Two Strategies. One Is a Trap.
Altman: There are fundamentally two strategies to build on AI right now. One — assume the model is not going to get better and build little things on top of it. Two — build assuming the models are going to keep getting better at the same pace.
It would seem that 95% of the world should be betting on the latter category. But a lot of startups have been built in the former. Then when we do our fundamental job — make the model and its tooling better with every crank — you get the 'OpenAI killed my startup' meme.
95% should bet on models improving. Instead, most build on today's limits and get steamrolled.
The 100x Filter
Altman: As an investor, ask the company whether a 100x improvement in the model is something they're excited about. We know the companies that come to us saying — when is the next model coming out? I want to be first to try it. It's going to be the best thing for my company.
Then there are companies we don't hear from on that. That's a pretty good delineation. If there's a clear path to how better underlying intelligence accelerates that product — they should be able to tell that story clearly.
Would a 100x better model excite them? The ones who can't wait for it — those are the ones to back.
The Medical Adviser Pitch
Altman: I talked this morning to an AI medical adviser company. They said — here's where the model's underperforming. But if it could get this much better on these metrics, we'd have all these other businesses. How soon can you do that? Here's how many people are dying every day you delay.
It was an effective pitch, actually.
Here's how many people are dying every day you delay. It was an effective pitch, actually.