Every major computing wave follows a familiar pattern. First come new primitives. Then come tools built around those primitives. The more important shift happens later, when products emerge that make the underlying complexity usable for far more people.

Glif is a good example of that shift.

When we first invested in Glif alongside a16z and USV in its US$17.5M seed round, what stood out was not simply that Fabian and Jamie were early to generative AI. Many smart teams were early. What stood out was that they seemed to understand where the category was likely going.

The long-term opportunity in AI was never going to be just access to models. Model capability improves. Costs fall. More of the underlying intelligence becomes abundant. When that happens, value tends to move up the stack.

In creative AI, the key question is not only what a model can do. It is whether a product can turn fast-moving model capability into something coherent, useful, and repeatable for real users.

That is a much harder problem than it appears.

With Glif, the interface becomes much simpler for the user while the system underneath becomes more capable. Most users do not want to manage orchestration. They do not want to think about which model to call, which tool to invoke, or how to stitch together multiple steps to get the result they want. They want to express intent clearly and have the product do the hard work underneath.

We believe many of the strongest AI products will be the ones that absorb complexity without reducing power. They will coordinate models, tools, and workflows behind the scenes while giving users an experience that feels simple, natural, and increasingly capable.

This is one of the distinctions we pay attention to as investors. In every platform shift, weak products expose system complexity to the user. Strong products hide it in the right places. The product becomes more powerful precisely because the user does not need to manage the machinery directly.

That is easy to say and hard to build.

It requires technical judgment, product taste, and a clear view of which abstractions are temporary and which are durable. It also requires understanding that commercialization is not something that happens after the product is built. In AI, as in deep tech more broadly, commercialization is often embedded in product choices from the beginning.

As models improve and the cost of intelligence continues to fall, this layer will matter more, not less. Glif is Claude Code for AI Creatives, and it shows what that future can look like.

-Brandon Zhao and Allen Lau

P.S. The Glif team asked Glif to create a brief ad about Glif, in one prompt, here it is: