-flash-preview"

manx — Mon, 27 Apr 2026 15:42:59 +0000

We probably want to compare the cartesian product of model+harness.

New comment by manx in "Servo is now available on crates.io"

manx — Mon, 13 Apr 2026 15:32:48 +0000

Agreed. Which other software does society need badly?

AI Fixes the Bullshit Asymmetry

manx — Tue, 07 Apr 2026 14:24:13 +0000

Article URL: https://www.konstantinschubert.com/2026/03/31/ai-the-bullshit-defense.html

Comments URL: https://news.ycombinator.com/item?id=47675887

Points: 2

# Comments: 1

Moore's Law for Everything (2021)

manx — Tue, 24 Mar 2026 15:39:37 +0000

Article URL: https://moores.samaltman.com/

Comments URL: https://news.ycombinator.com/item?id=47504292

Points: 2

# Comments: 0

Cliodynamics: History as Science

manx — Tue, 24 Mar 2026 15:14:40 +0000

Article URL: https://peterturchin.com/cliodynamics-history-as-science/

Comments URL: https://news.ycombinator.com/item?id=47503891

Points: 1

# Comments: 0

Fragile States Index

manx — Tue, 24 Mar 2026 14:52:57 +0000

Article URL: https://fragilestatesindex.org/

Comments URL: https://news.ycombinator.com/item?id=47503510

Points: 2

# Comments: 0

New comment by manx in "Floci – A free, open-source local AWS emulator"

manx — Sun, 22 Mar 2026 01:33:37 +0000

What matters more is if there is good QA.

Ask HN: How do you deal with long Covid?

manx — Thu, 12 Feb 2026 05:03:38 +0000

What is your understanding of the root causes of the symptoms? What works for you, what doesn't? What was your biggest insight?

Comments URL: https://news.ycombinator.com/item?id=46985070

Points: 4

# Comments: 2

Google deprecates Gemini-2.5-pro

manx — Thu, 05 Feb 2026 11:33:37 +0000

Article URL: https://ai.google.dev/gemini-api/docs/deprecations

Comments URL: https://news.ycombinator.com/item?id=46898546

Points: 1

# Comments: 0

New comment by manx in "Stop Doom Scrolling, Start Doom Coding: Build via the terminal from your phone"

manx — Wed, 07 Jan 2026 03:14:28 +0000

Also, you can have NixOS instead of debian: https://github.com/nix-community/nixos-avf

New comment by manx in "Grokipedia and the coup against reality"

manx — Tue, 28 Oct 2025 19:58:25 +0000

Wouldn't it make sense to have these wikipedias created by different llms, and compare them to expose their biases?

New comment by manx in "Testing a compiler-driven full-stack web framework"

manx — Wed, 08 Oct 2025 18:51:51 +0000

It's not only the amount of code but also the quality of the available code. If a language has a low barrier to entry (e.g. python, javascript), there will be a lot of beginner code. If a language has good static analysis and type checking, the available code is free of certain error classes (e.g. Rust, Scala, Haskell).

I see that difference in llm generated code when switching languages. Generated rust code has a much higher quality than python code for example.

Mobile NixOS

manx — Thu, 02 Oct 2025 21:08:49 +0000

Article URL: https://mobile-nixos.github.io

Comments URL: https://news.ycombinator.com/item?id=45455583

Points: 8

# Comments: 0

New comment by manx in "Learning Languages with the Help of Algorithms"

manx — Sun, 21 Sep 2025 10:08:08 +0000

I fully agree with this approach! 5 years ago I built a prototype to execute that same concept of language covering. But instead of just using words, I used n-grams. It ii trained on subtitles to model spoken language. Combined with sqlite in the browser to get the next sentence with the most impact.

github here: https://github.com/fdietze/ravioli

prototype deployed here: https://raviolio.web.app/

New comment by manx in "Learn Rust the Right Way"

manx — Mon, 15 Sep 2025 07:03:11 +0000

I think memory safety is a weak argument for rust and only works against C/C++, as you mention. For me, the strongest arguments against GCed languages are static analysis goodies that other languages don't have. Like checked immutably by the compiler, the borrow checker which forces you into a certain architecture, fearless concurrency etc. All those lead to fearless refactoring, which is a very strong point for me.