When you say Sora can lead to 0 to 1 progress in robotics do you mean AVs specifically or robotics in general? As far as I understand, RL training for robotics requires physics accuracy and Sora doesn't have physics accuracy according to OpenAI: "The current model has weaknesses. It may struggle with accurately simulating the physics of a complex scene, and may not understand specific instances of cause and effect."
People are clever, and its me speculating 1 to 2 generations down the line. Lots of data augmentation doesnt need to be perfect, and I bet this will be close enough for a lot of use cases.
When you say Sora can lead to 0 to 1 progress in robotics do you mean AVs specifically or robotics in general? As far as I understand, RL training for robotics requires physics accuracy and Sora doesn't have physics accuracy according to OpenAI: "The current model has weaknesses. It may struggle with accurately simulating the physics of a complex scene, and may not understand specific instances of cause and effect."
Link -https://openai.com/sora#:~:text=The%20current%20model%20has%20weaknesses.%20It%20may%20struggle%20with%20accurately%20simulating%20the%20physics%20of%20a%20complex%20scene%2C%20and%20may%20not%20understand%20specific%20instances%20of%20cause%20and%20effect.
People are clever, and its me speculating 1 to 2 generations down the line. Lots of data augmentation doesnt need to be perfect, and I bet this will be close enough for a lot of use cases.
Exciting time ahead for Robotics!