After 3 incredible years at @tinybird, it's time for my next adventure. I really enjoyed other people's tales of what companies are like from the inside, so here are my reflections on Tinybird: rbarbadillo.github.io/tinybi…

ML4 is now open-weight! Frontier capabilities on code agents & cyber, way more efficient and compact than its Preview version three weeks ago (!) and cherry on the cake: technical report telling its epic backstory. Check it out: huggingface.co/mistralai/Mis…

nada mejor

Robert Caro on loneliness:

If only there were a mechanism for sorting thing by value

Welcome to The New Standard. Super special evening in Paris with @raycast @linear and @MistralAI 🧡

Mistral Large 4 (high) uses 82K output tokens per Intelligence Index task and sits on the open weights Pareto frontier, while open weights models that score higher use 113K or more. Compared to the Preview, Mistral Large 4 (high) uses 39% fewer output tokens per Intelligence Index task

Running in Mistral Vibe, Mistral Large 4 (high) ranks second of the current open weights models on the Artificial Analysis Coding Agent Index (48), between GLM-5.3 (max) in Opencode (49) and DeepSeek V4.1 Flash 1017 (max) in DeepSeek Harness (46), while using about half as many total tokens per task (8.5M vs. 15.9M and 16.7M)

me this week playing around with le chonk

it's out 🐈 huggingface.co/mistralai/Mis…

ML4 is now quite token efficient, solving one of the major issue from the preview. If you've tried it before give it another spin, you'll be surprised

Estoy consternado con la tragedia de los trenes de Toledo. Parece que varias fuentes confirman el fallecimiento de mi querida amiga Fátima Báñez que viajaba en el primer vagón del Iryo. Escribo esto con enorme tristeza. Fátima era una mujer de increíble vitalidad y energía, lista como nadie, pragmática, con mucho sentido del humor, cariñosa, luchadora, nada sectaria y con enorme sensibilidad por las cuestiones sociales. He conocido a mucha gente en política. Si tuviera que decir de alguien que estaba en política por las razones correctas, esa era Fátima. Nos conocimos en la negociación de gobierno. Nos emparejaron para negociar las políticas sociales. Nos hicimos amigos inmediatamente y logramos muchas cosas. De aquél acuerdo, la medida social implementada más importante fue la primera gran extensión de permisos de paternidad que fue posible gracias a su empeño y compromiso. Desde entonces no hemos parado de hacer cosas juntos. La más relevante para mi y los investigadores de EsadeEcPol fue su apoyo instrumental desde CEOE a nuestra investigación experimental de tutorías de matemáticas para niños vulnerables. Como Ministra de trabajo tuvo que lidiar con un panorama muy difícil y junto con su querido Juan Pablo Riesgo pusieron en marcha reformas difíciles que facilitaron la recuperación del empleo. Se va demasiado pronto una mujer excepcional. Echaremos mucho de menos a personas como ella en política. Y en lo personal, la recordaré siempre con un enorme cariño. Todos mis pensamientos están con su familia y sus hijos de los que hablaba siempre con tanta alegría y orgullo. Todo mi cariño también para todas las víctimas de esta horrible tragedia.

Le Chonk is out of the bag 🐈 We're throwing Chonky Parties in our SF, NYC and London offices to celebrate the launch! You can expect some Champagne 🍾, food, good music, and to meet the team behind our latest release. 📍 SF: luma.com/viqhwgcw 📍 London: luma.com/xvuvl3mc 📍 NYC: luma.com/1qk99m3e Please sign up! Registration is mandatory.

Mistral is SO back: Inside Le Chonk, the West’s Biggest Open Model. My conversation with @PierreStock who runs the research organization at @MistralAI 0:00 Cold open 0:59 The Western open-weight comeback: Beam, Le Chonk, Muse, Nemotron 4:09 The US shutdown of Anthropic’s models - “AI is like electricity” 6:07 Why “Le Chonk” 7:56 “Mistral fell behind”: the comeback from the inside 10:29 Training on 3,800 GPUs in Mistral’s own datacenters 13:24 Meta keeps its best model closed, Mistral opens Chonk 15:40 Chonk vs Beam, Muse Spark Open and the Chinese models 18:44 Why Chinese labs are still at the frontier 21:08 Pierre’s story: FAIR, Yann LeCun, the “Mistral mafia” 25:14 From compressing networks to a trillion parameters 27:30 Mixture of experts explained: the hospital analogy 32:25 10-trillion-parameter models are coming 35:29 What Mistral borrowed from DeepSeek 36:46 One model, several reasoning modes 39:15 Native multimodality: satellites, manuals, computer use 41:56 Data, synthetic data and model collapse 46:32 Reinforcement learning as an exam 49:59 Where the efficiency comes from 52:18 Le Chonk is still training right now 53:39 What RL trains on: coding, cyber, finance, chips 58:13 Partial rewards and looping penalties 59:58 Did RL teach anything new? 1:03:16 The sandbox story: the model that pinged its trainer 1:07:44 Open weights vs cyber refusals 1:12:05 Timestral: time series on a CPU 1:14:18 AI doing AI research 1:16:22 Coming up: Mistral Large 5

Next Wednesday, 28 October, we’re bringing a group of founders together for an open conversation about Europe’s role in AI: what we’re building here, what’s holding us back, and what we can do about it. We’ll have a round table with @danipolymath from @SupersonikAI and Victor García from @NeuralTrustAI, followed by an open conversation where everyone can join in and ask questions. Then food and drinks. The event will take place at @Itnig, in Barcelona. If you’re interested, join via the link below! luma.com/zmqoj24c

FULL INTERVIEW: CEO of Mistral @arthurmensch on building or surrendering, European AI independence, funding gigawatt scale infrastructure in Europe, Le Chonk, AGI, the singularity, and existential risk. 00:24 - Build or Surrender 01:13 - Value in AI application, model and infra layers 03:57 - Mistral's full-stack strategy 05:56 - Open-source & European AI independence 08:31 - Funding gigawatt-scale infrastructure in Europe 11:40 - Overcoming pessimism & sovereign AI 13:26 - AGI, the Singularity, and existential risk concerns 14:50 - AI risks: agent security, control, and misuse 16:16 - The story behind the model name "Le Chonk"

🎙️ Introducing Voxtral 3 ! A new family of SOTA speech models: ASR: Mini & Small for batch transcription, plus real-time streaming TTS: expressive, multilingual speech, batch + streaming Voice agent SDK: low-latency conversations + tool calling Available via API or on your own infrastructure !

Prototyping is cheaper than ever, so I tried to imagine the simplest possible ClickHouse service. The result is a side project called ClickHamlet. You run: clickhouse client --host cloud.clickhamlet.com --secure --user <name> --password And that's it. No sign-up form, real ClickHouse in the cloud.

Introducing Voxtral 3: state-of-the-art Transcribe and Speech models powering Mistral’s new voice agents API. 🧵

1/ En abril me uní a Cursor para trabajar en Origin, su ‘git forge’, un poco antes de que SpaceXAI lo adquiriera. Cuando terminé en Unity/Plastic SCM en 2023 pensaba que no iba a volver a desarrollar un control de versiones,

Should Europe leave frontier AI to the Americans and the Chinese? The idea has gained ground in Europe. It is a fundamental mistake. Build or surrender: for now, the choice is still ours. Full article below.