@mattturck

VC at @FirstMarkCap. Host: MAD Podcast; Organizer: Data Driven NYC, Author: MAD Landscape.

 New York
mattturck.com
Joined February 2011
  • Tweets20.8K
  • Following2.5K
  • Followers146K
2.5KMedia
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
9
38
3
252
19,770
Woah they actually made a movie about the GPU shortage
6
15
1,488
Matt Turck retweeted
Florian Douetteau
@fdouetteau19h
More context on today’s release, Headless. Every company has developers using Claude Code or Cursor to build data pipelines, ML models and custom agents. It’s fast and it’s awesome. But then: where does that work go? Often, just a Python file on a laptop. That’s problematic.
3
3
15
1,558
Coding agents made AI building fast. Now the bottleneck is understanding what got built. @dataiku Headless lets developers build data pipelines, ML models, and custom agents from Claude Code or Codex. The output is visual, not 2k lines of code. tinyurl.com/2f2zy3dc #DataikuHeadless
5
4
1
19
1,341
Matt Turck retweeted
LM Studio
@lmstudio23h
LFM3 is live in LM Studio! Fits in about 1 GB of memory. Model page: lmstudio.ai/models/lfm3

LFM3

LFM3-VL-2B is Liquid AI’s 1.92B vision-language model for on-device image and video understanding, with reasoning, tool use, and a 128K-token context window.

lmstudio.ai
Liquid AI
@liquidai23h
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence. > 6 model instances, across text, audio, vision, and omni modalities, generative and decision models. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding > 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages. > 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language. > 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input. > fine-tune on 1 GPU. > a new tri-hybrid architecture > architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒 Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device. 1/ 🧵
33
88
4
1,025
51,459
Matt Turck retweeted
Gradium
@GradiumAI22h
80K+ downloads and counting for Phonon. Here are @neilzegh and @lmazare answering some questions from the community. gradium.link/phonon
7
10
2
43
3,100
Matt Turck retweeted
Ben Lerner
@ben_lern22h
Today, we're announcing Nitro: a formally verified agentic system for compiling your most complex SQL workloads into Rust. Our design partners are running millions of dollars of critical workloads through this and seeing 10x faster, 3x cheaper ETL. Reach out to try it.

10x Faster, 3x Cheaper ETL

3
6
2
57
10,559
Usual startup advice: too many co-founders is bad!!! Too much dilution! Founder friction risk! Analysis paralysis! Also: * Anthropic: 7 co-founders * OpenAI: 11 co-founders * Databricks: 7 co-founders
61
12
6
261
20,138
Anthropic S-1 highlights video ONE SHOT (just one prompt, no editing) using @synthesiaIO Syren
🤖 Made with AI
5
6
43
7,248
Matt Turck retweeted
Ivan Burazin
@ivanburazinOct 26
Times Square looks good.
13
14
7
253
12,576
Matt Turck retweeted
Gradium
@GradiumAIOct 26
We’re releasing a set of free tools powered by Phonon. The first is Gradium Read and Transcribe, a Chrome extension that reads any webpage out loud. It runs on Phonon, our 40M-parameter multilingual model, and handles everyday text. For hard cases like acronyms, phone numbers, and dates, you can switch to the Gradium API, which handles them by default. gradium.link/read
3
11
64
3,248
Particularly evil of Anthropic to file their S-1 on a Monday morning when all VCs are stuck in Monday team and partner meetings, making us unable to provide immediate #thoughtleadership with those S-1 teardowns that everyone is dying to read
37
8
2
561
19,721
Good neighborhood
Odin
@JoinOdinOct 25
The Top 50 VC Firms Ranked by the long-term value created by primary investments per dollar at risk. The top 10: 1. @shv 2. @foundersfund 3. @iaventures 4. @gigafund 5. @Lowercase 6. @pelion_vp 7. @usv 8. @benchmark 9. @BessemerVP 10. @DFJvc Read more: blog.joinodin.com/p/the-top-…
4
7
1
117
29,373
Matt Turck retweeted
Karanbir Singh
@kbsinghOct 23
Big fan of the @MistralAI potential and what they can enable for Sovereign AI
Matt Turck
@mattturckOct 22
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
1
3
1,593
Matt Turck retweeted
Ben Lerner
@ben_lernOct 23
"Open source is dead and I killed it" - our CTO @kuza55 wrote up how he saved us $100k by self-hosting Grafana. With, basically, two days of work. OSS businesses have some soul-searching to do in the age of AI. Read more here: espresso.ai/post/self-hosted…

Open Source is Dead and I Killed It: How We Self-Hosted Grafana

Self-hosting Loki, Prometheus, and Tempo cut our monitoring bill by $6.5k a month in about 2 days of engineering.

espresso.ai
4
2
1
6
1,617
Matt Turck retweeted
Rashad
@corporayshidOct 23
i finally got on cheeky pint
2
2
30
4,164
Matt Turck retweeted
Tina He
@fkpxlsOct 23
More Daytona @PaceCap If the Daytona team were an LLM, people would write papers on how to train them. It’s only fitting that so much RL is now running on their computers, when the founders themselves are RL machines.
1
7
1
44
6,531
Matt Turck retweeted
Max Mynter
@MaxMynterOct 23
Closing Slack for a small twitter sesh and seing @PierreStock wasn't what I expected, but it's a great watch if you wanna know what we're on about @MistralAI (he knows what he's talking about)
Matt Turck
@mattturckOct 22
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
1
20
2,140
Matt Turck retweeted
Angela Liu
@angeliuuuuuOct 23
some personal news: i’ve joined @FirstMarkCap, where I'll be working with exceptional founders and leading the growth investing strategy alongside @abimbhet and @vverawang. FirstMark has backed some of my favorite companies of this cycle: @vercel, @ClickHouseDB, @WeAreLegora, @tryramp, @modal, @ro, @function, to name just a few, and i’m grateful to join a team I've long admired: @rick, @amishjani, @mattturck, @adamn, @DavidWaltcher, and the rest of the FirstMark family!
40
3
5
187
32,889
Matt Turck retweeted
LM Studio
@lmstudioOct 22
Introducing: Read Aloud 🔊 Powered by a tiny local TTS model, the excellent Phonon by @GradiumAI.
19
30
6
258
16,362
This conversation with the very insightful @PierreStock of @MistralAI is also available on Spotify, Apple Podcasts and here on YouTube: youtu.be/nxAoCPsGfe8?si=o0ks…

Mistral Is SO Back: Inside Le Chonk, the West’s Biggest Open Model

Mistral AI is back in the frontier AI race. Mistral Large 4, nickna...

youtube.com
1
6
1,447
Load more