Is the model actually following the prompt, or is it just taking the most reasonable action? @felixwyw tests this with pushing and stacking blocks. As usual, these objects and tasks were not seen in the training data!
A simple example of physical prompt steerability: same environment, different prompts, different behaviors.
Thanks @JagdeepBhatia8 for the suggestion.
Read more about GEN-1.5 in our blog post in the comments below.
Physical prompting is an elegant idea!
One request for the @GeneralistAI team: can we see an example of *physical prompt steerability*, where different prompts induce distinct behavior in the *same* environment?
I'm curious whether GEN-1.5 is listening to the physical prompt or
5 days to #Actuate26! Down to our last few expo passes (will probably sell out by Saturday) so get your ticket today!
Excited to announce Tuesday's panel: Data Wars: Human Data vs UMI vs Teleop.
Moderated by @ryancjulian (@NVIDIARobotics), with Naveen Kuppuswamy (@FlexionAI),
Any particular end-effector has tasks it can do easily, and other tasks that may be entirely impossible. The advantage of being a robot is that you can use whatever tool is best for the job.
Each hand is a different sensorimotor interface by which GEN-1 experiences the physical world. Scaling pretraining across thousands of these interfaces teaches GEN-1 a universal physical commonsense that transfers to new hands and new ways to grasp, push, pull, twist, and more.
Come see live demos of GEN-1 doing dexterous manipulation tasks at #Automate2026 — the largest automation show in America, with our friends @Universal_Robot and @FlexivRobotics
📍 UR @ at Booth #1250
📍 Flexiv @ Booth #2844
For early access to our models, please reach out 📧
