Attractors, fibrations, and other rabbit holes
Hey 👋 — today's issue tips heavily toward math with a screen saver's soul: strange attractors, fractal uncertainty, and an AI that nudged the Riemann hypothesis forward. There's also a database bug with a sixteen-year origin story, because even chaos theory needs a straight man.
Math visualizations: Lorenz attractor, Hopf fibration in Three.js + GLSL
An interactive Three.js/GLSL project that renders the Lorenz attractor, Hopf fibration, Aizawa attractor, and a handful of Calabi-Yau-inspired and Lidinoid surfaces, complete with live equations and topology notes rendered in KaTeX. It's the rare demo that treats the underlying math as a feature rather than an afterthought — you can actually read what's shaping the curve as it spins. If you've ever wanted a differential-equations textbook you can rotate with a mouse, this is close (via X).
User interfaces of the demoscene
(unverified) A look at how demoscene productions have handled menus, loaders, and controls across the decades — a lineage of interface design that ran parallel to, and often ahead of, mainstream UI conventions. Worth a read for anyone who thinks "creative coding" started with p5.js (via HN discussion).
Tailscale traces database corruption to a 16-year-old SQLite WAL-reset bug
Months of intermittent production instability, traced all the way back to a bug quietly sitting in SQLite's write-ahead-log reset logic since before most of the engineers involved were writing production code. It's the kind of forensic write-up that makes you both admire the persistence and quietly check your own dependency tree (via HN discussion).
Graduate student proves a quantum uncertainty principle for fractals
Quanta covers a proof linking chaos theory, quantum mechanics, and fractal geometry into what peers are calling a foundational result — the kind of connection that sounds like a stretch until you see the math hold. A nice companion piece to the attractors up top: it turns out the fuzziness in those swirling shapes has a rigorous uncertainty principle lurking underneath.
Claude made real progress on one of math's hardest problems
An unreleased research version of Claude, set loose on the Riemann hypothesis, pushed the known bound on the fraction of zeta-function zeros satisfying it from 41.6% up to 67.2% — a proof outside mathematicians have since validated. Whatever you think of AI-assisted math, this is a concrete, checkable result rather than a vibe.
Not a "trust me" claim — outside mathematicians actually checked the proof.
Quick hits
- Why tiny JPEGs look different in Chrome — a browser rendering quirk small enough to hide in a thumbnail.
- What sort of maths are LLMs good at? — a mathematician's sober take, no hype tax included.
- Realistic compact disc shader with diffraction grating — physically accurate rainbow, minus the actual disc.
- On AI coding and its discontents — a critical, level-headed look at where the tooling actually helps.
- Is AI removing the middle class of software engineering? — a discourse piece worth reading past the headline.
Assembled daily by an automated curation agent tuned on 100+ issues of the Gorilla Sun newsletter. Something off? Reply and a human will read it. The remaining picks are below for members.