Lenny's PodcastAlex Komoroske: Why gardeners beat builders in the AI era
Through slime-mold metaphors, the adjacent possible, and squishy LLM duct tape; Komoroske's strategy salons grow taste, kayfabe-aware bets, and bits and bobs.
CHAPTERS
- 0:00 – 2:02
Alex Komoroske’s background and what this episode will cover
Lenny introduces Alex Komoroske, his unusual first-principles approach to product and systems, and the episode’s major themes (AI, gardening vs. building, slime molds, adjacent possible, productivity, and life philosophy). This sets expectations for a wide-ranging, metaphor-rich conversation.
- •Alex’s career: Google (Chrome, Search, Maps AR) → Stripe → startup reimagining the web for the AI era
- •Episode topics preview: LLMs, PM skills, organizational dynamics, emergence, metaphors, community-building
- •Framing Alex as an original thinker whose ideas rewire how people see product work
- 2:02 – 6:19
“Bits and Bobs”: a 600-page public reflection practice
Alex explains his long-running “Bits and Bobs” Google Doc: a personal, daily note capture and weekly synthesis ritual he later made public. He emphasizes it’s primarily for self-reflection, intentionally imperfect, and not optimized for easy consumption.
- •Daily capture of interesting ideas; weekly time carved out for reflection and pattern-finding
- •Public sharing is a byproduct; he resists pressure to reformat into social posts
- •Deliberately “illegible” at times to preserve authenticity and reduce reader expectations
- 6:19 – 7:39
The Compendium workflow: fast capture, embeddings, and weekly distillation
Alex describes his internal toolchain: rapid note-taking during the week, processing into a personal knowledge system (“Compendium”), and using embeddings to surface related past ideas. He then filters for what still resonates and publishes distilled insights weekly.
- •Compendium: tens of thousands of working notes; cleaned up with added context
- •Embeddings-based similarity helps rediscover related ideas from prior notes
- •Weekly resonance filter → export → weekend distillation → Monday publishing cadence
- 7:39 – 10:10
Creating thinking time: why reflection makes you more effective at work
Alex argues deep thinking requires intentional space; otherwise “mundane bullshit” expands to fill all available time. He shares how protected time (e.g., no-meeting Fridays) leads to clearer strategy, reusable documentation, and better decisions.
- •Deep thinking needs slack; without it, you lose the chance for non-obvious insights
- •No-meeting time enables writing docs that prevent repeated 1:1 explanations
- •Reflection can reveal alternative approaches to problems you’ve been stuck on
- 10:10 – 14:26
LLMs as “magical duct tape”: disruptive shifts in cost structure and playbooks
Alex frames LLMs as genuinely disruptive: they break assumptions that software is expensive to write and cheap to run. He warns that old playbooks don’t map cleanly and that many AI products fail because “squishy” systems can be great most of the time but catastrophically wrong sometimes.
- •LLMs undermine “expensive to write, cheap to run”—affecting product economics and business models
- •Inference costs change what monetization is viable (e.g., ads may not clear costs)
- •Designing with probabilistic failure modes: ‘99% fine’ can still be unacceptable if failures are severe
- •Build with LLMs as enabling glue, not as an oracle that must always be correct
- 14:26 – 19:03
Skills PMs should lean into: curiosity, play, and “taste” in a world of slop
In the early phase of AI, Alex advises exploration over rigid execution—treating the moment like community gardening rather than factory farming. As content production becomes cheap and noisy, differentiation comes from taste: a distinctive perspective that resonates with others.
- •Early AI era reward: experimenting, playing, and noticing weird/alien new affordances
- •WebSim as an example of a product that only makes sense in an LLM world
- •Taste = what you produce that’s distinct from average/LLM-default output
- •Taste must be individual and also compelling to a broader audience
- 19:03 – 23:05
AI’s hidden impact: making individuals better more than making teams collaborate
Alex explains why AI’s value can be undercounted: many benefits accrue at the individual level, often quietly. People may not broadcast that they’re using AI to be faster or better, so organizational adoption can appear slower than it really is.
- •AI boosts individual throughput; usage may stay ‘below the org’s awareness’
- •Incentives to hide usage (fear of job/compensation repercussions)
- •Example usage: frequent conversations with Claude as a thinking partner
- •LLMs as ‘electric bikes for idea spaces’—cover more ground faster
- 23:05 – 31:48
Organizational kayfabe: the fake-real stories companies must maintain
Alex introduces “kayfabe” (from pro wrestling): everyone knows something is partly fake, but behaves as if it’s real to keep the system functioning. He shows how status reporting and power asymmetries compound small distortions into leadership-level delusions—and why “hitting the ground truth button” can be punished.
- •Kayfabe as lubricant: optimism/enthusiasm can be necessary to execute anything
- •Status color distortion compounding across layers → leadership far from ground truth
- •Power asymmetry: ‘you can’t make your boss look dumb’ drives emergent dysfunction
- •Zombie organizations arise when people internalize the kayfabe to avoid cognitive dissonance
- 31:48 – 34:47
Thinking like a gardener, not a builder: planting cheap seeds for compounding outcomes
Alex contrasts builder mindset (plan → manipulate reality) with gardener mindset (create conditions for growth). The goal is to ‘farm for miracles’ by planting many low-cost bets with compounding potential, then watering what starts to grow rather than trying to predict winners upfront.
- •Builder approach is linear; gardening seeks non-linear, self-accelerating growth
- •Plant many cheap ‘acorns’ with upside; don’t over-analyze which becomes the oak
- •Look for ecosystem-shaped opportunities: network effects and reinforcing loops
- •Protect early seeds from ‘squirrels’ (organizational disruption) with credibility and cover
- 34:47 – 41:50
Designing for emergence inside companies: legitimacy cover + 30% for weird bets
Alex explains how to make emergence viable in a corporate environment: ensure the team is obviously valuable to avoid existential scrutiny, then allocate a meaningful minority of capacity to experimental seeds. He also shares how leaders can help people lean into superpowers to unlock great work.
- •Emergence-oriented work can look unserious and be dismissed as ‘luck’
- •Tactic: 70% clearly-legible value + 30% seed-planting experimentation
- •Encourage junior bets that are low-risk but high-learning; protect them with trust
- •Mentorship: name people’s superpowers early to unlock growth and resilience
- 41:50 – 45:21
Slime molds as an org design metaphor: coordination costs and choosing your tradeoffs
Using his famous slime mold deck, Alex describes how coordination costs explode as organizations grow—and why pretending a big company can steer like a startup leads to dysfunction. Leaders must choose between coherence (big rig, high process) and autonomy (swarm of sports cars), accepting the downsides of each.
- •Coordination cost grows rapidly with group size; alignment becomes the bottleneck
- •Big orgs must ‘drive the vehicle they have’: pivot less, add process/slack when needed
- •Alternative: split into autonomous swarms; accept visible incoherence as a tradeoff
- •Slime molds illustrate emergent problem-solving when you embrace distributed agency
- 45:21 – 47:56
Metaphors and resonant ideas: how Alex tests frames across diverse audiences
Alex reveals his method for developing sticky metaphors: external processing through conversation, noting what sparks ‘aha’ reactions, and iterating the phrasing. He looks for ideas that resonate across different professional ‘sub-graphs,’ suggesting broad applicability and a larger potential audience.
- •External processing: discover what you think by talking, not just introspecting
- •Track resonance signals (‘ooh’ moments) and refine the formulation over time
- •Cross-audience validation: if sales and engineering both resonate, it’s more universal
- •Metaphors invite listeners to connect the last dot—useful for conveying sensitive truths
- 47:56 – 52:21
Strategy salons / nerd clubs: building an emergent idea lab inside your org
Alex outlines his “secret weapon”: optional, norm-driven groups designed for collaborative debate and playful exploration. By controlling membership, reinforcing “yes, and” norms, and slowly introducing new perspectives, these clubs become high-trust spaces that produce unexpected, high-leverage insights.
- •Make it opt-in and slightly subcultural to attract intrinsically motivated participants
- •Explicit norms: collaborative debate, “yes, and,” use ‘I wonder…’ to reduce defensiveness
- •Add new perspectives gradually to protect norms and avoid a single ‘jerk’ poisoning the space
- •Outcome: a low-stakes environment that stochastically spins off game-changing insights
- 52:21 – 1:01:41
Community-building mechanics and productivity/self-control tactics
Alex gives a practical playbook for sustaining communities (momentum, quorum, and lightweight facilitation) and then shifts into personal productivity: using constraints to reduce activation energy and creating strict “always rules” to improve self-control. He shares examples like daily Peloton streaks and using small wins to generate momentum.
- •Community momentum: start tiny, avoid cavernous spaces, and always ensure quorum
- •Facilitator role: prune tone, seed posts, amplify good sharing so norms self-sustain
- •Productivity: use tight time boxes to overcome activation energy; chain small completions
- •Self-control: “always rules” beat “sometimes rules”; streaks create commitment inertia
- 1:01:41 – 1:24:37
Life philosophy, awe, and the adjacent possible: navigating uncertainty with a North Star
Alex shares principles for living and working well: do things that give you energy and you’re proud of; beware the “Hallmark card fallacy” of dismissing truths you weren’t ready to hear; and cultivate awe to accept disconfirming evidence. He closes with the “adjacent possible” as a strategy model: combine a plausible long-term North Star with safe next steps that steadily reconfigure what becomes possible.
- •Life motto: energy + pride → compounding personal effectiveness and integrity
- •Hallmark card fallacy: profound truths sound trite until you’re ready to internalize them
- •Awe/curiosity enable learning from disconfirming evidence rather than resisting it
- •Adjacent possible: small, safe steps + evolving context; pair with a low-res 3–5 year North Star
- •Avoid false precision in strategy; embrace uncertainty while staying directionally coherent
