| url | https://humanloop.com/blog/openai-plans |
|---|
Notes from a private meeting Sam Altman had with ~20 developers in May 2023, leaked (then partly retracted, then re-published) by Humanloop’s Raza Habib. The clearest single window into how Altman was thinking about strategy in the months after ChatGPT’s hypergrowth.
The two killer claims
1. ChatGPT is the strategy. Altman said OpenAI “would not release more products beyond ChatGPT.” His reasoning was historical: great platform companies need a killer app, and being a customer of your own API is how you make the API better. ChatGPT was that killer app.
2. Plugins don’t have PMF. Altman was unusually blunt: “the usage of plugins, other than browsing, suggests that they don’t have PMF yet.” He went further — “a lot of people thought they wanted their apps to be inside ChatGPT but what they really wanted was ChatGPT in their apps.”
That second sentence is one of the most quietly devastating product diagnoses of the AI era. Plugins were a hypothesis about distribution. Altman called it within months.
Why it matters here
This piece is the smoking gun for treating ChatGPT as a deliberate aggregator play: own the demand, then let the platform follow. It’s also a real-world demonstration of Vohra’s leading-indicator thinking — Altman wasn’t waiting for plugin revenue numbers to roll in. He looked at usage patterns and called the lack of PMF early.
See chatgpt-pmf for how this fits into the broader case study, and openai-windows-play for Ben Thompson’s parallel argument that platforms come from popular products, not the other way around.