FINALLYYYYY !!!!!!! BUT PLEASE @SebLecornu, DO NOT DISTRIBUTE THESE COURSES TO FREAKING PROMPT TRAINERS WHO DISCOVERED AI LAST WEEK. I propose for example to offer AI students in master's, bachelor's or doctorate the possibility to go do 1 hour
AI
-
Jensen Huang compares an AI agent to a worker in a workshop
By
–
Jensen Huang’s simplest explanation of an AI agent: a worker in a workshop.
— NVIDIA (@nvidia) 19 juin 2026
The model thinks. The harness gives it form. Tools and skills let it act. And the runtime gives the agent a place to get work done. pic.twitter.com/mrahjDICRQJensen Huang's simplest explanation of an AI agent: a worker in a workshop. The model thinks. The harness gives it shape. Tools and skills allow it to act. And the execution environment gives the agent a place to perform its work.
-
Google DeepMind AI Co-Scientist accelerates scientific discovery
By
–
Google’s #AI Co-Scientist Aims to Accelerate Scientific Discovery
— Ronald van Loon (@Ronald_vanLoon) 19 juin 2026
by @GoogleDeepMind#AIAgents #GenerativeAI #ArtificialIntelligence #MachineLearning pic.twitter.com/cynavBIWP6Google’s #AI Co-Scientist Aims to Accelerate Scientific Discovery
by @GoogleDeepMind #AIAgents #GenerativeAI #ArtificialIntelligence #MachineLearning -
AI cascade detects fruit mislabeling at self-checkout counters
By
–
The "banana trick" at self-checkout works by ringing up a £4 avocado as a 40p onion The fix is a simple AI cascade. @Ultralytics YOLO detects each item, a classifier (EfficientNet, MobileNetV4, etc) tells you what it actually is, and then cross-checks against the POS scan
-
Up to 95% Token Reduction Without Code Changes
By
–
UP TO 95% TOKEN REDUCTION WITHOUT CHANGING THE CODE A Netflix engineer just open-sourced Headroom, and it’s one of the smartest ways I’ve seen to cut LLM costs. It wraps Cursor or Claude in a local proxy to compress your payload before it hits the LLM: → Intelligently shrinks
-
The true primitive: authority protected by proof
By
–
Exactly. 'Impossible to unprison' is the wrong goal. The true primitive is authority protected by proof. GoalOS assumes that every model, agent, and evaluator can fail: generate → verify → contest → canary → monitor → cancel The capacity
-
Founders test models on unsolved problems: completing work vs making work
By
–
Very good point. See something similar talking to founders with very specific hard problems to solve – they test each new model for the ability to churn through their list of unsolved problems. In both cases, completing work v making work.
-
Research on AI undermining versus supporting thinking and learning
By
–
I talk about the research on when AI undermines, versus supporting, thinking and learning here:
-
Anthropic CEO observes a $62k YouTube strategy leak
By
–
El CEO de Anthropic viendo que un tipo usó Claude para construir un sistema de YouTube de +$62,000/mes y acaba de filtrar toda la estrategia GRATIS https://t.co/pXx0sWiGbZ pic.twitter.com/Yj0FgC6PzP
— Nico (@nicos_ai) 19 juin 2026Anthropic CEO seeing that a guy used Claude to build a YouTube system making over $62,000/month and just leaked the entire strategy for FREE
