Thibault Quillard

@rupertmanfredi Interested as well :)

planning my day with Hermes & our Television gui

Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946. For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids. An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better. This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.

@zephyr_z9 Which one is not a shitco?

Makes me even more bullish about neoclouds

@zephyr_z9 Expect Meta to go there soon w/ their new app strategy

Are Chinese microdramas the next big consumer play 🤔🤔🤔

former 20 millions d'actifs à l'IA en 5 ans est exactement le genre d'annonce qui sonne brillant en commission parlementaire et qui produit 0 effet réel je dis bien 0 effet sur le terrain parce qu'on a oublié quelque chose de fondamental qui est que: l'IA n'est pas une compétence en soi c'est un amplificateur des compétences cognitives qui préexistent dans votre cerveau, savoir lire profondément, savoir raisonner par premiers principes savoir formuler une intention claire ou même savoir reconnaître un raisonnement bancal autant de muscles intellectuels qui se construisent uniquement par 20 ans de friction réelle avec des livres, ddes problèmes mathématiques exigeants et des projets concrets sur lesquels on échoue / itére avant de réussir je crois que le drame de cette annonce c'est qu'elle traite l'IA comme un nouveau excel qu'il faudrait apprendre à manipuler en 6 modules de formation alors alors que le vrai sujet c'est de reconstruire dans le système éducatif français les fondations cognitives qui permettent ensuite à un adulte d'utiliser l'IA comme un véritable allie démultiplicateur ce n’est que mon avis MAIS sachez que sans ces fondations vous obtenez exactement le contraire de l'effet recherché, des millions d'actifs qui délèguent leur pensée à la machine et perdent en 18 mois ce qu'il restait de leur autonomie cognitive, c'est l'inverse exact d'une politique souveraine c'est une politique de zombification industrielle financée par les impôts des contribuables et ce que cette annonce va concrètement produire dans les 24 mois est facile à prévoir, l'apparition massive d'une nouvelle génération de Brivael, ces opportunistes qui ne tweetent plus eux mêmes mais qui font tweeter leurs agents IA pendant qu'ils encaissent en vendant des formations et des calls à 999 dollars de l'heure parce que chaque fois que l'etat français sort une enveloppe budgétaire avec le mot innovation dedans c'est exactement le profil qui se rue dessus pour la capter on aura donc dépensé plusieurs milliards d'euros publics pour enrichir des grifteurs qui n'ont eux mêmes jamais rien construit dans leur vie pendant que les vrais bâtisseurs deeptech par ex continueront de galerer pour financer leur première levée je suis persuadé que ce dont la France a vraiment besoin n'est pas une formation IA pour 20 millions d'actifs, c'est une refonte profonde de l'école primaire et du collège pour reconstruire les capacités d'attention de lecture de raisonnement et de créativité de la jeunesse française c'est aussi un investissement massif dans la recherche fondamentale en mathématiques, en physique et en biologie, c'est un soutien acharné aux bâtisseurs qui font de la deeptech et du hardware en France plutôt qu'aux vendeurs de formation digitale mettez vous vous en tête que qu'à la fin des fins ce qui déterminera la position de la France en 2050 ce n'est pas le nombre d'adultes qui savent prompter ChatGPT, c'est plutôt le nombre de cerveaux capables de produire la pensée originale que la machine ne pourra jamais générer toute seule

Just hooked my agent to order a bottle of wine every time we sign a new deal at @daimo Sending USDC on @tempo through @mpp Merchant receive it in USD thanks to @stripe

We updated our 4 company values this week to keep up with how the company has changed. Here's what I shared with the team internally. I hope it could be helpful for other companies.

There is nothing about this that is amazing. It was predicted endlessly, and those predictions were dismissed by people citing the fact that big companies opposed these measures. They opposed them because while their relative position is stronger, the pie is now smaller.

España tiene claro el camino: electrificación, energías renovables y autonomía energética. Desengancharnos de los combustibles fósiles no es solo una apuesta climática: también es una apuesta por nuestra soberanía y nuestra competitividad. Y esa transformación necesita de más inversión en interconexiones y en descarbonización.

I'm a manager at @OpenAI, but with GPT-5.5 I'm a more effective IC than I've ever been. I can now write CUDA kernels like a pro. I can rely on it to run my research experiments. And we know how to make it much more powerful from here.

Les choses avancent, mais pas assez vite. Jamais assez vite. J’ai donc pris un engagement clair : accélérer, une bonne fois pour toutes, la réalisation de nos projets industriels sur l’ensemble du territoire. Nous avons montré au monde que nous étions capables de rebâtir Notre-Dame en cinq ans. C’est cet esprit que je veux retrouver aujourd’hui. Aujourd’hui, nous identifions 150 grands projets, 150 cathédrales industrielles, partout en France que nous allons faire avancer beaucoup plus rapidement. Cela représente 71 milliards d’euros d’investissement et plus de 32 000 emplois, au service de notre souveraineté. Et cela s’accompagne de très nombreuses simplifications pour faire mieux et plus vite. Nous accélérons.

@josiezayner Can I order/send DNA from Spain?

I am all for doing all the work to sequence your genome at home but my company The ODIN is now offering to sequence any animal genome(including humans) at 30x We give you basic data analysis and complete control of your data, nothing fancy for a good price.

Hard to unsee the future from demos like this

@ashebytes thanks for sharing ! Could you add swipe to change slide for Mobile use ? :D

for anyone interested, these are the slides I put together & livestream chat ashe.ai/slides/april-a… x.com/ashebytes/stat…

When I saw our team's evals of Kimi 2.6, I thought "ok, things are gonna get interesting now". This is the first open-weight model that plays like a top-class agentic model. Watching it go through ambiguous and meticulous chained tool work successfully puts it squarely in the wheelhouse of Opus 4.6. We're looking at an open weight model, but with much cheaper direct inference provider pricing. For a subclass of our eval set, it's outperforming GPT 5.2. We're about to undergo a gigantic industry shift. Open weight is no longer for those who fine tune, those who want on-prem. It's an actual, reliable option for it's quality/price/latency profile for difficult agentic work. It's not perfect. It's token hungry, relatively slow, and can get stuck in “thinking loops". But those are things we can engineer around. For value it is, and how it positions itself against major labs, this is a dramatic day for open weight models. We sprinted as a team and worked closely with @FireworksAI_HQ to get this to our customers on day 0. No one should wait to try out a change like this. Try it yourself and tell me where it's working for you.

I just finished reading palantir’s manifesto & I need you to understand what you’re actually looking at because this is the MOST important document the tech world has produced this year most people came away thinking «wow what a thoughtful essay about patriotism and technology »…I came away thinking this is the most elegant justification for corporate capture of the state apparatus ever written & I want to walk you through why krp opens with «silicon valley owes a moral debt to the country that made its rise possible » & frames the entire document as a call to civic duty, but read between the lines and what he’s actually saying is that the engineering elite should be embedded inside the defense and intelligence apparatus of the nation, he’s describing exactly what palantir has already done and dressing it up as patriotism «the question is not whether AI weapons will be built, it is who will build them and for what purpose »sounds like a warning but it’s actually a sales pitch, he’s telling every gov on earth that the choice is binary either you buy from us or your adversaries will build it without you, this is the oldest arms dealer rhetoric in history wrapped in SV vocabulary « hard power in this century will be built on software »is the key sentence of the entire manifesto because this is where karp reveals the real thesis, he’s saying whoever controls the software layer of national defense controls the nation itself & if you’ve been following my threads you know that palantir’s gotham and foundry platforms are already plugged into the intelligence feeds the satellite data, financial transactions & communications of dozens of govts worldwide through a single ontological knowledge graph that creates a technological dependency so deep that migrating away would mean rebuilding the entire institutional memory of the organization from scratch this is vendor lockin at the scale of nation states and I’m personally convinced it was designed this way from the beginning «we should applaud those who attempt to build where the market has failed to act » is karp defending palantir’s expansion into every domain the gov used to handle itself, policing immigration, military targeting intelligence analysis public health, everywhere the state retreats palantir advances and what was once a government function becomes a private service that the government can no longer perform without plantir’s permission and here’s what I think makes it even more concerning, these systems are increasingly autonomous meaning the AI layer is making targeting recommendations threat assessments & resource allocation decisions that humans inside gov are rubber stamping without fully understanding the underlying logic a bureaucrat inside the pentagon / DGSI sees a recommendation from the system & approves it because the system has been right 97% of the time and questioning it would require technical expertise that no one in the room has, this is algorithmic governance wearing the mask of human decision making «the atomic age is ending, a new era of deterrence built on ai is set to begin »is the MOST chilling sentence in the document because karp is explicitly saying that ai based deterrence will replace nuclear deterrence as the organizing principle of global power, and whoever builds that ai deterrence layer owns the 21st century the same way whoever built the bomb owned the 20th & he’s telling you plainly that palantir intends to be that builder «national service should be a universal duty » & « we should only fight the next war if everyone shares in the risk »sounds noble until you realize that he is proposing a system where citizens serve the state & the state is operationally dependent on palantir, the public bears the risk and palantir captures the value, soldiers fight wars planned by algorithms they can’t audit built by a company they can’t vote out

@ashebytes @menhguin Love it ! Did you try @bumpbyamo by @amoamoamo ? They have the same video avatar style 🔥

consider hogwarts-maxxing your personal website

@jasonrgriffin Context at 1M is big enough to be usable IMO It's just you want the code to do what deterministic code does well, and you want the skill to do what LLMs do well

@simonlast Great trackrecord 🙏💪 do you have a projection for 2026? Tks for the notion push with ai agent making it accessible to a broad range of users

Some AI predictions for 2025: Coding agents that can do end-to-end software tasks become actually useful, reliably doing tasks that would take an engineer at least one hour. Adoption is slower than many expect because it requires a significant change in behavior (much more so than autocomplete). Models will become strongly superhuman at math, but it won’t have much of a direct effect. We’ll see extremely cheap and reliable small reasoning models. They’ll be very good at coding. This will unlock use-cases that involve a lot of fan-out. The next large scale-up in model size will succeed in all the labs, but it’s unclear to me whether they will actually be released. I think probably yes just because it would make the lab look like they are ahead, even if the cost tradeoff doesn’t make sense and the main trend is inference compute. The gap between top labs narrows further, with OpenAI, Anthropic, and Google very competitive. Google in particular seems to be catching up and I expect more convergence. Meta successfully replicates the reasoning paradigm, but remains behind the top labs and isn’t the best cost/performance tradeoff. Always-on ambient AI hardware devices hit the market, but are not widely adopted yet.