Every, a 30-person media and product company, has built a copy-editing agent trained on 30,000 historical edits from its editor in chief, Kate Lee. The goal is to distribute one employee's editorial judgment across the entire organization. CEO Dan Shipper also states that AI now writes essentially all of Every's code, while the company doubled its headcount from 15 to 30 over the past year.
Shipper is direct about two tensions worth reading for. First, Every publishes critical reviews of frontier models from labs like Anthropic and OpenAI, the same companies whose APIs power Every's products: Cora, Sparkle, Spiral, and Monologue, all bundled into a $20 monthly subscription. Their pan of Claude Sonnet 5, 'a model pitched for everyone impresses no one,' illustrates how the company navigates that conflict. Second, Shipper claims most writers are already using AI in their workflows and simply not disclosing it.
The interview's most substantive argument is Shipper's case for why Every keeps hiring despite aggressive automation. He frames AI as trained on the residue of human expertise, capable of recombining what already exists but blind to what comes next. That single claim does real work against the assumption that automation and headcount move in opposite directions. The full conversation is available on the Platformer podcast and YouTube.
[READ ORIGINAL →]