Simon Willison: Anthropic and OpenAI Have Found Product-Market Fit — In Coding Agents, Not Chat
Published 2026-05-27Ingested 2026-05-28AI-Assisted DevelopmentHigh
Summary
Simon Willison argues that the April 2026 enterprise-pricing shift at both Anthropic and OpenAI — moving heavy users from heavily-discounted seat-based subscriptions to API-aligned usage pricing — confirms that coding agents (Claude Code, Claude Cowork, Codex) are where the labs have actually achieved product-market fit, not consumer ChatGPT. His own usage at API rates would cost ~$2,180/month vs. the ~$200 he pays under subscription; both labs have decided that arbitrage cannot continue. Willis
Radar Context
Claude CodeGitHub CopilotDevin
Alignment: Reinforces current position
Related Positions: AI-Assisted Development Tooling, Enterprise AI Delivery, Multi-Model Multi-Vendor
Related Partnerships: Anthropic (Claude), Microsoft (GitHub / Copilot), Cognition (Windsurf / Devin)
simon-willisonclaude-codecodexcopilotproduct-market-fitenterprise-pricingagentic-codingubermicrosoftanthropic-revenue