Skip to content
Better HN
Top
Best
Ask
Show
New
Jobs
Search
⌘K
0 points
EGreg
4y ago
0 comments
Save
Share
Sounds like how AlphaGo beat Stockfish or Rybka… massive precomputing vs alphabeta search in real time
0 comments
2 comments · 1 top-level
top
newest
oldest
sillysaurusx
4y ago
· 1 in thread
Hmm. AlphaGo and AlphaZero were trained using self play. I don't think there was any precomputing, unless you count training as precomputation.
EGreg
OP
4y ago
Precomputation on a huge cluster for a long time
j
/
k
navigate · click thread line to collapse
