How to Get the Most Out of Fable 5 and GPT-5.6 Sol
Summary
This episode explores strategies for maximizing the potential of frontier models like GPT-5.6 and Claude 5. It covers official prompting guides emphasizing boundary setting, iterative steering, and efficient compute usage. The host also analyzes expert advice on increasing ambition, managing unknowns in agentic workflows, and implementing automated loops to unlock high-leverage, complex tasks.
Topics discussed
Intro, sponsors, and episode context
Overview of new frontier models and interaction patterns
Setting boundaries for tenacious models
Iterative steering and real-time collaboration
Efficient prompting for Chat vs. Work modes
OpenAI best practices: removing old instructions
Using AI for high-leverage impact work
Sponsors: KPMG, Blitzy, Retool, HyperAgent
Managing unknowns in agentic coding
Meta-prompts for harness optimization
Loop strategies for iterative improvement
Conclusion: Ratcheting up ambition with new models
Listen ad-free on Castria