2 Comments
User's avatar
richardstevenhack's avatar

While I continue to think "tokenmaxxing" is just another attempt at "lines of code" productivity measurements - which was debunked decades ago - this line: "The data paints a clear picture: you are overpaying if you default to frontier models for everything" is something I've been saying.

But no, for almost everyone, it's "All Claude, All The Time" - despite the idiocy.

Ari's avatar

True! It's crazy how many people haven't even seriously tried open-weight models.