OpenAI CEO Sam Altman believes AI is dismantling one of the oldest barriers to entrepreneurship - the need for technical skill. In his latest conversation, Altman argued that founders no longer need to know how to code to turn an idea into a working business, as long as they understand their customers well.
Coding is no longer the gatekeeper
For much of the startup era, an idea alone rarely went far without an engineer to build it. Altman, in an interview with Axios, said that dynamic is now reversing, with AI tools capable of handling much of the technical heavy lifting that once required a co-founder or hired developer. People who are skilled at spotting problems and reading customer needs, he said, can now build and ship a product largely on their own.
Reflecting on how the industry once viewed such people, Altman admitted, "We would make fun of these (idea) people." He added that the shift now favours them. "All of a sudden it's like the revenge of the idea guys." He was careful to note that technical talent still matters, but said he's now willing to back founders who 'can't code at all' provided they have a sharp read on their market.
One-person companies become more realistic
Altman also said AI is changing how small a business can be while still functioning effectively. Tasks that once needed a team with varied specialisations - from support to operations - can increasingly be handled by AI systems, letting a single founder or a very small team run more of the business themselves.
He described this as one of the most favourable moments in recent memory to launch a startup, pointing to OpenAI's GPT-6 Astra model as an example of AI capable enough to help someone "start and run a small business effectively." Altman compared such models to having a bench of capable employees spanning multiple functions.
AI makes niche business models viable
According to Altman, AI is also opening the door to smaller, more specific business ideas that wouldn't have made financial sense earlier. He cited an example of software built for a narrow customer base - sold to around 50 customers at roughly $500 a month - as the kind of venture AI now makes practical.
He extended the same logic to game development, saying AI can cut the cost and effort of building games meant for a small, specific audience, potentially making such niche projects commercially viable for the first time.
