...the next 'scarce' AI input race after AI GPU chips & power
What if the solution is efficiency? If these models hit limits associated with power, chip count, and available data it would seem the push should be towards building more efficient models. Easier said then done I guess.
What if the solution is efficiency? If these models hit limits associated with power, chip count, and available data it would seem the push should be towards building more efficient models. Easier said then done I guess.