Brian Gerkey

So there are lots of different ways you could tackle robotics and try to provide some kind of value to customers.

We think about it from the perspective of there's pretty good hardware out there that can handle a lot of applications.

Not all applications, but I like to say that in many cases, the hardware is willing and the software is weak.

There are...

You've got really great robot arms, really great sensors.

And what's happened over the last several years is if we think about manufacturing here, so there's lots of different robot applications.

At Intrinsic, we're focused on manufacturing.

So in that world, you've got pretty great arms.

You've got pretty great grippers, although that's an area where we could talk about what improvement looks like.

You've got pretty great sensors.

Now, historically, what's been difficult is writing applications that are able to deal with variability.

So the traditional approach to manufacturing has been, I mean, you see robots used in manufacturing all the time, but they're only used in places where you're producing in such high volumes

And you're doing the same thing again and again and again to where it's economical for you to like physically engineer out all of the variability so that the robots operate basically blind.

And so that what that I mean, that's worked well for those those domains, but that leaves so much else.

Like the latest I remember seeing a figure, I think, in the US, something like 80 percent of manufacturing facilities, if you just count them, have no automation at all.

And it's because many of them are smaller and they have a lot of variability.

They manufacture different things day to day.

They change the product that they're making.

They might be a contract manufacturer.

a manufacturing operation that's manufacturing something on behalf of somebody else.