Tag: Learning
-
AI’s Relentless Proactivity
relentless proactivity is the defining trait of this new generation of Mythos-class models. If you set them a goal and give them a way to get there, even inadvertently, they will figure it out. — Read on simonwillison.net/2026/Jul/22/openai-cyberattack/
-
Opinion | The Problem With Google’s A.I. Overview – The New York Times
When an A.I. answers your search query in three seconds, the window closes before curiosity can deepen. You got what you came for, but you also lost what would have turned curiosity into learning: the adjacent article you might have read, the resulting tangent you might have followed, the connection between two ideas with no…
-
Why The Best Player Alive Barely Runs – Chunking
They’re chunking positions of people and angles of legs and spins of balls in order to understand immediately what’s going on and what might happen next. The more patterns you absorb in any domain, the less effort it takes to read what’s happening and to predict what’s coming next. So when Messi walks, he’s not…
-
IA : « Réhabilitons le brouillon, la rature, l’essai, l’erreur, non comme signes d’échec, mais comme conditions de la pensée »
La perfection algorithmique exerce ici une séduction dangereuse. Elle promet de supprimer la faute, l’hésitation, la lenteur, la honte de mal formuler. Mais une société qui ne tolère plus l’erreur finit par ne plus tolérer l’humain. L’enfant qui apprend, l’étudiant qui cherche, le chercheur qui doute, le citoyen qui écrit maladroitement une idée juste : tous…
-
Hoard things you know how to do – Agentic Engineering Patterns – Simon Willison’s Weblog
The key idea here is that coding agents mean we only ever need to figure out a useful trick once. If that trick is then documented somewhere with a working code example our agents can consult that example and use it to solve any similar shaped project in the future. — Read on simonwillison.net/guides/agentic-engineering-patterns/hoard-things-you-know-how-to-do/
-
Linear walkthroughs – Agentic Engineering Patterns
Frontier models with the right agent harness can construct a detailed walkthrough to help you understand how code works. — Read on simonwillison.net/guides/agentic-engineering-patterns/linear-walkthroughs/
-
Opinion | What a Baby’s Laugh Actually Tells Us – The New York Times
The signature belly laughs seen in the video above are involuntary, bursting forth during genuine, uncontrollable amusement. This type of laughter is driven by the brain’s limbic system, structures crucial for emotion, memory and motivation. But by 6 months, our lab has found, infants can intentionally produce a laugh. This ability comes not from the…
-
Claude Code starter projects
jasmine sun on X: “@TheStalwart yeah I made a list of Claude Code starter projects and “visualize a random CSV” is #1 or you can export and visualize your own data — iMessage, Goodreads, etc! https://t.co/UAnJMsCK1q” / X — Read on x.com/jasminewsun/status/2015495431737029085
-
Blind to the solutions we were never taught to see
Programmers are trained to see everything as a software-shaped problem: if you do a task three times, you should probably automate it with a script. Rename every IMG_*.jpg file from the last week to hawaii2025_*.jpg, they tell their terminal, while the rest of us painfully click and copy-paste. We are blind to the solutions we…