You are using Claude Fable 5 wrong
TL;DR
Fable 5 is most useful when it runs systems, not one-off prompts: Greg's best examples chain tools like ElevenLabs, Whisper, FFmpeg, Remotion, and Figma MCP to edit a launch video end-to-end instead of just asking for a single output.
A landing page tournament beats asking for one version of copy: His prompt generates 8 landing pages, has 5 judges score them, then merges the winner, producing copy so strong he says it beat what he'd expect after years serving millions of users.
The best startup prompt starts with pushback, not code: By telling Fable to interview him like Mark Zuckerberg, Sam Altman, or Brian Chesky, Greg got challenged with lines like "That's a horoscope answer, Greg," and ended with a tighter product spec and V1 architecture.
Low effort may be the alpha for routine work: Citing Morgan Linton and Factory.ai's Droid, Greg says people should orchestrate Fable 5 with cheaper models because Fable low can outperform Opus high while saving tokens before API-based pricing kicks in.
Fable 5 gets especially strong on dense business documents and history: Greg highlights 1 million token context, table-reading, and chart analysis as the reason it can audit contracts, scan churn and support tickets, and even build an operating manual from years of your own notes.
He sees immediate businesses in synthetic focus groups, 48-hour internal software, and contract refund audits: Examples include charging brands $3,000 to test ads with persona panels built from 500 reviews, selling custom tools to med spas for $5,000, and taking 25% of savings found in vendor contract audits.
The Breakdown
Greg Isenberg argues that most people are wasting Fable 5 on flashy demos when it can already act like a brutal product strategist, copy tournament judge, contract analyst, and custom software builder. His core bet: the real edge is pointing this model at money problems now, before pricing shifts to API tokens and the easy experimentation window closes.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
Cheap Models, Hard Tasks
Most agent workflows route every step to the frontier model by default. The bill scales with how chatty the agent gets, even when most steps don't need that brain.

Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.

Playbook
The Art of Tasteful Prompting
Learn how tasteful prompting helps you move beyond generic AI output by shaping context, style, and judgment from the start.