Andy Halliday

There are two suitors.

One of them is Meta, which has done a bunch of moves and reversals and poaching.

And it seems like that as soon as somebody joins them, half of them run away, run away, run away.

It's so dysfunctional.

And yet, it's very powerful.

And on the other hand, you've got OpenAI, which has continued to be, I think, a magnet for really high-end aspirational leadership in what's happening in AI.

And it seems that it's a very...

motivating and intellectually stimulating place to work because there are so many elements to it and all of them are doing incredible things.

Now, you're saying, where am I going to go?

And it's such an obvious choice.

I mean, why would anyone take a personal risk in their sort of choice of partner and go to meta at this point, just because of all of the red flags that have emerged.

So I just wanted to say that, but now I want to switch.

I want to go ahead and comment on that, but I won't talk about Google MCP, the web MCP thing, how it relates to agents and,

And why that's also a very important announcement.

It just seems like... Yeah, there's a vibe there, right?

Okay, so WebMCP, it was announced by Google, okay?

And this is moving us closer to the ability of agents to work kind of seamlessly across the web on websites that you currently...

you know, navigate to and employ.

So what's the difference?

Like we've talked on this show before about how the early browser agents, Comet, Atlas, et cetera, were doing what amounts to pixel analysis to try to figure out how to operate a website and then would take actions at the pixel level on the screen.