[laughs]

So Magentic, multi-agentic.

Um, it started off, as I said, um, thinking about how agents can work together to drive, uh, individual applications.

Um, and so like our initial experiments in this space, um, we started with, um, uh, what we called Magentic 1, which was like a team of agents that could, um, operate and complete complex tasks on the web.

And that type of system, you know, had an orchestrator agent that maintained some short-term memory and could direct other agents which had different skills.

So we had a web surfer that could browse the web, file surfer that could interact with files, et cetera.

Um, then, uh, once, you know, we, we got that working to some degree, like we started thinking about, okay, like if this is gonna be used in the real world, um, humans will have to be overseeing these agents.

Do you think, do you think that agents are now at a very rudimentary form? How do you think they'll, they'll evolve in that sense?