Anthropic Developer Shares Fable 5 Prompting Tips: Find Your Own Blind Spots First
Anthropic developer Thariq Shihipar shares Fable 5 prompting techniques that teach you to identify your own blind spots before trying to outsmart the model — unlocking dramatically better results.
A New Era of Prompting: Why You Are the Problem, Not the Model
Anyone who codes with AI agents knows the problem well. You craft your prompt, the plan seems clear, but the result isn't quite right. According to Anthropic developer Thariq Shihipar, with Claude's latest model, Fable 5, this is less and less a problem with the model itself and more a result of the user's own blind spots.
Shihipar argues that Fable 5 is the first model where output quality is genuinely limited by the user's ability to clarify their unknowns. This represents a radical shift in thinking: instead of asking "how do I trick the model into giving better results," we should ask ourselves — "what don't I know that I don't know?"
The Four Categories of Knowledge That Transform Your Prompts
Shihipar breaks knowledge into four distinct categories. Known Knowns are what is already stated in your prompt. Known Unknowns are questions you know you haven't answered yet. Unknown Knowns describe knowledge so obvious that you would never write it down. The critical category is Unknown Unknowns — things you haven't considered at all.
This last category represents the single biggest obstacle to effective prompting. When you don't know that you don't know something, you can't tell the model to account for it.
The Specificity Dilemma: Too Precise vs. Too Vague
One of Shihipar's most important observations is that being too specific is just as bad as being too vague. When you are overly detailed, Fable 5 may rigidly follow your instructions even when a change of course would make more sense.
"When you don't account for your unknowns you fail both ways," Shihipar writes.
The Blindspot Pass: Your First Step Before Any Project
Shihipar offers a specific technique he calls a blindspot pass — ask Claude to identify your unknown unknowns. This works especially well when you're working in an unfamiliar part of a codebase.
Conclusion: The Future Belongs to Those Who Find Their Blind Spots
With the arrival of Fable 5, AI prompting is no longer just a technical skill — it is becoming a practice of self-awareness and metacognition. The more powerful models become, the more critical it is for users to understand the boundaries of their own knowledge.