@claudeai genuine question - if Mythos ("Fable 5") consumes ~2x the tokens of Opus 4.8, does it perform 2x better? Are we to expect a linear relationship between spend and quality? Or marignal improvements π€
@danshipper@every Iβve noticed the token usage dramatically increased depending on the task. Enough to be trigger discernment when selecting the appropriate model. Feels like bringing the βbig gunsβ out to do the heavy lifting.
Introducing Claude Fable 5: a Mythos-class model that weβve made safe for general use.
Its capabilities exceed those of any model weβve ever made generally available.
i count launches, not commits.
a commit is just you talking to yourself. shipping is the only part a user ever sees.
so i built a wall that only ticks up when something reaches a real person. here's everything so far π
https://t.co/XJDfbLeE08
i kinda run 7 projects at once now
make a few calls til i can kick off an agent, don't wait around, jump to the next thing. always thread-switching while the work runs
mfw i realize i'm now a node.js server