Episode Summary
Executive Summary: Howie Liu argues that legacy software companies must be reinvented from first principles in the AI era, not merely retrofitted with features. At Airtable, he’s pushing a CEO-as-IC model, faster experimentation, deeper product involvement, and a reorg into fast-thinking and slow-thinking teams. His core thesis: AI-native success requires hands-on use of models, collapsed role boundaries, prototype-first execution, and product-led discovery.
Main Topics: CEO as individual contributor in the AI era (Priority: 5/5): Howie explains why he has become much more hands-on—coding, testing AI products, shaping UX, and personally driving AI initiatives—because AI shifts require intimate product understanding rather than distant oversight. Rebuilding Airtable as an AI-native product (Priority: 5/5): He argues Airtable must be re-founded around AI, with the product experience, business model, and workflow rethought from scratch while still leveraging existing no-code primitives as an advantage. Fast-thinking vs slow-thinking organizational structure (Priority: 5/5): Airtable reorganized into a fast-moving AI platform group and a slower, infrastructure-oriented group so the company can ship new AI capabilities weekly without sacrificing foundational systems work. Play, experimentation, and prototype-first culture (Priority: 4/5): Howie pushes employees to spend time exploring AI tools, cancel meetings if needed, and share actual prototypes instead of docs, because AI products are best understood through hands-on experimentation. Role convergence across PM, engineering, and design (Priority: 4/5): He believes AI makes cross-functional skill growth more important: PMs should gain design sensibility, designers should understand technical constraints, and engineers should think more about product and business outcomes. Evals, vibes, and product discovery (Priority: 4/5): Howie says early AI product development should start with open-ended experimentation and vibes before formal evals, which become more useful once the form factor and use cases are clearer. Founder mode, humility, and staying close to the craft (Priority: 3/5): He frames his experience as a lesson in not stepping too far away from the product details, emphasizing that leaders should maintain product passion, humility, and gratitude while scaling.
Key Arguments: A company with a decade-old mission should ask: if it were founded today, how would it execute with AI-native workflows? If it can’t answer that convincingly, it may need to be sold or relaunched. CEOs in AI companies need to act more like contributors again: using models daily, testing products continuously, and staying close to the mechanics of the work. AI changes are fast enough that product teams must operate on weekly experimentation cycles, not traditional quarterly roadmap planning. Airtable’s existing primitives—real-time collaboration, no-code building blocks, layouts, automations—are an advantage because they can serve as a domain-specific language for agents. The best AI products are experiential; demos, functional prototypes, and real usage teach more than decks, PRDs, or marketing claims. Cross-functional fluency is becoming table stakes: PMs should prototype, designers should understand technical feasibility, and engineers should think about user outcomes. Evals are important, but only after a product direction is discovered; early on, open-ended testing and judgment matter more than rigid measurement. Legacy process-driven management can prevent breakthrough innovation because it encourages incremental thinking instead of holistic product leaps.
Data Points: Airtable company age: 13 years - Howie describes Airtable’s history as stretching back about 13 years. AI usage frequency by Howie: Hourly / multiple times per hour - He says he uses AI products constantly, not just daily. Airtable AI inference cost ranking: #1 most expensive user internally and, for a long time, globally across Airtable customers - Howie says he intentionally runs large, expensive AI workloads to extract strategic insights. Inference cost example: Hundreds of dollars per exercise - He cites spending hundreds of dollars on transcript analysis as trivial relative to potential value. AI adoption at Airtable: Half the company or half the EPD org - He says roughly half of Airtable’s product/engineering/design organization is working on AI capabilities. ChatGPT weekly usage: 700 million weekly active users - Howie references OpenAI’s reported scale to show how fast AI products can reach massive distribution. ChatGPT adoption share: 10% of humans on Earth weekly - He uses the figure to underscore the magnitude of ChatGPT’s ramp. Deep research cost: On the order of $1+ per research call - He mentions this in the context of using deep research to replace manual research work. AI team cadence: Weekly shipping cadence - Fast-thinking AI platform teams are expected to ship new capabilities nearly every week. Model/product history: GPT-3.5 era to current models - He contrasts early autocomplete-style coding tools with current agentic app-building systems.
Pivotal Quotes: "If you were literally founding a new company from scratch with the same mission, how would you execute on that mission using a fully AI native approach? If you can't, then you should find a buyer." — Howie Liu: On how legacy companies should evaluate whether they can truly adapt to AI. "If you want to cancel all your meetings for like a day or for an entire week and just go play around with every AI product that you think could be relevant to Airtable, go do it." — Howie Liu: On how he wants employees to learn AI tools through experimentation and play. "I think you need get like decently good at all three." — Howie Liu: On the future of PM, engineering, and design roles becoming more cross-functional.
Implications: For builders and leaders, AI winners will likely be the companies that redesign workflows, teams, and product UX around model capabilities—not those that merely add AI features. Expect more role blending, faster shipping, and stronger pressure on executives to stay deeply hands-on.
About Lenny's Podcast
Lenny Rachitsky interviews world-class product leaders and growth experts about building products and growing careers.