The 'Astra can read and write weights to improve Qwen' post is misleading · honnibal.dev
The experiments that read raw weights and the ones that fix Qwen use completely different techniques.
Discover the resourceWe hand-pick product and design resources worth your attention, and deliver them to your RSS reader or inbox.
#438 Interactions instantanées: 3 méthodes pour designer une expérience différenciante
WhatsApp et Superhuman ont su se différencier notamment grâce à un sujet invisible : la performance. Rapide en apparence et en profondeur, on explore ensemble les solutions pour atteindre ces attributs forts. 📚 Pablooo.club, la librairie de screenshots iOS & Web collaborative et gratuite : https://pablooo.club 📰 La newsletter des meilleurs contenus design sélectionnés à la main : https://partageons.romainpenchenat.com/#follow 💬 Discutons ensemble sur LinkedIn www.linkedin.com/in/romain-penchenat/ 🔊 Jingle par Studio Module : www.studio-module.com This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit parlonsdesign.substack.com/subscribe
Discover the resource
Sparks of Weight-Space Programming: LLMs can natively read & write neural network weights
We find that frontier LLMs are surprisingly good at natively reading and writing neural network weights, well past what we’d expect from humans or conventional algorithms.
Discover the resource
A select should tell you something
Customizable select isn’t just a prettier dropdown. Five before/after demos show how options can carry what people need to decide.
Discover the resource
AI Design Critique Makes Human Reviews Better
AI design critique is not a replacement for a human design review. Run it first, brief it with a clean Figma zone and real project context, and the human session can spend its time on the decisions that actually matter. It improves quality and speed at once: more complete...
Discover the resource