Umair Akbar

ml/ai

[Karine Hsu @karine_hsu]

wow sick new billboards from openai

[Anthropic @AnthropicAI]

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.

The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.

Access to all other Claude models is not affected.

We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.

Read our full statement: https://t.co/bwn0sximKZ

[Andrej Karpathy @karpathy]

This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November), peaking especially for long problem-solving sessions on very difficult problems. You can give it a lot more ambitious tasks than what you're used to, the model "gets it" and it will just go, and it's never felt this tempting to stop looking at the code at all (but don't do this in prod!). The model still has quirks that people will run into and the safeguards are configured to be a little too trigger happy for launch, which can hopefully be tuned over time.

I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing substantially. You can ask for anything - explainers, visualizers, dashboards, bespoke single-use apps (e.g. a full wandb that is hyper-specific just for your project), you can 10X your test suite, auto-optimize code, run giant research projects with custom HTML for the results, anything! "Free your mind" (Matrix ref). Really looking forward to all the things people build!

[Claude @claudeai]

Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.

Its capabilities exceed those of any model we’ve ever made generally available.

[Daksh Gupta @dakshgup]

boris cherny goes on a podcast every three months and says something like “i’ve stopped breathing now i just wrote a breath.md” and the next day everyone in sf stops breathing

[Anthropic @AnthropicAI]

Anthropic has confidentially submitted a draft S-1 registration statement to the Securities and Exchange Commission.

Pending completion of SEC review, this gives us the option to pursue an initial public offering.

Read more: https://t.co/onGZAhRLvD

[Andrej Karpathy @karpathy]

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

[reed @reed]

watch @waymo dodge a human pulling a blind unprotected turn

[Lamborghini @Lamborghini]

We are proud to introduce Lamborghini Fenomeno Roadster: the most powerful open-top ever created by Lamborghini. Limited to 15 units, powered by an iconic 1080 CV naturally aspirated V12 hybrid HPEV.

Engineered with an aerospace-inspired carbon fiber monofuselage, advanced active aerodynamics, and capable of 0–100 km/h in just 2.4 seconds, Fenomeno Roadster delivers pure performance with uncompromising driving emotion.

A V12 symphony.

#Lamborghini

[DeepSeek @deepseek_ai]

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.

🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.

Try it now at https://t.co/GCdiMzk1Dl via Expert Mode / Instant Mode. API is updated & available today!

📄 Tech Report: https://t.co/drlDrxkYtp
🤗 Open Weights: https://t.co/T13Y8i7SDM

1/n

[Brivael Le Pogam @brivael]

Aujourd'hui grosse discussion avec mes ingés (chez Argil) sur pourquoi Elon a viré le LIDAR de ses voitures autonomes. Choix radical, moqué pendant des années, et comme d'hab il avait raison depuis le début.

Le LIDAR c'est un laser qui balaye l'environnement et crache un nuage de points 3D. Sur le papier tu obtiens la géométrie exacte du monde. Dans la vraie vie c'est une verrue technologique collée sur le toit parce qu'on sait pas faire mieux avec la vision seule.

Problème numéro un : ça rajoute une modalité dans le training du modèle. Ton réseau doit apprendre à fusionner vision + lidar + radar + ultrasons. Chaque capteur en plus c'est une source de désaccord à arbitrer, pas une source d'info supplémentaire. Sensor fusion artisanale = dette technique permanente.

Problème numéro deux, la bitter lesson de Rich Sutton : scaler le compute sur une seule modalité bat systématiquement les architectures bricolées à la main. Tesla a dropé le radar, puis les ultrasons, est passé full end-to-end vision. Leur courbe sur les edge cases s'est accélérée APRÈS, pas avant. Waymo fait l'inverse et reste stuck en ops géofencée.

Problème numéro trois, le plus fondamental : le LIDAR voit la géométrie, pas la sémantique. Il sait qu'il y a un truc, pas ce que c'est ni ce que ça va faire. Les derniers 9 de fiabilité sont des problèmes de cognition, pas de perception brute. Un capteur de plus résout rien, il ajoute du bruit.

Sébastien Loeb balance une 208 T16 à 180 dans un chemin boueux corse sous la pluie avec zéro LIDAR. Deux yeux, un cerveau. L'évolution a donné des yeux aux prédateurs pendant 500 millions d'années, pas des lasers. Il y a une raison.

Le LIDAR c'est l'équivalent du marxisme appliqué à l'économie. Une solution planifiée, centralisée, qui prétend modéliser explicitement ce qui doit émerger d'un système distribué et adaptatif. Tu remplaces l'intelligence par de la mesure, la compréhension par de la donnée, l'émergence par le contrôle. Ça rassure les ingénieurs qui veulent tout spécifier en amont, exactement comme la planif rassurait les économistes soviétiques. Et ça échoue pour les mêmes raisons : la réalité est trop riche pour être capturée par un capteur, comme elle est trop riche pour être capturée par un plan quinquennal.

La vraie intelligence, celle de Hayek comme celle de Tesla, c'est de faire confiance à un système qui apprend de l'expérience plutôt que de tout pré-encoder. L'élégance d'une solution c'est son rapport signal sur complexité. Le LIDAR explose le dénominateur.

Défendre le LIDAR en 2026 c'est préférer empiler des hacks plutôt que résoudre le vrai problème. C'est de la feignasserie intellectuelle maquillée en rigueur d'ingénieur. Les mêmes gens qui défendaient les systèmes experts en 2012 contre le deep learning. Ils finiront pareil.

Never bet against end-to-end. Never bet against la simplicité. Never bet against Elon.

[Tesla North America @tesla_na]

Tesla Insurance update

With the latest version of Safety Score (v3.0), every mile you drive with FSD Supervised enabled will receive a score of 100.

This allows you to maintain a higher average safety score over time, resulting in lower monthly insurance premiums.

Applies to new policies in Indiana, Tennessee, Texas, Arizona, Virginia & Illinois

[Amazon Web Services @awscloud]

Announcing Amazon S3 Files.

The first and only cloud object store with fully-featured, high-performance file system access.

Learn more here. https://t.co/rNuWa5Rsi2

[gabriel @gabriel1]

https://t.co/iwuydUvF5A

[OpenAI @OpenAI]

5.4 sooner than you Think.

[Alex Wa @_djdumpling]

new blog! What methodologies do labs use to train frontier models?

The blog distills 7 open-weight model reports from frontier labs, covering architecture, stability, optimizers, data curation, pre/mid/post-training + RL, and behaviors/safety

https://t.co/88heRH4TcO

[Oliver Prompts @oliviscusAI]

ElevenLabs just lost its moat 🤯

Someone just dropped Voicebox, and it clones any voice from just a 3-second audio clip, running 100% locally on your machine.

100% Open Source

[Google Gemini @GeminiApp]

Gemini 3.1 Pro is here: A smarter model for your most complex tasks.

Building on the Gemini 3 series, 3.1 Pro is a step forward in reasoning. It's designed for tasks where a simple answer isn’t enough, taking advanced reasoning and making it useful for your hardest challenges.🧵

[Alex Albert @alexalbert__]

Sonnet 4.6 is here. It's our most capable Sonnet model by far, approaching Opus-class capabilities in many areas.

Very excited for folks to try this one out. The performance jump over Sonnet 4.5 (which was released just over four months ago) is quite insane.

[Claude @claudeai]

Introducing Claude Opus 4.6. Our smartest model got an upgrade.

Opus 4.6 plans more carefully, sustains agentic tasks for longer, operates reliably in massive codebases, and catches its own mistakes.

It’s also our first Opus-class model with 1M token context in beta.