factual

Embodied AI Systems Are Stitched, Not Trained End-to-End

Projects like Google's SayCan plug pretrained language models into robots to translate natural-language instructions into actionable sub-tasks using a vision-language model, but the key limitation is that these systems are not trained end-to-end through world interaction—they bolt together passively-pretrained modules—raising a deep question about whether such modular stitching, versus learning from the ground up via rich multimodal interaction, is the right model of cognition.

factualpending

Speaker

Raphael Milliere

Evidence Quote

“there's perhaps a surprisingly deep question here about whether this kind of system is at all useful to think of what might be happening in human cognition”

Source

230 | Raphaël Millière on How Artificial Intelligence Thinks— Sean Carroll's Mindscape
Created: 6/13/2026, 12:19:51 AM

Your notes

Loading notes...