The discussion showcases Avery as a model for an AI-native company. They employ a 'Head of AI Operations' to build automated workflows, and their product team of two engineers manages 15 AI agents (Claude Code instances) to build products without handwriting code.
Dan Shipper posits that as AI increasingly handles execution, the most valuable human skills will shift to those traditionally held by managers. This includes evaluating work, allocating resources, and providing high-level direction to AI agents.
A key assertion is that successful, company-wide AI integration is almost entirely dependent on the CEO's personal commitment. If the leader is not actively using AI tools daily, the organization is unlikely to realize significant productivity gains.
The conversation counters the narrative that AI will eliminate entry-level roles by highlighting its power as a learning accelerant. A junior employee was able to make a year's worth of progress in two months by using AI to instantly apply feedback and never repeat mistakes.
The episode details a shift from writing code to managing a team of specialized AI agents. Developers at Avery use multiple agents (Claude, Friday, Charlie) like a team of 'Avengers', each with different strengths, to generate, review, and refine code.
Keep pulling the thread on Dan Shipper.