Dan Shipper is the co-founder and CEO of Every, a company combining AI journalism, software, and training. Through its products, his Chain of Thought column, and the AI & I podcast, he investigates how coding agents change what small teams can build and how organizations capture what they learn.
From founder to AI-native publisher
While studying at the University of Pennsylvania, Shipper co-founded Firefly with Justin Meltzer and Patrick Leahy. The browser-based customer-support startup received the first investment from First Round Capital’s Dorm Room Fund and was acquired by Pegasystems in 2014, after which Shipper worked on its collaboration technology inside the acquiring company.
He later created Superorganizers, a publication about how creative people organize their work. In April 2020, he and Nathan Baschez combined Superorganizers and Baschez’s Divinations newsletter, establishing the partnership that became Every. The business grew from a subscription bundle into a laboratory for AI-native company building, with products including the Cora email assistant, the Spiral writing application, the Monologue dictation app, and the collaborative editor Proof.
- Compound engineering: In an article coauthored with Kieran Klaassen, Shipper describes a development cycle of planning, agent execution, review, and codification. Teams preserve lessons in instructions, specialized agents, documented patterns, and automated checks so completing one task makes subsequent work easier. Klaassen developed the methodology at Every; Shipper helped articulate and popularize it.
- AI-native company building: Shipper argues that universal adoption changes coordination, not simply coding speed. Agents can adapt features between repositories and technology stacks, accelerate onboarding, enable cross-product pull requests, and help teams replace speculative proposals with working demonstrations. His AI Engineer talk draws on Every’s experience assigning substantial applications to individual engineers supported by coding agents.
- Agent-native architectures: His guide to agent-native applications, coauthored with Claude, treats agents as active software participants with access to tools, files, and application capabilities. Practical requirements include persistent state, interrupted-session recovery, and workflows spanning multiple steps.
- The two-slice team: Shipper explored whether one person working with agents could handle responsibilities associated with a conventional product team by building Proof, a collaborative Markdown editor for people and AI agents, while running Every. The product records who contributed individual changes, making authorship visible when humans and agents edit together.
