New paper: Tell Me What To Learn: Generalizing Neural Memory to be Controllable in Natural Language. We built a neural memory system that lets you tell an AI — in natural language — what to remember and what to ignore over time. 🧠👇

A Brief History of Intelligence by @maxsbennett is utterly wonderful in its clarity. A must read for anyone in both natural and artificial intelligence bookshop.org/p/books/a-brief…

So excited that this is happening. To my knowledge, this will be the most well funded neuroscience-inspired AGI initiative out there.

V cool paper

If only everyone in AI would read this book. amazon.com/Brief-History-Int…

My friend @AndySack wrote a cool book, some great frameworks for anyone working on AI applications or applying AI in a business setting amazon.com/AI-First-Playbook…

The Orthogonal Bet. It's a smart maneuver. But it's also the name of my friend Samuel Arbesman's latest podcast. Here are some sample episodes, (including one from yours truly) on a -different- take of one of the most important issues of our time: Embodied Intelligence. Congrats Sam! @drmichaellevin Understanding Embodied Intelligence (podcasts.apple.com/us/podcas…) @maxsbennett A brief history of Intelligence (podcasts.apple.com/us/podcas…) @tarinziyaee Artificial Life and Robotic Evolution (podcasts.apple.com/us/podcas…)

^ @maxsbennett

Remember Moravec's paradox? "Corner cases". "Long tails". You can play whack-a-mole with them. Or, you can confront their root causes directly. Our new paper "Evolution and The Knightian Blindspot of Machine Learning" argues we should do the latter.