Claude, On Empty
Running out of Claude isn't a limit problem — it's a routing problem. Five moves that keep Fable available when it matters.
Claude, On Empty
Running out of Claude isn't a limit problem — it's a routing problem. Five moves that keep Fable available when it matters.
Every claim in this video was fact-checked against the source chapter and primary sources before rendering.
youtube.com/@NikBearBrown
Transcript
A lightly formatted transcript of the video, with timestamps so you can jump to any section on YouTube.
[0:00] I know people who stop working for days because they ran out of Claude. Not because the work stopped, because the model did. That's not a limit problem. That's a routing problem. The real cost isn't tokens. Run dry on Tuesday and the work just waits for the reset. The wait is the price and it's self-inflicted. Where did it go? One chip, the flagship effort on high for everything. Hard problems, email drafts, renaming files, all billed at genius rates. Flagship by default feels safe, but you're spending judgment on chores. The model that can hold your whole architecture in its head does not need to rename your files.
[0:40] Route by task, the bulk work, running skills, batch renders, long refactors, hand it to Opus and Claude code. That lane is built for the long haul. Save Fable for the words that get reused, the skill, the plan, the tricky call. Author with your best model once, the quality freezes into the file and every cheap run inherits it. Then Thursday's hard problem arrives and Fable is still there. Same budget, different week. The routing on one page. Fable writes what gets reused, Opus runs it. Medium effort for the everyday, high only when it earns it. Usage is a budget. Spend it where only your best will do.
[1:24] If this saves your Thursday, subscribe. Nick Bear Brown.
