How we restructured Airtable’s entire org for AI | Howie Liu (co-founder and CEO)

Howie Liu is the co-founder and CEO of Airtable, the no-code platform valued at around $12 billion. After a viral tweet declared “Airtable is dead” based on incorrect data, Howie led a radical transformation: reorganizing the entire company around AI, becoming an “IC CEO” who codes daily, and achieving over $100 million in free cash flow.What you’ll learn:1. The “fast thinking” vs. “slow thinking” team structure that lets Airtable ship AI features weekly (inspired by Daniel Kahneman)2. Why Howie uses AI hourly (not daily) and is Airtable’s #1 inference-cost user globally3. Why CEOs must become ICs again in the AI era (and how to restructure your calendar to make it possible)4. Why “playing” with AI tools should be mandatory—Howie tells employees to cancel all meetings for a week to experiment5. The specific skills product managers, engineers, and designers need to develop to succeed in the AI era6. Why evals can kill innovation (and when to use “vibes” instead)—Brought to you by:LucidLink—Real-time cloud storage for teamsDX—The developer intelligence platform designed by leading researchersClaude.ai—The AI for problem solvers and enterprise—Where to find Howie Liu• X: https://x.com/howietl• LinkedIn: https://www.linkedin.com/in/howieliu/• Email: howie@airtable.com—Where to find Lenny:• Newsletter: https://www.lennysnewsletter.com• X: https://twitter.com/lennysan• LinkedIn: https://www.linkedin.com/in/lennyrachitsky/—In this episode, we cover:(00:00) Introduction to Howie Liu and Airtable(04:05) The “Airtable is dead” viral tweet controversy(08:07) The rise of IC CEOs(10:57) AI’s paradigm shift in product development(16:27) Specific changes Airtable has made(21:38) Fast- and slow-thinking teams(32:57) The emergence of new form factors in AI models(34:48) Airtable’s vision and philosophy(40:20) Empowering teams with AI tools(46:50) Encouraging experimentation and play(50:55) Cross-functional skills in product teams(01:03:35) The importance of evals and open-ended test