The AI arms race gets a twist
Mira Murati’s Thinking Machines Lab is making its first move, and it’s not the usual “our model is bigger than your model” flex. Instead, the startup is betting on smaller, more customizable AI systems — basically the app-store version of the frontier-model world.
That matters because the AI market has started to look a lot like early smartphone land: a few giants set the pace, then everyone else fights over the features that make the product useful in real life. If Thinking Machines can offer something easier to tailor, it could chip away at the idea that only the biggest labs get to define the future.
Why investors should care
This isn’t just startup theater. It’s a reminder that AI competition is spreading beyond raw scale and into usability, flexibility, and deployment costs. That can pressure the moat narrative around the biggest players while also expanding the market for companies that help businesses fine-tune, host, or integrate these models.
Big picture
You’re watching AI shift from a heavyweight boxing match to a weirdly crowded neighborhood bake-off. The prize is no longer just having the smartest model — it’s being the one people can actually plug into their workflow without needing a PhD and a cloud bill the size of a mortgage.