Steam Frame starts at $1059

XCancel service is suspended until further notice

OpenAI bots knew about the RubyGems caching vulnerability

An atlas of periodic solutions to the three-body problem

What are you working on? What have you been curious about lately?

Ask HN: What are you working on? (September 2026)

Pion, an agent designed to run any company autonomously

EuroBirdPortal – Live bird movements across Europe

Distributed Systems Classics (2017)

Microsoft patches Windows and Excel – breaks audio, remote access, and paste

Principles for Fast Tokio Applications

A Beginning for Mathematics

Dario, Please

Oracle’s 6am layoff emails hit staff amid new wave of cuts

How my e-reader lost its stripes

Show HN: Neobrutalism.dev – Just added Base UI support and added new color theme

Amazon vs. Perplexity – U.S. Court of Appeals for the Ninth Circuit

Notes on gotchas while migrating 35kb preprompts from Opus to self-hosted Ollama

In November and December 2025, inspired by Simon Willison’s pelican-riding-a-bicycle benchmark, I had some then-current LLMs create SVGs from thirty similar prompts, such as “Generate an SVG of an octopus operating a pipe organ.” Simon mentioned that experiment on his blog [1].<p>Nine months have passed and much stronger models have been released, so I tried the experiment again today. The linked site shows the results.<p>Running ten of the prompts through six models at OpenRouter cost about twenty dollars, so I stopped there for now.<p>[1] <a href="https:&#x2F;&#x2F;simonwillison.net&#x2F;2025&#x2F;Nov&#x2F;25&#x2F;" rel="nofollow">https:&#x2F;&#x2F;simonwillison.net&#x2F;2025&#x2F;Nov&#x2F;25&#x2F;</a>

Show HN: Pelican-bicycle alternatives

Why don't machine learning research agents overfit?

Adversarial Fashion Makes a Statement on AI Panopticon