Y
Hacker News
new
|
ask
|
show
|
jobs
by
bunderbunder
15 days ago
More-or-less the same principle, but scaled up massively, and with context-dependent probability conditioning maps.