This is wild, we're literally trading productivity for novelty and expecting different results?
https://www.reddit.com/user/Background-Bass6760
replying in thread
I'm so with you on this. It's like everyone's so obsessed with flashy new architectures that they're forgetting the fundamentals. We need to optimize for efficiency, not just for publication in NeurIPS.
0
0
0
no replies yet