À base de popopopop, mais pour le Hip-Hop je développe — former CEO @prototypoApp + @getinyo - @techstars alumni - Head of Product @mindflow_io
Lyon
Joined October 2011
- Tweets4.3K
- Following691
- Followers848
- Likes5.8K
Yannick Mathey retweeted
Meet Mistral Large 4, aka Le Chonk.
• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.
Open weights release end of October.
🤖 Made with AI
Great point of view, please start to love european efforts rather that bashing them. Go teams @MistralAI @Aleph__Alpha
Sorry, Seb, usually lots of love for your work, but how is Europe now having n=2 players in the LLM game a generational fumble for Mistral? Can we please just celebrate having more teams who try and build models, rather than immediately using this as a reason to hate on either? Europe‘s market is big enough for several sovereign players (btw, the world is also a pretty nice TAM) and different models (from my understanding Mistral is working on a much bigger model).
Yannick Mathey retweeted
Opus 5.5 designing LEGO 👀
I asked it to design a Microduck I can build with real LEGO pieces. It:
> designed it life-size using 1113 real LEGO parts
> verified: 3,204 connections, 0 collisions, every step buildable, centre of mass inside the feet 🤯
> made a 141-page LEGO-style booklet (237 steps)
> priced every piece in the browser and prepared the orders on BrickLink
Yannick Mathey retweeted
eventually everything turns to sand ...
Johann Wolfgang von Goethe, Zur Farbenlehre (Theory of Colors). 1810. via @MeditateColor
Yannick Mathey retweeted
« Navier Strokes »
Image by @OpenAI
xitter.cf/OpenAI/status/20973746…
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.
The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.
The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
Yannick Mathey retweeted
Our Head of Engineering, Gregory Roubinet, put a swarm of AI agents to the test. He built a simple agent that sends commands to a server and returns results, then connected it to a "supervisor" agent tasked with acting as a pentester.
First attempt: it refused. No authorization, no action, exactly the kind of guardrail you want.
So he added a human-in-the-loop step: one Slack message, one click to approve, and the flow picks up right where it left off. From there, the agent got to work, running nmap scans, checking services, and flagging an exposed MySQL instance in real time.
From there, adding more capability was just as simple: a couple of extra tools (Slack updates, Jira ticketing), one line in the prompt each time.
It's a good look at what agent swarms can actually do in practice: composable, tool-connected agents that get real work done, with a human still holding the risky calls.
Full video here: youtu.be/BjQfvEa_rqg?si=uL99…
"The romantic version of loops is that you write the loops and a thousand agents build your company overnight. The production version is that you write the loops, and most of your job is making sure they halt."
Counterintuitive, but so true. Harness FTW.
Yannick Mathey retweeted
Mistral Medium 3.5 is interesting less for the benchmarks and more for the positioning. Look at who they're comparing against: Kimi, Qwen, GLM, Claude (Sonnet). Not GPT, not Gemini. And i dont mean that in a negative way!
With Aleph Alpha being acquired by Cohere last week, Mistral is now the only non-US, non-Chinese lab still in the frontier conversation. At 128B dense with open weights, they're making a different bet than the Chinese MoE models in that chart (which activate only 17-40B params despite being 400B-1T total).
Mistral is trading inference efficiency for consistency. The Collie score (95.8, best in class by a wide margin) tells you where they're aiming: not raw reasoning, but the most reliable model to actually follow instructions in production. That's a European enterprise pitch, not a benchmark race.
Very solid release from Mistral!
Mistral Medium 3.5, a new flagship model in public preview by @MistralAI that merges instruction-following, reasoning, and coding into a single 128B dense model with a 256k context window and configurable reasoning effort. It's a new default model for Mistral Vibe and Le Chat. Released as open weights, under a modified MIT license.
Yannick Mathey retweeted
Not a native English speaker or writer.
With the help of the brilliant Notion AI (@NotionHQ), I finally wrote THE Startup Manifesto I’d been dreaming of. 42 lessons forged in the trenches of startup life.
Feels incredible to share it today →
thestartupmanifesto.com
🚀 Just dropped: THE Startup Manifesto. 42 lessons learned the hard way, building @Mindflow_io. No theater. No dogma. Just scars, sweat, and practices that actually work.
→ thestartupmanifesto.com
Think “Startuper’s Guide to the Galaxy.” 🧵👇
Insight number 18 will surprise you 😉
🚀 Just dropped: THE Startup Manifesto. 42 lessons learned the hard way, building @Mindflow_io. No theater. No dogma. Just scars, sweat, and practices that actually work.
→ thestartupmanifesto.com
Think “Startuper’s Guide to the Galaxy.” 🧵👇
The way we roll @ @Mindflow_io 👌
