Published “Smooth Exponentials for Robotics”
Eric shared predictions about robotics progress, looking at how more builders, broader training data and stronger AI models could shape the next few years.
Eric Jang writes about AI and robotics, creates machine learning tutorials, and is the author of AI is Good for You and a former VP of AI at 1X.
At 100 nudges, fangea sends Eric’s team an official invite.
🔒 Star boosts open once Eric claims this page.
Eric’s happenings, collected from public sources, plus their backstory and Starbook, where fans write.
Eric shared predictions about robotics progress, looking at how more builders, broader training data and stronger AI models could shape the next few years.
At the chalkboard with Dwarkesh Patel, Eric explained Go, tree search, neural networks and self-play. They also discussed what coding agents can already do for AI research.
dwarkesh.comWatch or read the conversationEric added AutoGo to their collection of tutorials. The dedicated tutorial page uses a wide canvas and asks mobile readers to rotate their device to landscape.
In “As Rocks May Think,” Eric described rebuilding AlphaGo with Claude and setting up a workflow that runs experiments, saves results and suggests what to investigate next.
Eric reflected on joining the company in 2022 and helping grow its AI work. They planned to spend the next few months learning, rebuilding deep learning papers and gaining a fresh perspective.
Follow Eric’s essays, tutorials and notes on AI and robotics.
Fans built this page. Claim it to connect with them and unlock Star boosts, or request a takedown if you’d rather it not exist.