Sam Altman on what most startups and VCs are getting wrong about the future of AI

“I think there are two fundamental strategies to build an AI startup right now. You either bet the technology is about as good as it’s going to be, or you bet that the technology is going to get massively better.”

He gives the example of building an AI tutoring company.

One path is to build a product that is only capable of serving 6th graders today but will eventually be able to serve 8th graders, 10th graders, and even PhD students as the base model gets smarter.

The other path would be to assume the base model won’t get much smarter and spend all of your effort handling edge cases to make your product work for 8th graders in a few limited subjects (e.g. history) right now.

“You’ll be really happy when GPT 5 comes out,” Sam says, if you assume the base model gets much smarter. And if you assume the model basically stays the same, you will be really sad.

He continues:

“My intuition would have been that 95% of entrepreneurs would [bet the models get smarter]. But it seems like 95% of entrepreneurs are [not]. And then you have this whole ‘OpenAI killed my startup meme.’.. If you’re doing a little thing to get it to barely work in one specific case, that’s probably a mistake.”

Full video: Harvard Business School “A fireside chat with Sam Altman OpenAI CEO at Harvard University“ (May 2024)

P.S. We’ve put together a YouTube playlist with every Sam Altman insight we’ve ever shared. You can watch it here: "Best startup advice from Sam Altman"