📈Willison: Anthropic and OpenAI Have Finally Found PMF
TL;DR
Simon Willison argues the coding agents — Claude Code and Codex — are the product-market fit moment for the labs, not ChatGPT. He documents personal usage of $1,199 on Claude Code and $980 on Codex in a single month against $200 in subscriptions.
Simon Willison argues the coding agents — Claude Code and Codex — are the product-market fit moment for the labs, not ChatGPT. He documents personal usage of $1,199 on Claude Code and $980 on Codex in a single month against $200 in subscriptions.
Key Points
Willison's April bill: $1,199.79 on Claude Code, $980.37 on Codex via API
He paid only $200 across subscriptions in the same period
Enterprise pricing for both coding agents now matches list API price (no enterprise discount)
GPT-5.5 (April 23) is priced at 2x the GPT-5.4 API rate, and is still selling out
Hit 692 upvotes on Hacker News within 10 hours
Why It Matters
If a single power user spends $2,180 a month willingly, the unit economics of agentic coding are no longer hypothetical. The ChatGPT-vs-Claude consumer battle is the wrong scoreboard.
Quick Facts
Frequently Asked Questions
Why does this matter?
If a single power user spends $2,180 a month willingly, the unit economics of agentic coding are no longer hypothetical. The ChatGPT-vs-Claude consumer battle is the wrong scoreboard.
What happened?
Simon Willison argues the coding agents — Claude Code and Codex — are the product-market fit moment for the labs, not ChatGPT. He documents personal usage of $1,199 on Claude Code and $980 on Codex in a single month against $200 in subscriptions.
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.