The Ash AI Tax
I kept noticing my AI agent burned more tokens on our Ash code than on plain Phoenix. So I measured it: ~2.6x more on Claude, ~1.9x on Codex — building the exact same feature.
Posts about large language models in practice: measurements, token bills and the places where they help or get in the way.
1 post with this tag.
I kept noticing my AI agent burned more tokens on our Ash code than on plain Phoenix. So I measured it: ~2.6x more on Claude, ~1.9x on Codex — building the exact same feature.