ffmpeg+zeromq+haskell
Joined March 2024
- Tweets3.1K
- Following165
- Followers2.5K
- Likes5.6K
Ce monde me rend fou. This is crazy. J’espère vraiment qu’on aura pas une frappe nucléaire. Je déteste ce monde! L’espèce humaine est une honte pour l’univers.
BREAKING: Iran is now distributing iodine tablets to communities within a two to three kilometer radius of its sensitive nuclear facilities over fears of a US nuclear attack, a Health Ministry official confirms to The Times.
"Yes, we are distributing iodine. Trump can no longer be trusted. He may go through with this crime," the official said.
"Everyone knows that very little time remains. If America does not use a nuclear weapon against Pickaxe Mountain, it will be a miracle," an Iranian analyst with close ties to the IRGC said.
Senior figures across the government, the IRGC and the regular armed forces expect Pickaxe to be among the principal targets of the coming war.
Iraq activated nuclear emergency plans and border radiation monitoring yesterday.
le seul domaine qui « tabasse » le computer science.
We were taught that Earth orbits the Sun in a closed circle. In reality, Earth has never returned to the same point in space once in 4.5 billion years.
Right now, the Sun is carrying Earth and every planet around the center of the Milky Way at 230 km/s (828,000 km/h), more than 18,000 kilometers every 80 seconds.
Because the Solar System is tilted 60.2° to the galactic plane and bobs up and down through the galactic disc roughly every 30 million years, the planets trace open 3D corkscrew helices across deep space rather than flat loops.
The last time the Sun stood where it stands today, one Galactic Year (230 million years) ago, the very first dinosaurs were only beginning to appear on Earth. Since forming 4.6 billion years ago, our Sun has completed only 20 laps.
ajout du modèle gemini 3.5 transcribe… ce modèle dembedding est TOP.
je viens juste de déployer le modèle embedding-gemma 2 sur un vieux android. ce modèle est incroyable, j’arrive à retrouver de vieilles photos des années 2020 sur bcp de sujets. le modèle peut même retrouver des discussions whastapp hh.
Presque tous mes projets se basent sur la programmation asynchrone et la bibliothèque ZeroMQ. Plusieurs threads communiquent via inproc ou IPC. Ce pattern permet de résoudre facilement des problèmes complexes.
Finalement c'était plus rapide que prévu. J'ai réussi à adapter mes anciens scripts avec le nouveau gpt-live. le modèle peut contrôler la rc car. plot twist sur la fin de la vidéo hh. le script sera sur mon github.
ce week-end insha Allah, je vous montrerai une démo de comment l’intégrer dans un téléphone. le script sera dispo sur mon github. maybe on pourra l’étendre sur le rc car 🤔.
ce thread :). voir tout ce potentiel en 2020 montre une bonne capacité d’analyse et de problem solving.
GPT3 writing code. A compiler from natural language to code.
People don't understand — this will change absolutely everything. We're decoupling human horsepower from code production. The intellectual equivalent of the discovery of the engine. player.vimeo.com/video/42681…
en fait, les modèles précédents 2023-2025, avaient du mal à jouer ces 3 roles(actor, evaluator, reflection) donc on était obligé de contourner ces limites via des hooks etc… mais maintenant, les modèles ont intégré ces 3 composantes. ils sont assez « smart ».
on a parlé de ça hier dans une autre discussion, vous n’avez pas besoin de skills, mcp, hooks etc… ça, c’était pour les versions d’avant. les nouveaux modèles comme opus 5 ont une bonne capacité d’auto évaluation. rien de mieux qu’un simple setup. pas de multi agent etc…!
"delete your claude.md file, delete your skills, and delete your hooks. Then see what the model does. It might surprise you."
- Claude Code creator at Anthropic Boris Cherny ( @bcherny )
"For people who aren’t building agentic products but are using Claude Code, every 6 months, delete your claude.md file, delete your skills, and delete your hooks. Then see what the model does. It might surprise you.
For Opus 5, we strongly recommend trying to delete all of these things because the model may no longer need the extensive instructions that were necessary for previous models."
- at Y Combinator Startup School 2026.
----
From "Y Combinator" YouTube channel, (full video link in comment)
Si jamais vous voulez suivre la session d’hier, vous pouvez cliquer sur le drive link, On va en faire d’autres. Next Time, @lo_jimy prendra la main, insha Allah.
Merci bro 🙏🏼 le meet est dispo ici drive.google.com/file/d/1PF4…
le code est sur github, n'hésitez pas à le changer, le modifier et faire ce que vous voulez avec hh. github.com/milkymap/agent_lo…
Insha ALLAH, 2h 30 de code(build advanced ai agent from scratch with gemini sdk). Samedi 5 Septembre 22h-00h-30 Time zone: Europe/Paris. soit 20h, heure Sénégal. meet.google.com/hxx-gdva-zsh
Insha ALLAH, 2h 30 de code(build advanced ai agent from scratch with gemini sdk). Samedi 5 Septembre 22h-00h-30 Time zone: Europe/Paris. soit 20h, heure Sénégal. meet.google.com/hxx-gdva-zsh
un thread à suivre…!
I’ve successfully run the full retained MaleCNS v1.0 fruit fly connectome, all 166,700 neurons, inside Minecraft, with its simulated neural activity driving a fly’s movement.
V1 Currently in development. Built with the help of GPT-6 Astra.
Props to the @OpenAI team and @thsottiaux for this release. Code and mod coming soon!
Humm, ça c’est pas de la mémorisation. ces modèles ont une bonne capacité de search, de recombinaison. il est difficile de prédire le futur…! arc agi 3 solved!
Ibraheem Khalil retweeted
GPT-6 Astra represents a step-function change in model capability for interactive reasoning problems. It scores 66% on ARC-AGI-3 using our standard harness, and nearly 100% with a continuous conversation harness and custom compaction, at a cost of roughly $360 per game.
In fact, the continuous harness version significantly outperforms our human baseline in action efficiency across almost all levels. When we examined the reasoning chains to understand how the model operates, we found it performing highly efficient, on-the-fly symbolic world modeling for each game and level. It goes as far as developing its own shorthand DSL to represent in-game situations -- essentially a game-specific algebraic notation.
Overall, Astra exhibits symbolic modeling behaviors we had previously only seen with sophisticated harnesses -- so harness capabilities are increasingly shifting into the model itself.
We see Astra as a major breakthrough in model intelligence.
Read our post on Astra and what these results mean: arcprize.org/blog/astra
