Our LLM cost model was double-counting batch and cache discounts (beforeyouship.dev) 1 points by indiegoing 3d ago ↗ HN
1 comment
[ 0.15 ms ] story [ 110 ms ] thread