The Tragedies of Reality Are Coming for You
In 2023, I was at an ML conference. The night was young, the drinks were flowing, and the topic turned to a question: “if you could take any machine learning subfield, and give all their resources to a different one, what are you killing and what are you boosting?”
In 2023, I was at an ML conference. The night was young, the drinks were flowing, and the topic turned to a question: “if you could take any machine learning subfield, and give all their resources to a different one, what are you killing and what are you boosting?” I don’t remember what they boosted, but one person said they’d kill robotics. When I pushed them on it, they said robotics progress is too slow and nothing happens relative to everything else. I think they were correct that robotics progress was slower than software-only machine learning subfields. But I would also add two…
saved by
related reading
- All Roads Lead to Robotics | Eric Jangevjang.com
- Many Small Steps for Robots, One Giant Leap for Mankindnotboring.co
- As Rocks May Think | Eric Jangevjang.com
- GitHub - adam-maj/robotics: A deep dive on the history of robotics and the future of humanoidsgithub.com
- Sporks of AGIsergeylevine.substack.com
- The Final Offshoringfinaloffshoring.com
- State of Robot Learning, December 2025vedder.io
- Software and Hardware for General Robotsblog.evjang.com
- Android Dreamsandroid-dreams.ai
- 14 Reasons Robotics is Hardsecondthoughts.ai
- Andrej Karpathy — AGI is still a decade awaydwarkesh.com
- How Can We Make Robotics More like Generative Modeling? | Eric Jangevjang.com