How do We Quantify Progress in Robotics? - by Chris Paxton
itcanthink.substack.com · 1,955 words · saved by 1 readers
Determining which robotics advances are real can be incredibly difficult
How do We Quantify Progress in Robotics? Determining which robotics advances are real can be incredibly difficult Chris Paxton Jan 29, 2026 30 4 7 Share Image from Gemini The release of the Imagenet dataset was a landmark moment for the nascent field of deep learning. This collection of what is now 14,197,122 images covering more than 100,000 concepts (“synonym sets” or synsets) was key in driving and assessing early progress in deep learning, in part because it signaled the ongoing shift from “better algorithms” to “better data,” which in turn unlocked much stronger algorithms. Quintessential
saved by
related reading
- What Do Robotics Leaderboards Tell Us About The State of Robot Learning?itcanthink.substack.com
- RoboArena: Distributed Real-World Evaluation of Generalist Robot Policiesarxiv.org
- State of Robot Learning, December 2025vedder.io
- Sporks of AGIsergeylevine.substack.com
- How Can We Make Robotics More like Generative Modeling? | Eric Jangevjang.com
- Frontier of General-Purpose Robotics (2025)adampatni.com
- Many Small Steps for Robots, One Giant Leap for Mankindnotboring.co
- Eric Jang: Robots Must Be Ephemeralizedblog.evjang.com
- [2503.24278] AutoEval: Autonomous Evaluation of Generalist Robot Manipulation Policies in the Real Worldarxiv.org
- Demystifying evals for AI agents \ Anthropicanthropic.com
- Robots Must Be Ephemeralizedevjang.com
- Building Worlds That Train Robotsworldlabs.ai