
Just published a write-up on my blog about the cloud coding agent I've been building recently. There's a lot of demos and technical details in it! stanislas.blog/2026/02/netcl…

I like building stuff with code github.com/angristan • feeding models @MistralAI

Just published a write-up on my blog about the cloud coding agent I've been building recently. There's a lot of demos and technical details in it! stanislas.blog/2026/02/netcl…

J'ai testé vite fait sur l'analyse de vulnérabilité de l'app laboucle.run et je suis assez bluffé. Il a trouvé deux failles critiques sans avoir accès à tous mes services alors que j'avais donné le même exercice à Fable5.2 la semaine dernière et il n'avait rien trouvé
Mistral Vibe is now available on Desktop! Run agents locally or remotely, and start shipping straight from your MacOS or Windows device. You can try the Pro plan for free if you’re building in Europe. Coupon code: LECHONK (yes, we added more after we ran out yesterday). Use the link in 🧵:

If you missed it, we just added more coupons! Get it before it's too late 🐈
Mistral Vibe is now available on Desktop! Run agents locally or remotely, and start shipping straight from your MacOS or Windows device. You can try the Pro plan for free if you’re building in Europe. Coupon code: LECHONK (yes, we added more after we ran out yesterday). Use the link in 🧵:

Mistral Vibe is now available on Desktop! Run agents locally or remotely, and start shipping straight from your MacOS or Windows device. You can try the Pro plan for free if you’re building in Europe. Coupon code: LECHONK (yes, we added more after we ran out yesterday). Use the link in 🧵:


Team decided to have some fun and customize the install of Vibe desktop. Get Vibe Desktop at mistral.ai/products/vibe

Try out the vibe desktop app on macOS and Windows. We are particularly proud of Mistral Large 4 capabilities for Cyber Defense.
Download the desktop app here: mistral.ai/products/vibe/

The vibe desktop app is here: vibe work, vibe code, local and cloud agents, same shared harness as vibe cli 💖
Oh, and Vibe desktop is here! Hand Vibe several tasks at once, watch the agents work side by side, then review their diffs and track pull requests in the same window.

3 months of free vibe 👀
To celebrate Mistral Vibe + Le Chonk becoming the world’s best open model-harness duo, we’re making Vibe free in Europe for the first 3 months. (Yes, it’s actually better than GLM / DeepSeek at coding). Use code LECHONK via link in 🧵. Limited time, first-come, first-served.

Oh, and Vibe desktop is here! Hand Vibe several tasks at once, watch the agents work side by side, then review their diffs and track pull requests in the same window.

To celebrate Mistral Vibe + Le Chonk becoming the world’s best open model-harness duo, we’re making Vibe free in Europe for the first 3 months. (Yes, it’s actually better than GLM / DeepSeek at coding). Use code LECHONK via link in 🧵. Limited time, first-come, first-served.

Vibe is AWESOME on chonky codebases and fixing vulnerabilities with precision. Vibe with Le Chonk score highest among open models on FrontierCode and Terminal-Bench, and take on security work other agents refuse.

I can’t believe this is the sort of society we are heading towards just because people are afraid to call it out for what it is
Weihnachtsmarkt in Berlin. Dieser Wahnsinn muss aufhören.

AI isnt stealing ur job, its exposing how fucking useless u were at it lmao. imagine spending years turning 20 minutes of actual work into an 8 hour shift just for a machine to expose the scam in 30 seconds. now suddenly automation is an ethical crisis. turns out ur most valuable skill was convincing someone to pay u six figures for doing fuck all

it's kinda sad, but i think it's well overdue that we start having a serious conversation about this and i don't think people understand how deep the problem actually is we're witnessing one of the most extraordinary engineering achievements in human history. we're literally catching skyscrapers falling from the sky and there are people whose first reaction is hate and resentment the problem is that their entire understanding of the world is built on assumptions that are fundamentally incompatible with progress. you could explain them rockets. you could explain the economics and how this could improve their lives. you can explain how technological progress works, you could even hold their hand explaining how this is the fundamental step to make our species multiplanetary. it doesn't matter. they see everything as a zero-sum game under the lens of envy. if someone achieves something extraordinary, someone else must have suffered for it. since everyone is equal, the only way to achieve this is by stealing and oppressing. if we spend money exploring space, that money must have been taken from the poor (even if the poors take much more than they give). if someone they don't like builds something incredible, the achievement becomes something to condemn. literally zero understanding of how anything works you cant even begin to reason with someone like that to change their mind, you would have to dismantle their entire understanding of reality and rebuild it piece by piece (that's assuming they're even willing to question it, which they are not) what makes everything worse is that these people vote and their vote is worth as anyone else's. if we're serious about progressing as a species, sooner or later we're going to have to figure out how to solve the problem.
If you ever want to Black Pill, Just go to Instagram and read the comments. I don’t know how humanity survives the Left.

Still good performance for ML4 on the new coding agent index!
Announcing Artificial Analysis Coding Agent Index v2.1, upgrading FrontierCode and Terminal-Bench. Claude Fable 5.2 and Claude Opus 5.5 share the lead at 58 The Coding Agent Index measures models in agent harnesses; v2.1 of the Index upgrades FrontierCode from 1.1 to 2.0 and Terminal-Bench from 4.0 to 4.1. Upgraded: ➤ FrontierCode-AA 2.0 (replacing FrontierCode v1.1): our implementation of Cognition's FrontierCode 2.0, comprising 132 real-world coding tasks graded on whether a repository maintainer would merge the agent's change ➤ Terminal-Bench 4.1 (replacing Terminal-Bench 4.0): the same 66 tasks, scored with updated verifiers that improve reliability and resistance to reward hacking Staying in the Index: SWE-Atlas-QnA-AA Key results: ➤ Claude Fable 5.2 and Claude Opus 5.5 share the lead at 58: Claude Fable 5.2 scores 58 at both xhigh and high in Claude Code. Claude Opus 5.5 (max) scores 58 in Claude Code and, paired with GPT-6 Luna High, in Cognition's multi-model harness Devin Fusion CLI. Claude Fable 5.2 (max) scores 56 as more of its attempts hit task time limits. Claude Sonnet 5.5 (max, 56) in Claude Code and Gemini 4 Carbon (high, 55) in Antigravity CLI follow ➤ Index scores are lower in v2.1, driven by the upgrade to FrontierCode-AA 2.0: models score lower by 5 points on average. Codex with GPT-6.1 Terra (max) is the only configuration which didn’t fall between versions, sitting within 1 point of its prior score ➤ Cost per task varies ~30x among configurations scoring 48+: from $1.51 for Codex with GPT-6.1 Sol (max, 52) to $44.22 for Claude Code with Claude Fable 5.2 (max, 56)