Top Stories

Decoding the Obfuscated Bash Script on a Uniqlo T-Shirt

1387 points · tris.sherliker.net

Someone noticed the “code” printed on a Uniqlo t-shirt wasn’t just decorative gibberish — it’s a self-evaluating, obfuscated bash script, apparently served through Akamai’s CDN and printed onto retail merchandise. The author walks through decoding it layer by layer, and the mix of supply-chain curiosity, mild security paranoia, and pure “wait, that’s real code?” delight is exactly the kind of puzzle HN loves to pile onto.

Chatto Is Now Open Source

987 points · hmans.dev

Chatto has been released as open source, and the community response was immediate and enthusiastic. Developers are drawn to well-built tools getting freed up under permissive licenses, and the thread is full of people digging into the architecture and imagining what they’d build on top of it. Open-sourcing a polished project is a reliable way to top HN, and this one landed hard.

John Deere Owners Win Right to Repair Under FTC Settlement

889 points · apnews.com

After years of being the poster child for locked-down equipment, John Deere has agreed under an FTC settlement to let owners repair their own machines. This is a landmark moment for the right-to-repair movement — farmers have been fighting Deere’s software locks and dealer-only diagnostics for a decade. The HN crowd, which cares deeply about ownership and hardware freedom, sees this as a precedent that could ripple far beyond tractors.

Grok 4.5

662 points · x.ai

xAI has shipped Grok 4.5, its latest frontier model, and the benchmarks and capability claims are getting scrutinized in real time. Every major model release now triggers a familiar HN ritual: comparing it against the current leaders, poking at its reasoning, and debating whether the improvements are real or benchmark-gaming. With the frontier moving this fast, each new version resets the conversation about who’s ahead.

TypeScript 7

638 points · devblogs.microsoft.com

Microsoft has announced TypeScript 7, the milestone release built around the native (Go-based) compiler that promises dramatic speedups over the old JavaScript-based toolchain. For a language that now underpins a huge share of web development, a 10x-class compiler rewrite is a big deal, and developers are eager to see what the performance gains mean for large codebases and editor responsiveness.

Rewriting Bun in Rust

623 points · bun.com

The Bun team announced they’re rewriting the runtime in Rust, moving off Zig — a decision that sparked plenty of debate. Zig’s pre-1.0 instability and Rust’s larger ecosystem and safety guarantees are the stated motivations, and Zig creator Andrew Kelley even weighed in with his own thoughts. It’s a rare public look at the tradeoffs of betting a major project on a young systems language, and the HN engineering crowd is dissecting every angle.

Cloudflare Drop

448 points · cloudflare.com

Cloudflare rolled out a new offering it’s calling “Drop,” and the launch drew a big crowd on HN. Cloudflare product announcements reliably draw attention because so much of the web runs behind its infrastructure, and developers are quick to evaluate how each new service fits into their stack — or whether it deepens their dependence on a single provider.

FAANG Simulator

430 points · abeyk.com

A satirical browser game that lets you “escape the rat race” by simulating life inside a big tech company struck a nerve with the HN audience. The humor lands because it’s aimed squarely at the readership — endless promotion ladders, performance reviews, and the treadmill of corporate tech life. Games that hold up a funny mirror to the industry tend to spread fast here.

Separating Signal from Noise in Coding Evaluations

219 points · openai.com

OpenAI published a piece on the hard problem of measuring how good coding agents actually are, digging into why benchmarks can mislead and how to build evaluations that reflect real engineering work. As AI coding tools become central to how software gets written, the question of how to trust the claims about them is increasingly urgent, and this sparked a thoughtful discussion about what a meaningful benchmark even looks like.