The PodKnowledge|← All Episodes
Based on Lenny's Podcast data
Lenny's Knowledge Sketch

Fast Thinking,
Slow Thinking

Howie Liu
Co-founder & CEO, Airtable · 13 years running the company · Betting on the IC-CEO role
AUG 31 2025
The Down Moment

"Airtable Is Dead"
(the Viral Tweet)

VIRAL TWEETwrong numbers,spicy takesALL-IN COVERS ITlater ran thecorrection episode
"This particular tweet went super viral and that was the perplexing part... Then the All In podcast covered it, which really gave it legs."
  • A tweet full of unsubstantiated numbers called Airtable dead; the media echo carried it
  • All In later issued a correction: "we got the numbers wrong, we're revising our view"
  • Howie sees it as a case study in what happens to the last generation of high-valuation companies in a new market
  • Underneath it, Airtable was quietly maturing through a very SaaS era into its gen-AI act
Framework

Two Groups, Complementary

FAST THINKING"AI Platform" groupnear-weekly capability shipsjaw-drop value eachSLOW THINKINGinfrastructure & scaleHyperDB, multi-100M recordsthe deliberate build
"You need fast and slow thinking to operate as a human. We now have two separate parts of the company. They complement each other."
  • Named directly after Kahneman's Thinking Fast and Slow — not better, just different modes
  • Fast thinking creates the top-of-funnel excitement and inspires new use cases
  • Slow thinking lets those initial seeds sprout and grow into large deployments, including enterprise
  • The failure mode Howie sees in AI-native peers: wide top of funnel, no infrastructure to hold what lands
Who thrives in fast thinkingSomebody who operates with a lot of autonomy, entrepreneurial by nature. Doesn't have to be a former founder, though Rippling literally acquihires founders for this. Howie hires for the trait, not the resume.
The Turn

The IC CEO Move

  • The trend Howie is at the front of: CEOs becoming individual contributors again — coding, building, leading initiatives directly
  • The old, disintermediated version made him feel farthest from the actual substance of the company
  • "Inspect at the topmost layer" mode can hide coasting for a long time; nobody notices until it's late
  • His own reversal: for the right kind of leader, the IC role is more fun and more invigorating
Chesky-style CPO overlap

Howie's read from Brian Chesky: a serious product company needs a CEO who plays a CPO role. You can't coast on scaling GTM around a product forever.

The Procter & Gamble test

Even at a CPG giant, you'd want a CEO who tastes the soup and sees how the product is experienced on the shelf. Delegate up to a point — but keep tasting.

Playbook

How Howie Rewires a Team

  • Cancel your meetings for a week and go play with every AI product that could touch your product; nothing replaces first-hand time
  • Move from feature-shaped teams to mission-shaped groups — incremental remits produce incremental thinking
  • Separate the near-weekly-ship group from the multi-quarter-infrastructure group, on purpose, not by accident
  • Refuse the framing that the CEO shouldn't be in the weeds; the substance is in the weeds
A step-function vision, not many small onesThe best way to innovate on the product is not to split incrementally across surface-area teams. It's a bigger vision of where the product needs to leap next.
Contrarian

Howie's Uncomfortable Takes

✗Feature teams ⇒ better productsINSTEAD →✓ Feature teams produce incremental improvements by definition. The step-function wins come from mission-shaped groups that can move across surface areas.
✗Mature CEOs delegate everythingINSTEAD →✓ The IC role is where the substance is. "Less decisions I could make the better" is the mode a coasting company operates under until it isn't.
✗One org can do both fast and slow work wellINSTEAD →✓ Split them on purpose. Fast thinking ships jaw-drop capability weekly; slow thinking builds the HyperDB that holds it. They complement each other; they don't share a team.
✗Wide AI top-of-funnel is enoughINSTEAD →✓ AI-native peers get tourist traffic and can't hold it. The slow-thinking infrastructure is what turns adoption into deployments that scale.
Based on Howie Liu's episode on Lenny's Podcast. All ideas on this page are from the episode.Watch on YouTubeFollow @howietl on X
GO DEEPER IN THE EPISODE
04:05 The “Airtable is dead” viral tweet controversy08:07 The rise of IC CEOs10:57 AI’s paradigm shift in product development16:27 Specific changes Airtable has made21:38 Fast- and slow-thinking teams32:57 The emergence of new form factors in AI models34:48 Airtable’s vision and philosophy40:20 Empowering teams with AI tools46:50 Encouraging experimentation and play50:55 Cross-functional skills in product teams01:03:35 The importance of evals and open-ended testing01:08:06 Key strategies for AI-driven success01:12:43 Counterintuitive startup wisdom01:22:21 Don't step away from the details that you love01:25:50 Advice for aspiring engineers and designers01:30:00 Lightning round and final thoughts
Every sketch here began as a transcript.

Send us yours and we’ll draw it. $1 a transcript.

Sketch mine →
𝕏︎ X / Twitterin LinkedIn📸 Instagram🔗 Copy link
0:00