Are Transformers Truly Foundational for Robotics?
This addresses the problem of inefficient AI in robotics for researchers and engineers, but it is incremental as it critiques existing methods without new results.
The paper questions the utility of GPTs in robotics due to high compute costs and training times, contrasting them with insect brains to propose biological lessons for improvement.
Generative Pre-Trained Transformers (GPTs) are hyped to revolutionize robotics. Here we question their utility. GPTs for autonomous robotics demand enormous and costly compute, excessive training times and (often) offboard wireless control. We contrast GPT state of the art with how tiny insect brains have achieved robust autonomy with none of these constraints. We highlight lessons that can be learned from biology to enhance the utility of GPTs in robotics.