InTowards AIbyLuhui Hu·6d agoJEPA and World Models: From Prediction to Physical IntelligenceWhy learning to predict in latent space may be more important than generating pixels — and what comes next for intelligent robots.
InCodeXbyLuhui Hu·Oct 18Recursive Self-ImprovementHow AI learns to improve itself — and what it means for world models and physical intelligence.
InThe CerebrusbyHari Harikrishnan·Sep 20Anthropic Flywheel: How Ecology, Psychology, and Technology Made UsA lens on human evolution resulting from a three-step cycle of change— ecological, psycho-physiological, and technological — boosted by…
InBetween the Lines InkbyMaria Diepgen·Sep 13Your Body Teaches Your Child Every DayWhy your Body Intelligence becomes your child’s greatest teacher.A response icon2A response icon2
InThe CerebrusbyHari Harikrishnan·Aug 14Persistence is What We Need: A Viability Imperative for AIScaling compute alone may keep true agency out of reach. Here is an alternative — make prediction subordinate to persistence.
InThe CerebrusbyHari Harikrishnan·Aug 11Physical Intelligence as Information-enabled System RegulationA Physics and Cybernetics-grounded Architecture for biological and non-biological Intelligence
Abhijeet Kadam·Jun 30Giving AI a Pair of Hands: How This Physical AI Journey BeganEvery innovation has a story, and this one began with a simple school project.
InData Science CollectivebySugun Segu·May 15Dissecting ACT: Action Chunking with TransformersCVAE, Action Chunking & Temporal Ensembles — Architecture to Jetson Thor
Nicholas Mitsakos·Feb 4AI in the Physical WorldAI is a monumental breakthrough, but it is merely a component — the “wheel” — that requires a robust physical system to become a functional…
