Laurie Voss: The real cost of software is now defining user intent

Summary

Laurie Voss argues that as AI reduces the costs of writing, reviewing, and operating code, the irreducible cost of software will shift entirely to understanding what users want and defining it clearly. Voss suggests that in a world of unlimited software demand, this represents the new 'full job.'

Why it matters

This approach transforms the AI-assisted coding narrative from one of 'less work' into a structural redistribution of where value and cost concentrate in the software lifecycle. It indicates that the bottleneck is no longer technical implementation, but rather product intuition, requirement clarity, and user experience design. These skills do not scale as easily as code generation.

Highlights

• Code writing costs have collapsed; review, correction, and operating costs are expected to follow suit
• The remaining cost falls per product: discovering user intent and defining it clearly
• With unlimited software demand, product definition is becoming the entirety of the engineering job

Critical take

This argument rests on the assumption that 'learning what people want' is the only remaining non-transferable cost. However, this assumption is accepted without scrutiny. In reality, code produced at scale by AI creates new cost categories, such as security audits, coordination overhead, and the maintenance of heterogeneous AI software. Voss's regular triad does not account for these items. Moreover, product intuition and user experience patterns are transferable to a significant degree across products, which weakens the claim of 'non-transferability.'