Stop Vision AI Failures

So what we, what we mean by infrastructure software is effectively we aren't necessarily a single use case or single app computer vision/vision AI company.

There are thousands of vision AI startups.

Ninety-nine percent of them, they kinda do a use case.

So Flock, for instance, is in the news a lot right now.

Their use case is they are what's called an ALPR, an automated license plate reader.

Not, you know, people have very strong opinions about that as sort of a tech category, but that's, that's really all they do.

There's other companies that try to catch shoplifters, and that's all they do.

And there's other companies that try, uh, to s- to see, you know, if you've got enough inventory on your shelves, and that's what they do.

And, and that's the majority of the, all those thousands of startups.

So they're all like, "That's all I do.

We, we have one use case.

It might be relevant to one or two verticals.

We built a product that does that one thing." What we are trying to do is, is not build really any apps or any use cases.

We're trying to build the software that's underneath all of them, so we actually want all those startups to be our customer.

And that's, that's, again, like I said, I promised you I'd answer the other question as well.

That's, that's literally why I joined.

So Kit was the founder, Kit Merker.

He was on the founding t-team at Google of Kubernetes.

Mm.

Kubernetes isn't an application in and of itself, but almost every application in the world at this point in time, at least if it's wasn't written in the '80s, anything that's been written recently, you know, Kubernetes is in there.

That...

And that's what infrastructure software is.

And I worked at Canonical, and our, our, our flagship product is Ubuntu, which is the world's most popular Linux.

Now, there are people who, like, use Linux as an application.

They, you know, they, they run their laptop on Linux, or they run their servers at home on Linux, and they use Ubuntu.