LLenny's Podcast
← All frameworks
StrategyScott Wu (CEO and co-founder of Cognition)

Stickiness Over Moats

Stop chasing barriers competitors can't cross; build accumulating value they can't easily replace.

Difficulty
Moderate
Time to result
~ongoing to results
Steps
4
Confidence
85%

In fast-moving AI markets, Wu argues true moats (barriers that stop competitors from entering) rarely exist, so the right question is stickiness: once someone likes your product, is it easy or hard to switch away? Stickiness is built through value that compounds with use — accumulated learning and multiplayer network effects inside a team.

Origin

Scott Wu's reframing of the standard 'moats and defensibility' question, echoing Cursor's Michael Truell that defensibility is more like being the best product people can freely switch between.

Core principles

  • 01In AI, no layer has a hard barrier preventing competitors from entering
  • 02Reframe defensibility from 'can they enter?' to 'will users want to leave?'
  • 03Value that accumulates with use (learned codebase/context) raises switching cost naturally
  • 04Multiplayer effects — the whole team teaching the agent — deepen stickiness beyond individual use

How to run it

  1. 1

    Drop the moat framing

    Accept that in AI you likely can't build a barrier that keeps competitors out; stop building strategy around one.

    Watch out Betting defensibility on an impenetrable moat in a space where models advance weekly is a losing plan.

  2. 2

    Optimize for wanting-to-stay

    Ask whether a user who likes the product is excited to keep using it, versus finding it just as easy to switch and learn a new one. Design so the answer favors staying.

    Pro tip The best products win on being the best experience, not on locking users out.

  3. 3

    Build accumulating, compounding value

    Make the product get more useful the more it's used — learn the user's codebase, stack, and process over time so leaving means losing accumulated context.

  4. 4

    Add multiplayer effects

    Let value accrue across the whole team — shared context, agents onboarding new hires, colleagues chiming in on the same session — so stickiness is organizational, not just individual.

In the wild

Devin learning the team over time

Wu likens Devin to an engineer who's been at the company five years versus day one — it accumulates a representation of your codebase, stack, and process, and gains multiplayer value as different engineers teach it and review its PRs.

Switching away means abandoning accumulated, team-wide context — stickiness without a hard moat.

Common mistakes

Confusing stickiness with lock-out

Trying to lock competitors out of building, rather than making your own product compound in value, misreads how AI markets actually defend.

Is it for you?

Best for

Founders and product leaders building in fast-moving AI categories where model advantages are fleeting

Not ideal for

Markets with genuine structural barriers (regulatory, hardware) where classic moats do apply

From the transcript

I think it's often less about moes and more about stickiness

57:30

once you have a product experience that you really like, are you excited to keep using that experience

58:00

as you use Devon and as your whole team uses Devon

58:00

From the episode

Inside Devin: The world’s first autonomous AI engineer that's set to write 50% of its company’s code by end of year

Scott Wu (CEO and co-founder of Cognition)