Are foundation model labs actually building better apps
Are foundation model labs actually building better apps
Description
Startups should target niche markets to avoid competition from large AI foundation model companies
Transcript
what we'll end up seeing is that if you're in one of these big proven high margin, high demand markets, like coding, so you got you're competing versus like Codex or you know Claude Code, if there's another vertical that ends up having that level of validation, that level of margin, that level of usage, you're going to have a quick, you're going to, you're going to boom, but then you're going to bust, because they're going to come for you and they're going to realize that's the next vertical that has open AI and Anthropic, we have to go after and they will win it. what I also believe though is there'll be enough smaller verticals that OpenAI and Anthropic will just never have the compute to be able to go after. And if you're in one of those, and you can maximize the models to be able to deliver an amazing and distribute an amazing product in one of those smaller, niche ones, yeah you can win and you can then, you know, create something that over time becomes, becomes great. Yeah. But it's, it's the tale of Icarus man, be careful of flying too close to the sun if you're building one of these, these products built on these these AI models. It is an interesting one because it makes me think why exactly are the model companies able to build better applications or software than the application companies like why is everyone using Claude Code or Codex versus a company that focus focuses in, cause like really all, all you need is the model. What, why can't they build something like better than them? Like I understand they got the model but, does it, does it just come with the hype, like oh, because I want to use Claude, I would just use like Claude.

