As AI models become cheaper and more capable, the true value shifts from execution to the human imagination needed to identify novel, high-impact tasks that AI can solve, exemplified by an engineer spending $40 on a complex optimization only a frontier model could handle. This highlights that while routine tasks can be commoditized by inexpensive models, pioneering new AI applications requires creativity, deep expertise, and organizational support to explore uncharted possibilities.
The video discusses a curious phenomenon in AI usage: as AI tools improve and become cheaper, outputs across users and competitors start to look increasingly similar. This convergence is not due to the tools themselves but because AI has commoditized execution—doing known tasks efficiently and cheaply. The real value, the video argues, shifts away from execution to imagination: the ability to conceive new, valuable tasks that AI can perform, tasks that no existing to-do list or process has captured.
The story of Mitchell Hashimoto, a respected engineer, illustrates this point. Hashimoto tested expensive frontier AI models against cheaper alternatives on routine coding tasks and found that all produced similarly acceptable results, with the cheaper models being far more cost-effective. However, when Hashimoto assigned a novel, complex optimization task that no one had previously identified, only the expensive frontier model could handle it effectively, taking two hours and costing $40. This task was not on any backlog or sprint but emerged from Hashimoto’s deep expertise and imagination about what AI could now achieve.
This example highlights that AI’s ceiling of value is not the model’s price or prompt engineering but the breadth of tasks one knows how to ask for—the imagination to identify new problems worth solving. Execution remains critical, but it is multiplied by imagination. Cheap models handle routine work, but frontier models are essential for pioneering new applications that redefine what AI can build. The video stresses that as execution commoditizes, the premium shifts to those who can creatively leverage AI’s capabilities to explore uncharted territory.
The video also draws parallels to historical technological shifts, such as the transition from steam to electric power in factories and the smartphone market disruption by Apple. In both cases, success came not from better execution alone but from reimagining the problem and redesigning systems around new capabilities. Similarly, companies and individuals must move beyond running old task lists faster and cheaper; they need to invest time and resources into “scouting” new AI possibilities and empowering those with context and expertise to experiment and innovate.
Finally, the video warns against relying solely on hiring imaginative individuals without embedding imagination into the broader organization. Imagination thrives when paired with deep context and permission to experiment. The ability to pose high-value, frontier questions to AI models without bureaucratic barriers is crucial. The recent rapid adoption and return of the Fable 5 model exemplify how those who invested early in imagining new uses continue to lead, underscoring that the future of AI value lies in human creativity combined with powerful execution tools.