Sporks of AGI - by Sergey Levine - Learning and Control
sergeylevine.substack.com · 2,273 words · saved by 1 readers
Why the Real Thing is better than the Next Best Thing
Training big models is really hard, and as the models get bigger and expand into new domains, it’s only getting harder. LLMs use lots of text data, while VLMs require data with text and images, and vision-language-action (VLA) models in robotics require lots of data of robots performing real tasks in the real world. This hits agents especially hard: whether you want to control a real-world robot or take actions to fulfill user requests on the web, data of real-world interactions with action labels can’t be obtained as cheaply as text and images from the web. It’s no wonder that researchers…
related reading
- Sporks of AGIsergeylevine.substack.com
- Sporks of AGIsubstack.com
- Sporks of AGIsergeylevine.substack.com
- State of Robot Learning, December 2025vedder.io
- Emergence of Human to Robot Transfer in Vision-Language-Action Modelspi.website
- Generalist - GEN-0 / Embodied Foundation Models That Scale with Physical Interactiongeneralistai.com
- Many Small Steps for Robots, One Giant Leap for Mankindnotboring.co
- ACT-1: A Robot Foundation Model Trained on Zero Robot Data | Sunday Robotics | The helpful robotics companysunday.ai
- A Steerable Model with Emergent Capabilitiespi.website
- How Can We Make Robotics More like Generative Modeling? | Eric Jangevjang.com
- All Roads Lead to Robotics | Eric Jangevjang.com
- How Claude Performs on Robotics Tasks \ Anthropicanthropic.com