3 ms·
Yea.. nah. Seriously. This "more cores more gpus more data more training" doesn't explain how. It just says more is better. How do you know how close to the a
by ggm 5y ago
Yea.. nah.
Seriously. This "more cores more gpus more data more training" doesn't explain how. It just says more is better.
How do you know how close to the asymptote you are before you get there?
There's no sign of emergent intelligence. There's heaps of value in NLP and image classification but intelligence is more ephemeral.
I'm sticking to no, and I've been there since the decade after Lighthill.