Bradley

That's absolutely correct.

And now we're getting, Pat, to the age of the AI agents.

Well, the AI agent is basically AI that can do something for you.

Automatically answering emails is a simple one that's available to everybody right now using Microsoft Pilot.

Certain emails, certain rules, it'll automatically respond for you with a simple answer or find an open place on your calendar.

That's an AI agent.

What he's talking about is the advance in the computational power and the library, also known as LLM, large language model, that it has to use.

that you could have a bad actor learns how to build a nuclear bomb much quicker or learns how to deploy a biotech poison much quicker.

And that what's happening is that now the agents and the collective intelligence behind the agent is sort of now the third, the fourth wave.

Chip, data center, agents, and now collective processing power and super AI powering used for bad.

Okay.

You can look back in history and saying, oh my gosh, the folks that invented the nuclear weapon, we need to regulate and control this.

We need to figure this out.

We're about to go to what will be the next wave, which is levels of control through oversight.

So if you think that Google listening on your device or your Alexa device to sell you ads is spooky, don't.

The back end of AI is going to know who's looking for what, and the government is going to, by default, have to be snooping on that to see who's using AI to find out what is the best stream in America to put a viral poison in to kill the most number of people based on snowpack, runoff, melting, and the speed of the stream.

People are going to be asking those questions.

Where's the best place for a suicide bomber to kill the maximum number of people in New York City on Black Friday?

Oh, well, it'll look at traffic.

So the government's going to have to see this, and that's what he's talking about.