And I'm curious what might've changed in pathology or an AI that made it necessary to rethink the platform from the ground up instead of continuing to add new features.

Yeah, it's, it's a great question and it's something that we've, we've given a lot of thought.

We didn't just wake up one day and say, we gotta build the fifth generation concentric and we're gonna whiteboard everything and then build it for six, 12 months and then release it.

This was an evolution that was ongoing for many quarters, probably one or two years.

We were seeing technology advancements, for example, large language models, image foundation models, some of which were generalized, some of which were pathology domain specific.

Mm-hmm .

and just seeing what people are publishing in the literature and realizing that we're now seeing an explosion of capabilities that previously would have taken months to years to build in a very high-risk way.

How do you reconcile those two things?