Founders Must Taste AI Soup Directly

Howie Liu believes that in the AI era, founders and leaders must become individual contributors again, getting their hands dirty with the technology they're building. This return to "founder mode" isn't just about micromanaging—it's about understanding the rapidly evolving capabilities that only direct experience can provide.

The AI paradigm shift requires leaders to be intimately involved with product details because AI isn't a one-time form factor change like desktop to mobile. With each new model release, novel UX patterns and form factors emerge. To capitalize on these capabilities, you can't view things from a 10,000-foot perspective—you must understand what's possible at a granular level.

Liu personally spends significant time using AI tools, not just Airtable's but competitors' as well. He's proud to be the highest-inference-cost user at Airtable, spending hundreds of dollars on experiments that might yield strategic insights. This hands-on approach helps him understand both the models' capabilities and the product form factors in which they can be deployed.

For teams to succeed in this environment, Liu restructured Airtable into "fast thinking" and "slow thinking" groups. The fast thinking team ships new AI capabilities weekly, focusing on creating jaw-dropping value. The slow thinking team handles more deliberate, infrastructure-heavy bets. This structure allows for both rapid experimentation and thoughtful scaling.

Liu encourages his team to "play" with AI tools, giving them permission to block out entire days or weeks just to experiment. He believes the most successful product people in this era will be those who can cross traditional role boundaries—PMs who can prototype, engineers with design sensibilities, designers who understand technical constraints.

The practical implication for both leaders and ICs is clear: you need to become more versatile. The baseline of cross-functional knowledge is rising. Liu advises finding personal projects that force you to use these tools in meaningful ways, not just clicking around but building something useful. This hands-on experience is the only way to truly understand what's possible and to develop the intuition needed for product development in the AI era.

When reinventing a company for AI, Liu suggests asking: "If you were founding a new company from scratch with the same mission, how would you execute using a fully AI-native approach?" Then determine if your existing assets give you an advantage or if you'd be better starting fresh. For Airtable, their no-code components provide reliable building blocks that their AI agent can manipulate, offering advantages over generating everything from scratch.