AI Is Changing the Rules of Entrepreneurship

Victor P. Seidel, Bret Greenstein, Thomas H. Davenport · article · 2026-09-09 source ↗ Reader 🖥️

practical · prescriptive trend-analysis essay (HBR feature) built around a coined-term framework (a six-dimension comparison table contrasting 'lean startup' with 'abundance entrepreneurship'), illustrated with named case studies and cited surveys, closing in normative advice for founders and incumbent-firm managers — Three-author byline mixing an academic voice (Seidel, entrepreneurship scholar) with practitioner/consulting voices (Greenstein, Davenport); argument leans on interview anecdotes and secondary citations (HBS working paper, YC data, a16z-adjacent survey, news coverage) rather than original research.

Structure

Key terms

• abundance entrepreneurship Abundance Entrepreneurship
  The authors' coined label for an AI-enabled entrepreneurial model in which a single operator, aided by autonomous AI agents, generates and runs many low-cost ventures in parallel, characterized by capability surplus rather than resource scarcity.
• lean startup
  Eric Ries's/Steve Blank's resource-scarce methodology (capital, MVP, customer development, iterative pivoting) used here as the historical foil that defines abundance entrepreneurship by contrast.
• illusion of progress
  False signals of business traction generated by AI outputs (simulated interviews, generated content, plausible-looking artifacts) that mimic real market validation without constituting it.
• signal filtering and judgment
  The comparison table's replacement 'key skill' for hypothesis testing: the capacity to distinguish genuine, scarce market signal from pervasive AI-generated noise when choosing among many parallel options.
• solo unicorn
  A company (here MEDVi) reaching unicorn-scale revenue/valuation while run by essentially one or two people via AI -- used descriptively, then undercut by the article's own reveal of MEDVi's regulatory failure.

Propositions

• explicitAI has drastically reduced constraints on both information-related resources (access to expertise and data) and labor-related resources (idea generation, analysis, task execution), shifting the primary entrepreneurial constraint from capital and access to founder attention and selection.conf: high
  extendsPodcast: The AI Joy Gap: Why Some Developers Thrive While Others Struggle · p6 — Both make the same bottleneck-relocation argument -- engineering/idea-generation productivity surges, so the constraint moves downstream to discovery and judgment -- but this piece applies it to solo-founder entrepreneurship rather than product organizations, generalizing the pattern to a new domain.
  supportsAndreessen Horowitz's Post · p2 — a16z's claim that leaner AI-augmented teams can execute against far larger visions ('more ambition, not more capital restraint') is the same constraint-shift this piece documents at the level of the solo founder and AI-run organization.
• explicitThe lean startup methodology (Ries, Blank, Toyota lean manufacturing) was fitted to a resource-scarce environment: capital, an MVP, extensive customer discussion, and iterative pivoting by small teams.conf: high
• explicit'Abundance entrepreneurship' differs from the lean startup across six dimensions: constraints, typical actors, experimentation logic, speed, organization, and key skills.conf: high
  extendsThe Agentic P&L: Beyond the Empire of Headcount · p16 — The table's 'ORGANIZATION: multi-agent systems' dimension and the abundance-entrepreneurship model of a founder plus coordinating AI agents is a solo-founder instance of the '3+N squad' structural primitive that piece proposes for the agentic enterprise generally.
• explicitThe entrepreneur's role is shifting from executing tasks and managing workflows to defining goals and constraints for an AI-run organization, as emergent platforms let a founder assign CEO/engineer/marketer roles to autonomously coordinating AI agents.conf: high
• explicitcontestedA study of four years of Y Combinator batches found AI-native-built startups were 25% smaller overall, with 13% more engineers but 15% fewer entry-level workers and managers.conf: high
  contradictsThe AI paradox: More automation, more humans, more work | Dan Shipper · p4 — This piece's YC-study finding that AI-native-built startups run 25% smaller (fewer entry-level workers and managers) sits in tension with Dan Shipper's report that hiring at his AI-native company is going UP, not down (headcount doubled in a year) -- though the tension may resolve by growth stage: lean at YC-scale founding, hiring again once product-market fit and scale hit.
• explicitStartup formation has surged alongside AI abundance: 1.56 million new U.S. business applications were filed from November 2025 through January 2026 -- the most of any three-month period since at least 2004 -- and 17% of founders across a global sample of over 1,000 AI startups create more than one business at once.conf: high
• explicitAbundance entrepreneurship faces three named unresolved challenges: an 'illusion of progress' from AI outputs that mimic traction without being real, sycophantic AI models that won't tell founders their ideas or businesses are failing, and commoditization of startup ideas as many founders converge on similar AI-generated concepts.conf: high
• explicitBuilding a business quickly with AI, little industry experience, and few employees exacerbates regulatory and business risk, as illustrated by MEDVi's FDA warning letter for misleading marketing claims despite (or arguably because of) being run almost entirely by two people via AI.conf: high
• explicitToday's AI is not always capable of successfully running businesses autonomously: AI agent-managed vending machines made 'questionable and aggressive' pricing, marketing, and product-layout decisions, and observers question whether AI-generated ('vibe coded') systems are robust enough to scale.conf: high
  supports#3: How to Build an AI-Native Startup from Day One · p8 — This piece's failure cases -- AI-managed vending machines making aggressive, questionable commercial decisions, and observers questioning vibe-coded systems' robustness -- are concrete evidence for Turing Post's design principle that AI-native 'does not mean fully autonomous; it means machine participation where it pays, human review where it matters': both examples ran into trouble precisely by skipping the human-review gate that principle prescribes.
• explicitAccess to AI-supplied expertise does not eliminate entrepreneurs' need for genuine domain expertise, and this need is highest in regulated industries like financial services and healthcare.conf: high
• explicitFounders should think of themselves as designers of systems and processes, not just products; they still need a competitive moat (proprietary data, experience, or curated expertise) and should experiment only in moderation so their attention isn't spread too thin.conf: high
  supportsThe Founder's Playbook: Building An AI-Native Startup · p7 — Both name the same moat composition for AI-native ventures -- accumulated depth via codified expertise, integrations, and curated data/experience -- rather than being first or relying on the underlying model.
• explicitHuman attributes like taste, accountability, judgment, clarity, and persistence are 'not available in abundance' and remain the differentiators AI cannot substitute.conf: high
• explicitEstablished organizations must also adopt abundance-entrepreneurship practices -- managing a portfolio of options and accelerating internal innovation -- beyond what they already gained from adopting lean-startup principles, in order to keep pace with AI-native competitors.conf: high
• implicitThe authors treat 'abundance entrepreneurship' as a durable, structural successor paradigm to the lean startup rather than a speculative or hype-driven phase, despite documenting significant, unresolved failure modes (illusion of progress, sycophancy, MEDVi's regulatory failure, unreliable autonomous agents) within the same piece.conf: medium

Arguments

Narratives

• Tibo Louis-Lucas, a French solo founder, used AI to build, market, sell, and refine five different software businesses with only 10 employees, claiming over $1M/month in combined revenue.
  → Proof-of-concept illustration for the abundance thesis -- feasibility narrative, not evidence of typicality or sustainability.
• Solo entrepreneur Nat Eliason created an autonomous OpenClaw agent named 'Felix,' gave it $1,000 in starting capital and payments access, and tasked it with building businesses; per Felix's own public dashboard, it created three businesses generating over $200,000 in total revenue.
  → Illustrates the extreme end of 'founder as goal-setter, not executor,' foreshadowing but not resolving the governance and accountability questions raised by giving an AI agent capital and payment authority.
• An unnamed founder generated most of her company's code with AI, had satisfied customers, but worried those same customers could simply generate their own replacement for her application.
  → Illustrates the erosion of code-based competitive moats and shortened windows of competitive advantage under abundance conditions.
• MEDVi, a telehealth GLP-1 provider run almost entirely by founder Matthew Gallagher and his brother using AI for code, marketing copy, ad creative, and customer service, reached $1.8B in annual sales -- then received an FDA warning letter for misleading claims implying its compounded drugs were FDA-approved, plus other questionable advertising practices reported afterward.
  → Cautionary counter-narrative that complicates the celebratory tone of the earlier examples and directly supports the 'expertise still matters' argument.
• AI agents put in charge of managing vending machines made questionable, aggressive choices in pricing, marketing, and product layout (per TechCrunch and WSJ reporting).
  → Supports the claim that today's AI is not always capable of running a business autonomously and well.

Questions addressed

• ✅ What is replacing the lean startup as the dominant entrepreneurial approach in the AI era?
  'Abundance entrepreneurship' -- parallel, low-cost experimentation via orchestrated AI agents, with the founder acting as goal-setter rather than executor.
• ✅ If AI removes resource constraints, what becomes the new bottleneck for entrepreneurs?
  Founder attention and the capacity to filter signal from noise when selecting among many generated options, per the comparison table's 'key skills' shift.
• ✅ Does AI abundance eliminate the need for domain expertise?
  No -- expertise remains necessary, especially in regulated industries, illustrated by Arya Labs's continued reliance on Dobrin's domain background versus MEDVi's regulatory failure.
• ✅ Are AI agents currently capable of running a business autonomously and well?
  Not reliably -- the AI-managed vending machine examples made poor commercial decisions, and vibe-coded systems are questioned for scalability and robustness.
• ⭕ Is 'abundance entrepreneurship' a durable structural paradigm shift, or could it be a bubble inflated by unverified self-reported success stories and cheap idea-generation, given the article's own case studies (MEDVi, AI vending machines) showing serious, unresolved failure modes?
  Not addressed -- the piece asserts durability and frames the shift as a genuine new era without weighing its own counter-evidence.
• ⭕ What happens to the many parallel low-cost ventures that are generated but not pursued in a 'portfolio of options' model -- is there a societal cost (market clutter, wasted compute, customer confusion) that scales with the number of abandoned AI-generated startups?
  Not addressed.

Gaps

• illogicalMEDVi is used to support a general risk claim ('building them quickly with AI, little industry experience, and few employees exacerbates the risk') from a single, dramatic anecdote, with no base rate or comparison group offered -- how many similarly AI-run, low-headcount companies in regulated verticals did NOT run into regulatory trouble is never asked.
  Read next: Look for a systematic study of compliance/regulatory incident rates among AI-native versus traditionally-built startups in regulated verticals, rather than single-case reasoning.
• incompleteThe celebratory examples (Tibo Louis-Lucas's claimed $1M/month, Felix's claimed $200K in revenue per its own public dashboard) rely entirely on self-reported figures with no independent verification -- yet self-reported traction signals are exactly what the article later warns readers to distrust under 'illusion of progress.' The article never acknowledges applying a looser evidentiary standard to the stories that support its thesis than to the AI outputs it tells founders to be skeptical of.
  Read next: Independent reporting or financial verification of Louis-Lucas's and Felix's claimed revenues; corpus material on self-reported startup metrics and traction-signal skepticism.
• incompleteNo engagement with what happens to the failed or abandoned ventures in a 'portfolio of low-cost experiments' model -- market clutter from proliferating similar AI-generated products, customer confusion, wasted compute/energy, or effects on customers of ventures later abandoned by their AI-agent operators.
  Read next: Look for critical commentary on the externalities of AI-driven startup proliferation, or environmental/compute-cost analyses of mass parallel venture-generation.
• uninformedNo mention of legal or IP questions raised by AI-generated business assets -- copyright status of AI-generated code, branding, and marketing copy, or liability when an autonomous agent like Felix independently spends capital and enters payment transactions.
  Read next: Legal scholarship on AI agent liability and the IP status of generative output; regulatory commentary on autonomous agents holding and spending capital.
• incompleteSycophancy and the 'illusion of progress' are named as core challenges, but the only prescribed mitigation is a general exhortation to 'leave their computers for the real world of commerce' -- no concrete practice (adversarial-critic prompting, real customer-validation gates, third-party audits) is offered, even though this is the crux of the 'expertise still matters' argument.
  Read next: Corpus material on adversarial-critic prompting patterns explicitly names the same unresolved failure mode -- see extract-eating-my-own-dog-food-ai-autonomy-framework#g5 ('model sycophancy despite hostile persona, critique that is adversarial in tone but generic in substance').

Closes gaps elsewhere

• The Founder's Playbook: Building An AI-Native Startup · g5 — That gap flags the playbook's case studies as uniformly positive, missing failure cases (agentic technical debt, model regressions, token-cost blow-up, being out-competed by the model provider) -- a survivor-bias-by-curation problem. This article PARTIALLY closes it: it supplies genuine, named AI-native-startup failure cases (MEDVi's FDA warning letter from AI-run marketing without regulatory expertise; AI-managed vending machines making poor commercial decisions) that break the uniformly-positive pattern. It does not close the specific failure modes the gap named (technical debt, model regressions, token-cost blow-up, provider encroachment) -- it supplies a different, complementary category: domain-expertise absence and AI unreliability rather than engineering/economic failure modes.

Related documents

• #3: How to Build an AI-Native Startup from Day One
• The Agentic P&L: Beyond the Empire of Headcount
• The Founder's Playbook: Building An AI-Native Startup
  AI-Native Company Human as Orchestrator of Agents Moats & Defensibility Near-Zero-Cost Code
• What Survived Contact
  Human as Orchestrator of Agents Near-Zero-Cost Code Product Discovery & Experimentation Small Teams & Compression
• Andreessen Horowitz's Post
  AI-Native Company Moats & Defensibility Near-Zero-Cost Code
• 11) AI-spesial med Vilde Opsal og Anders Sveen
  Multi-Agent Orchestration Near-Zero-Cost Code Vibe Coding
• Steve Yegge's Vibe Coding Manifesto: Why Claude Code Isn't It & What Comes After the IDE
  Multi-Agent Orchestration Small Teams & Compression Vibe Coding
• What's an Orchestrator—and Why Does Software Need One?
  Human as Orchestrator of Agents Multi-Agent Orchestration Vibe Coding
• Your Work Team Is Now a 'Pod' and Your Co-Workers Are AI Agents
  AI-Native Company Human as Orchestrator of Agents Small Teams & Compression
• Why cultivating agency matters more than cultivating skills in the AI era | Max Schoening (Head of Product, Notion)
  Small Teams & Compression Taste & Judgment