Recently I’ve been thinking about how AI has reshaped everyday life and the way we communicate. I wrote down some thoughts.
thoughts.rarewood.dev/ai
gg @iamakulov @kurtextrem you guys rock
We spent a year working with the Lighthouse team on how it simulates slow phones. Lighthouse 13.6 fixes two flaws that made fast Framer sites look slow in PageSpeed Insights. Your site was already fast. Now your score shows it. Full write-up: framer.com/blog/lighthouse-1…
Cédric retweeted
There's a buzzy new paper about GPT-6 Astra being able to directly interpret model weights (prompts full of floating point numbers) - I tried it with Qwen 3.8 27B on my laptop and found that, after 45+ minutes, that model could achieve the same thing simonwillison.net/2026/Nov/3…
Harness 2 for Framer Agents is here. Agents start making changes as they work through your prompt, see progress sooner and finish bigger tasks faster.
It’s more capable, up to 33% cheaper, and long tasks are 20% faster. Read the deep dive by @hemlok_ framer.com/blog/harness-2/
Cédric retweeted
Être chanceux ça invalide pas ton taf. J’ai eu de la chance dans mon parcours et j’ai bossé. Vous voulez qu’on vous suce la bite parce que vous auriez une éthique de taf hors normes. Non à un moment ou un autre vous avez eu de la chance. C’est tout frère
Cédric retweeted
This is the coolest video you will watch today!
@Zai_org's GLM 5.3 Flash NVFP4 (full local on 4 sparks) on a full @BalatroGame playthrough and WINNING in realtime!
It took me a good amount of time to build a nice harness for this, but now we have it.
Which model should I try next?
We are releasing → classcad.ai today: AI driven embeddable CAD. If you ever enjoyed my OSS projects (zustand, react-spring, r3f, pmndrs), these all came out of CC. 🍻
A small 🇨🇭 team has been working on CAD for so many years, release seemed always out of reach. This is a huge milestone for us and i'd appreciate a RT. 🙏
Cédric retweeted
Evercut is live!
capture screenshots. record your screen. edit everything. share in seconds.
cinematic zooms, overlays, voiceovers, OCR, annotations, transcript editing and much more...
start for free with 7 days trial.
Cédric retweeted
TypeGPU + ruNNtime + Jev @typesafeai is a very fun combo :D
ruNNtime gives me efficient local inference, TypeGPU lets inference and rendering share GPU resources directly with zero copy. That’s 3 separate NN inferences plus rendering, all happening in realtime
Since we control the pipeline, Jev can just sit in the middle and add the semantic bit.
camera + mic → Moonshine + YOLO26 + DepthART → Jev → lights, shadows and bloom











