I'm not mixing them up because I don't know the difference.

But one of them, and I'll put in the notes, I believe it was Gartner.

had a research report that they published after they did an objective analysis using an MIT framework, I think it was, and perhaps some other stuff, but just a very objective analysis of, I think it was at least 1,200, if not 1,300 plus technology vendors that suggested via their marketing and otherwise that they were agent or agentic.

And just to kind of steal the thunder of the report, if you want to guess, you can guess, but it was 1% of those technology vendors that passed a very simple and objective definition of what is or isn't agentic.

I only bring it up to be hopefully somewhat funny, but also just to remind us that we're all trying to figure this out as we go and the world's moving really fast.

So the reason I mentioned that is that also something that's automated versus autonomous versus the ability to think versus the ability to have a human in a loop and on the loop and all this other stuff, there's so many considerations that you really need to work through before you really start stepping into The most important question to start, which is why am I doing this? Right? What is the point? So I think it's incumbent upon all of us, including the deeply technical rock stars that I love spending time with is like, let's not talk about the technology right first.

Let's just talk about what are we trying to achieve? What are the business outcomes? And so I find the people that start there tend to be the people that have a lot more success in making foundational choices and decisions around what type of infrastructure, let's call it data infrastructure slash compute, and then data modeling infrastructure, which steps into the Snowflake and the Databricks environment, et cetera, and of course, TACNode, before then getting into the area of, okay, what does the orchestration layer look like? What kind of a relationship do I want to have directly or indirectly with the models and frontier stuff? And then do I want to build one agent? Do I want to build a multi-agent framework? How specialized are they going to be? What's the relationship between the agent and the human and in the loop, on the loop, et cetera, et cetera.

and familiar with the work of management