Martin Casado

Yes.

Uh, uh,

And so I'm just using, I'm using just like Python browser automation.

I'm actually using Selenium.

And then I'm actually, there's actually a skill that emits just exactly what you want to know so that the agent is not reading HTML and consuming all those tokens all the time.

And then at Fivetran, for example, we are working right now on a Salesforce administration agent to,

to integrate, to basically do continuous integration of small changes into our Salesforce, which is very labor intensive right now.

And it also does not use browser automation because the Salesforce CLI we found does everything we need to do.

We'll use browser automation if we need to, if we come to something that can't be done with a Salesforce CLI.

But to their credit, the Salesforce CLI is quite comprehensive.

Pretty much anything you can do in the UI, you can do with a CLI command.

And the agents seem to already know how to use it.

So can I do a $1 bet with you for when we're on this podcast in five years?

Yeah, I think they'll just hit the APIs.

You know, in theory, it seems like they're an unnecessary layer.

Because these agents are great at calling APIs and calling command line tools.

So why bother having this other layer?

In practice, when you sit down and actually try to build systems, MCPs do solve important problems, particularly authentication and just like discoverability of what's available.

So even though you can say kind of from first principles that maybe they shouldn't exist, when you actually sit down and write a real system that's accessing context, you almost always end up sticking an MCP server into it.

Yeah, and there's just also a lot of little affordances in the AI tools that are built around MCP, like user...