Aug 31, 2025 · 1h 40m · lennys-podcast
How we restructured Airtable's entire org for AI | Howie Liu (co-founder and CEO)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Airtable co-founder and CEO Howie Liu joins Lenny Rachitsky to discuss how legacy software companies must reinvent their strategy, leadership, and operational structures for the AI era. He outlines Airtable's transition to a hands-on 'IC CEO' model, the division of engineering into fast and slow thinking teams, and the emergence of multidisciplinary hybrid roles.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Lenny holds 17.5% of the talking time here. How this is scored →
Lenny as informed peer 4.3 Guest teaching 5.2 Guest disagreement 2.0 Lenny pushing back 1.2
speaking balance: gold is Lenny, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:03:55 Rejecting premature evals in favor of vibes
Howie explicitly challenges the widespread consensus from Anthropic and OpenAI leaders about starting with evals, arguing that rigorous benchmarks early on stifle creative product discovery.
Hardest push from Lenny ▶ 5:02 Confronting viral bankruptcy narrativeLenny directly confronts Howie with the viral 'Airtable RIP' tweet alleging the company had over-raised and was financially underwater.
Biggest teaching moment ▶ 57:43 Correcting Lenny on multi-disciplinary baselineWhen Lenny suggests PMs and engineers only need to adopt one adjacent discipline, Howie directly reframes the standard, asserting that builders must attain baseline competence across all three product disciplines.
Lenny holds their own ▶ 28:44 Synthesizing AI acceleration frameworksLenny draws on his conversation with OpenAI's Nick Turley, validating and enriching Howie's thesis by framing velocity and experiential discovery around maximal acceleration.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Lenny as informed peer | Guest teaching | Guest disagreement | Lenny pushing back | Why |
|---|---|---|---|---|---|---|
| Addressing Misinformation and Viral Internet Narratives | 4 | 4 | 3 | 2 | Lenny brings up a viral tweet alleging Airtable was in serious trouble financially. Howie debunks the claim by explaining the inaccurate metrics and how viral misinformation spreads, while Lenny chimes in with cultural commentary on Twitter dynamics. | |
| The Return of the Hands-On IC CEO in Tech | 4 | 5 | 2 | 1 | Lenny introduces the emerging trend of the 'IC CEO' and cites Howie's role as chief tastemaker. Howie expands in depth on why software CEOs must remain directly involved in low-level UX and architectural details during platform shifts. | |
| Leveraging High-Volume AI Inference for Strategic Insights | 5 | 5 | 2 | 2 | Lenny brings in an insight from Dan Schipper about measuring CEO AI usage. Howie takes the idea further, explaining how he acts as a heavy consumer of LLM inference compute to run deep qualitative analysis across massive corporate datasets. | |
| Airtable's Reorganization: Fast Thinking vs. Slow Thinking | 4 | 6 | 2 | 1 | Lenny asks how Airtable adjusted its org structure to move with AI startup speed. Howie outlines his Kahneman-inspired restructuring into 'fast thinking' (AI platform shipping weekly) and 'slow thinking' (core infrastructure and scalability) pillars. | |
| Building AI-Native Workflows and Rich User Experiences | 4 | 5 | 2 | 1 | Lenny mentions his firsthand test with Airtable's Omni builder. Howie explains his UX philosophy on why standard chat interfaces under-merchandise AI power and need rich graphical affordances and generative UI widgets. | |
| Velocity, Experiential Adoption, and the Evolution of PLG | 6 | 5 | 2 | 2 | Lenny connects Howie's points with insights from OpenAI's Nick Turley regarding pace and product discovery. Howie compares PLG approaches in AI, analyzing ChatGPT's unprecedented weekly active user adoption curve against enterprise sales motions. | |
| Vibe Coding Limitations and Airtable's Modular Architecture | 4 | 6 | 3 | 1 | Howie addresses the wave of AI app builders, critiquing full code-generation for business apps due to context collapse and bugs. He argues Airtable's modular database primitives provide a reliable domain-specific language for AI agents. | |
| Cultivating Product Intuition Through Personal Experimentation | 4 | 5 | 1 | 1 | Lenny synthesizes the key themes across legacy companies transitioning to AI. Howie elaborates on his personal habit of rapid weekend experimentation across diverse AI tools to sharpen intuition about emerging form factors. | |
| Sponsor Break: Anthropic Claude | 4 | 5 | 1 | 1 | Following an ad read, Lenny asks how Howie sets expectations across engineering, design, and PM roles. Howie shares his mandate allowing staff to drop meetings for entire weeks to prototype and play with new AI capabilities. | |
| Learning by Doing and Practical Implementation of Field Agents | 4 | 6 | 3 | 1 | Lenny suggests that PMs and engineers only need to pick up one adjacent skill. Howie gently pushes back, insisting modern product builders must develop a functional baseline across product, engineering, and design simultaneously. | |
| Navigating AI Product Discovery: Vibes First, Evals Second | 5 | 7 | 4 | 2 | Lenny brings up the standard industry emphasis on rigorous evals. Howie challenges this conventional wisdom, arguing that novel AI product discovery must start with open-ended 'vibes' before premature evals restrict the exploration space. | |
| Breaking Down Functional Silos Across the Entire Organization | 5 | 5 | 2 | 1 | Lenny summarizes actionable guidelines for organizational transformation. Howie supplements the summary by advocating for dismantling functional silos across sales, solutions engineering, and marketing to build full-stack autonomy. | |
| Founder Mode, Holistic Leadership, and Staying Close to Craft | 5 | 6 | 2 | 1 | Howie reflects on Chesky's 'Founder Mode' and the danger of traditional operational advice that forces CEOs into detached management. Lenny and Howie discuss balancing granular involvement with avoiding counterproductive micromanagement. | |
| Empowering Non-Technical Builders in the AI Era | 4 | 5 | 1 | 1 | Howie gives an encouraging closing argument that anyone can become a versatile builder today thanks to conversational AI tutoring. Lenny enthusiastically agrees, reminiscing about the painful hacks of early web development. | |
| Lightning Round: Media Picks, Craftsmanship, and Life Philosophy | 3 | 3 | 0 | 0 | In the lightning round, Lenny and Howie enjoy an amiable back-and-forth discussing sci-fi literature, Seth Rogen's new show, vintage Japanese textile manufacturing, and living with humility and gratitude. |
Statements from this episode (25)
Liu: Viral 'Airtable is dead' tweet used completely false revenue metrics
“Basically none of it was true. And I mean, the surprising thing to me was how viral this tweet went when, frankly, like I actually looked back at this person's other tweets. I think they worked at CB insights. And the irony is like the whole point of that busi…”
Liu: All-In Podcast issued a correction after misstating Airtable's revenue numbers
“They did also issue a correction though all in did a follow-up episode a few, few I think weeks later saying like, hey, like, you know, we got the numbers wrong, like you know, we were revising our case and kind of a view on Airtable.”
Liu: Post-Elon Twitter Is Better Due to Bolder Product Execution
“I would say net net, I like The post-Elon Twitter more than the pre-Elon Twitter because it's just bolder, and like, I, you know, I guess I really admire bold product execution where you're not just kind of stuck to, like, the current laurels, and they've made…”
Liu: Airtable and Figma both spent 2.5 years building before launching
“I think the same is true for Figma, which you know, actually like had a very parallel timeline to us. Like we both were founded around the same time, both spent two and a half years building the product like hands on you know, that early team before launching.”
Liu: Every software product has to be refounded for AI
“Regardless, I think that now we're entering this moment where like every, I mean, certainly every software product, in my opinion, has to be refounded because like AI is such a paradigm shift It's not even like, just like the shift from desktop to mobile or on…”
Liu: Product leaders must play directly with underlying AI primitives
“Like, AI is something you have to play with, and ideally, you're playing with both the, like, kind of packaged up, you know, app or solution that you've built with it, but you're also playing around directly with the underlying primitives. You're using The mod…”
Liu: I am personally the highest inference-cost user of Airtable AI
“I'm pretty sure I'm still the I just checked this recently, but like I take pride in being the number one most expensive An inference cost user of Airtable AI not just within our own company, but I think for a long time I was globally across all our customers …”
Liu: Half of Airtable's engineering, product, and design org works on AI
“Now we have a weekly, Basically sprint check in on all of our AI execution stuff, which now is like half the company or half the EPD org is working on AI capabilities.”
Liu: Feature-based team structures create incremental rather than mission-driven thinking
“You tend to think incrementally when everyone's remit is actually like a feature that they incrementally improve by definition, as opposed to thinking about like a mission or like a outcome goal, right? That might need to, you know coordinate, you know, dramat…”
Liu: Airtable's HyperDB handles multi-hundred-million-record datasets
“Our data store, HyperDB, that now can handle, like, multi-hundred million record data sets.”
Liu: Many AI-native startups struggle to turn tourist traffic into durable retention
“A lot of the challenge for many of the AI native companies I've seen Is that they could have, like, a very wide top of funnel, like, get all of this AI tourist traffic, you know, a lot of interest, a lot of, like, kind of, like, you know, early usage, but then…”
Liu: Airtable continued hiring in EPD during both company layoffs
“Even when we had to do, ah, two riffs, right, that, that significantly, you know, kind of reduced our head count, you know, we had just, like, way too quickly grown and overscaled the business at a certain point but even when we did our riffs, we were still ac…”
Liu: Airtable is adding code generation to its Omni conversational agent
“We're doing this new thing that's about to ship where, you know, not only can you describe the app you want to build and then iterate on it with, you know, kind of our conversational agent Omni, but and it builds it with like the existing air table platform ca…”
Liu: AI products under-merchandise capabilities with blank chat boxes
“What's under executed today in the world of AI products is like, there's so many awesome capabilities of AI and most of them are really under merchandise and there's like very poor actually visual or otherwise metaphors or affordances given to users to help re…”
Liu: Airtable made its AI agent the default interface for the entire product
“And in our case, like we just like made the entire product experience AI centric, right?
Like it used to be that like, you know, we had kind of this like secondary thing that you could ask questions to the assistant sidebar.
We now made our agent the default w…”
Liu: Pre-AI founders unable to leverage existing assets should sell their companies
“And if you can't really introspect and say like, look, I think I'm better off doing this with the pieces that I have for my existing business and product, then I think you should sell, right? Like you should find a buyer. For that company and then go and like,…”
Liu: Generating full business apps from code causes bugs and context collapse
“And yet if you have an agent that has to generate every single bit of that app from scratch from code, it's going to be very unreliable. There's going to be bugs. There's going to be data and security issues. And then you're also going to have a context collap…”
Liu: Non-technical users get stuck in v0, Lovable, and Replit when reprompting fails
“Whereas if you're not technical, you can't inspect the code underneath a V zero or lovable or Revlin app, right? Like it's just kind of opaque to you. And if you can't reprompt it to get what you want, you're kind of stuck.”
Liu: Airtable employees may cancel all meetings for a week to test AI
“If anyone wants to just literally block out a day or frankly, even a week and like have the ultimate excuse, like you can use, like, you know, you can say that I told you to do it, right? Like if you want to cancel all your meetings for like a day or for an en…”
Liu: Multidisciplinary hybrid roles have a major advantage over specialized tech roles
“There's a strong advantage to any of those Three roles who can kind of cross over into the other two, right? Like kind of the hybrid unicorn types, right? So if you're a designer who. Can be just technical enough to kind of be dangerous and understand a little…”
Liu: OpenAI launched a ChatGPT Deep Research API costing roughly $1+ per call
“There is now. There is now... It ends up being and I think they only recently exposed it. It ends up being like something on the order of like a dollar plus per research call”
Liu: Novel AI Product Discovery Should Start with Vibes, Not Evals
“For a completely novel product experience or form factor, you should actually not start with evals and you start with vibes, right? Meaning like, you know, you need to go and just kind of test in a much more open-ended way. Like, does this even work? Like, you…”
Liu: Airtable Is Building an AI Crawler Agent to Compile Entity Lists
“We're implementing this new capability that can use basically a long running AI crawler agent that goes and researches the web, you know, for a specific type of object or entity. So it's a little bit different from deep research, similar to deep research, but …”