Robotic platforms are a few iterations away from it at most
Robotic platforms are a few iterations away from it at most
Prompt engineering is an extremely simple form of behavior authoring. I'm curious about how you arrived at the conclusion that newer models make prompts and prompting strategies less relevant.
> Researchers from Microsoft tested GPT-4 on medical problems and found "that GPT-4, without any specialized prompt crafting, exceeds the passing score on USMLE by over 20 points and outperforms earlier general-purpose models (GPT-3.5) as well as models specifically fine-tuned on medical knowledge (Med-PaLM, a prompt-tuned version of Flan-PaLM 540B).
[0] https://en.wikipedia.org/wiki/GPT-4#Medical_applications
That’s a pretty concise description of behavior authoring that functionally work