Louis van Proosdij
25K posts
@LvP
Entrepreneur, intrapreneur, new-tech/digital senior exec, digital, strategy, startup mentor, modest "Expert Generalist", dots connector
Terence Tao has won every award mathematics can give a human being. Fields Medal. Breakthrough Prize. MacArthur Genius Grant. He is widely regarded as the greatest living mathematician. Not one of. The greatest. He just said something that should terrify every university on Earth. Tao: “We live in a particularly unpredictable era. I think things that we’ve taken for granted for centuries may not hold anymore.” Not years. Not decades. Centuries. The assumptions governing who gets to contribute to knowledge have been in place longer than most nations have existed. Tao just told you those assumptions are dissolving. Tao: “The way we do everything, not just mathematics, will change.” This is not a man who deals in hyperbole. He builds arguments the way he builds proofs. Piece by piece. Nothing unverified. When he says everything, he means everything. Tao: “In math, you previously had to basically go through years and years of education, be a math PhD before you could contribute to the frontier of math research.” That was the contract. You give a decade of your life to an institution. You grind through coursework, committees, dissertation reviews, postdoc rotations. Then maybe you get to touch the boundary of what’s known. The entire system was built on that bottleneck. Time was the gate. Credentials were the key. Tao: “Now it’s quite possible at the high school level that you could get involved in a math project and actually make a real contribution because of all these AI tools.” A high schooler. Contributing to frontier mathematics. The same frontier that used to require a decade of institutional obedience to even approach. He said this about math. He already told you this applies to everything. AI didn’t just speed up the path. It removed the path entirely. The university sold you a ten-year toll road. AI just paved around it overnight. The toll booth operators haven’t realized yet that no one’s coming. Tao: “In many ways, I would prefer the much more boring, quiet era where things are much the same as they were ten years ago, 20 years ago.” This is the line that should haunt you. The smartest mathematician on the planet would rather this wasn’t happening. He is not selling this. He is not positioning himself for a funding round. The acceleration is so violent that even the mind best equipped to process it would prefer it stopped. If Tao is uncomfortable, you should be paying very close attention to your own assumptions about what’s coming. Tao: “The things that you study, some of them may become obsolete or revolutionized, but some things will be retained.” That word “some” is doing enormous work in that sentence. It means the rest won’t be. Entire fields that people spent their careers building will collapse. Not slowly. Not politely. And Tao is telling you he can’t predict which ones survive. Tao: “You should be open to very, very different ways of doing science, some of which don’t exist yet.” Most people will scroll past this. It’s the most important line in the entire clip. He’s not saying learn new tools. He’s not saying adapt your workflow. He’s saying the methods themselves haven’t been invented yet. The frameworks don’t exist. You cannot prepare for what hasn’t been created. You can only build the kind of mind that doesn’t break when the ground shifts beneath it. Tao: “It’s a scary time, but also very exciting.” He said scary first. Every tech founder says exciting first and mentions risk as a footnote. Tao reversed it. When the most brilliant mind of a generation leads with fear and follows with possibility, that is not optimism. That is a man telling you the truth about what’s coming while still choosing to walk toward it. The people who survive the next decade won’t be the ones with the best credentials. They’ll be the ones who stopped mourning the world that was and started building for the one that doesn’t exist yet.
Scrum a ruiné une génération de devs. Les solutions clés en main d'agents IA vont ruiner la suivante. Pour exactement la même raison. Le pattern : → un problème fondamentalement contextuel → quelqu'un vend une méthode générique → tout le monde l'applique sans réfléchir → ça marche pas → on blame l'exécution au lieu de la méthode Les meilleures boîtes n'ont jamais appliqué Scrum by the book. Elles ont construit leur propre façon de fonctionner, adaptée à leur contexte, leur culture, leur stade de croissance. Et elles l'ont fait évoluer en permanence. Avec les agents, c'est exactement pareil. Chaque boîte a une archi différente, une stack différente, des process différents. Brancher un agent générique sur tout ça et espérer que ça marche, c'est la même illusion que de coller Scrum sur une équipe et attendre que la productivité explose. Nous en ce moment on construit Argil Forge. C'est notre harness interne. On a wrappé Claude Code avec toute notre infra : → Playwright pour le browser → Tauri en desktop app → un backend qui orchestre nos services → des worktrees git dédiés avec des ports séparés par service → des callbacks d'auth auto-générés Résultat : un agent qui a le même setup qu'un ingé onboardé chez nous. Pas un outil générique posé dans un terminal qui bosse en silo. On commence par le produit, mais le plan c'est d'étendre à toute la boîte : ops, marketing, ads, SEO, legal. → Ship une feature → la landing se crée → Nouveau use case → la campagne se lance Et tout ce qui a vraiment de la valeur là-dedans, c'est ce qui est propre à nous : → Le context graph de la boîte → La mémoire globale de l'orchestrateur level 0, qui porte la vision d'ensemble → La mémoire de chaque agent spécialisé, qui accumule l'expertise de son domaine → Une decision trace commune entre toutes les entités, pour que chaque décision prise par un agent soit visible et traçable par tous les autres C'est pas générique. C'est le reflet de comment notre boîte fonctionne. Et c'est un truc qu'on va owner et maintenir de manière profonde, parce que c'est ça qui permet à terme de faire tourner la boîte quasiment en autopilote. L'objectif dans quelques mois chez Argil, c'est qu'il reste deux métiers : → Maintenir et faire évoluer ces systèmes → Prendre les décisions qui reposent sur de la taste, et corriger celles de l'agent quand il se trompe Et quand l'agent a suggéré 10 décisions d'affilée validées par l'humain, on considère que ce domaine commence à pouvoir tourner en autopilote. Dans le futur, les meilleures boîtes owneront leur agent infrastructure comme elles ownent leur culture et leur orga. Ça se construit, ça se maintient, ça évolue. Les autres achèteront le "Scrum des agents" et se demanderont pourquoi rien ne marche.
HAPPY BIRTHDAY! Barron Trump, the son of President Donald Trump and First Lady Melania Trump, turns 20 today.
Whether the next president is a Republican or a Democrat, can we have someone with at least a little class next time? Because this is embarrassing.
BREAKING: “Beautiful Army blonde” with over 1 million followers & frequently pictured with Trump revealed to be AI generated.
🇭🇺🇺🇸 Thank you for your support, Mr. President @realDonaldTrump!
When even London’s mayor @SadiqKhan @MayorofLondon says the UK cannot afford to drift between Trump’s tariffs, geopolitical shocks and the long shadow of Brexit, it’s time to listen — in Britain and beyond. Sadiq Khan’s message is clear: the illusion of sovereignty has come at a real cost, with weaker growth and less influence. In an unstable world, standing alone is not strength. Rebuilding a close partnership with Europe is not nostalgia — it is a strategic necessity.
