Will OpenAI kill all startups? Why the answer is no

Executive overview

Founders fear that OpenAI and similar labs will absorb every viable AI startup opportunity. History says otherwise: every major technology shift — electricity, the internet, mobile, cloud — created more businesses than it destroyed, with startups disproportionately advantaged over incumbents.

OpenAI's stated goal is AGI, not building AI-powered CRMs or vertical tools. That gap is the opportunity. LLMs are at least as foundational as cloud computing or the App Store; the right response is to build real products, not cargo-cult the trend.

The founders best positioned to win are domain experts who recognise this as their moment, and CS-native builders excited by what's now possible — not people chasing fast money.

Why history favours startups in technology shifts

• Every major platform shift (farming, electricity, internet, mobile) increased the total number of businesses.
• The faster and larger the technology change, the more startups are advantaged over incumbents.
• Incumbents were disrupted quickly during the internet era; the same dynamic applies to AI.
• Someone smart right now is identifying real problems that couldn't be solved 6–12 months ago.

Cargo culting vs. genuine AI product building

• Cargo culting: adding "AI" to a pitch without it meaningfully improving the product or helping customers.
• Real AI integration drives measurable retention gains and makes products easier to monetise.
• The App Store analogy applies: early high-quality mobile apps were genuinely valuable, even amid hype.
• Cloud computing was also overhyped — yet dismissing it meant missing a fundamental shift.
• Fintech and crypto attracted builders chasing money; AI is attracting builders excited by the technology itself.

Two types of founders drawn to AI right now

• Domain experts in ML who have spent years on the foundation: they can now see the future they predicted arriving.
• CS-native generalists who treat LLMs as a powerful new tool and are racing to master it early.
• Both groups resemble early iPhone jailbreakers — learning something new before it was mainstream.
• Founders motivated primarily by fast money are notably absent, which improves the quality of the cohort.

Thinking in second-order effects

• The obvious AI plays (OpenAI competitors, thin API wrappers) are likely low-hanging fruit OpenAI can absorb.
• The real opportunities are second-order: Uber wasn't the obvious consequence of the iPhone — maps were.
• Starting as a thin wrapper is fine if it's a starting point, not a destination; Dropbox began as a thin wrapper on S3.
• "I can build this in a weekend" is a red flag from an observer, not a reason to stop building.
• The goal is to make something customers genuinely love, not to impress critics on the timeline.

Why building on existing LLMs is a structural opportunity

• OpenAI's first, second, and third priority is AGI — not building tools that improve people's daily workflows.
• That gap is permanently open until AGI is actually achieved.
• The original vision for OpenAI (as a YC nonprofit) was to be enabling infrastructure so other founders could create value on top.
• Using today's LLMs to improve lives and businesses is exactly what the underlying technology was designed to enable.