Andrey Fradkin
@publiclibrary
Andrey Fradkin
@publiclibrary
"the question of whether a computer can think is no more interesting than the question of whether a submarine can swim."
- edsger dijkstra
It discusses the scaling of large language models, focusing on compute budgets, scaling laws, inference efficiency, and implications for training and model size optimization in various applications.
vladfeinberg.com