21 August 2026
Newsletter CEO discusses balancing criticism with dependence on AI labs
First reported
Platformer ran this on .
- Every, a newsletter covering AI, relies on early access to models from Anthropic and OpenAI to test products and write about them.
- Every's CEO Dan Shipper says the publication can publish critical reviews despite this dependence because AI companies cannot fairly evaluate their own products.
- Shipper argues Every's value comes from acting as a neutral third party that readers can trust more than the companies themselves.
How it was covered
PlatformerCasey Newton
CEO Dan Shipper discussed the tension between publishing critical vibe checks of frontier models from Anthropic and OpenAI, whom Every depends on for early access and uses to build products. Shipper argues Every's role as an objective arbiter is durable because model companies cannot credibly evaluate their own products, positioning the publication as a trusted intermediary.