The man behind Claude Code says you’re comparing AI costs to the wrong thing

Boris Cherny, creator of Claude Code, argues that companies evaluating AI investments are using incorrect comparison metrics. Rather than focusing on traditional cost-per-task calculations, organizations should reassess how they measure AI's value proposition and ROI.

Cherny's critique addresses a fundamental misalignment in how enterprises approach AI cost-benefit analysis. Rather than comparing AI expenses to previous infrastructure or labor costs, companies should evaluate AI against the specific outcomes it enables. This distinction matters because traditional cost comparisons often miss productivity gains, quality improvements, and entirely new capabilities that weren't previously measurable or available. The shift reflects a broader maturation in AI adoption where early-stage cost obsession gives way to outcome-focused investment strategies. Companies using legacy benchmarking frameworks risk underestimating AI's value and making suboptimal allocation decisions. This perspective stems from observing real-world deployments where businesses achieve disproportionate returns when they reframe AI evaluation. The implications extend beyond procurement to strategic planning—organizations that properly measure AI impact gain competitive advantages in speed, capability, and scalability. Cherny's messaging resonates in an environment where AI adoption is accelerating but many enterprise decision-makers remain uncertain about true ROI. For developers and AI service providers, this highlights the importance of articulating value creation rather than merely defending pricing. The industry is transitioning from purely efficiency-driven narratives to outcome-based justifications. This shift could reshape how AI services are packaged, priced, and sold across enterprise segments.

• →Companies are using outdated cost comparison frameworks when evaluating AI investments
• →Outcome-based metrics provide better assessment of AI value than traditional cost-per-task calculations
• →Proper AI evaluation requires measuring productivity gains and new capabilities rather than just expense reduction
• →Misaligned measurement frameworks could lead enterprises to underfund or underutilize AI solutions
• →The shift toward outcome-focused evaluation represents AI market maturation beyond early adoption phases