the sidewalk. But they had to do a test where they put mannequins on the sidewalk like homeless people who had passed out from fentanyl. And then they had to test the robot to see if it would go around the bodies, the homeless people lying on the sidewalks. And I think it did, but we're just assuming from the video that what they were testing was if the robots could go around the humans. What we didn't see is if they're testing the robot to kill the humans and then go around them. I don't know. We just saw the bodies. We don't know how they died, but it might be the first step in the robots destroying all humanity.

Speaking of that, Perplexity, the AI company that I've said good things about in the past — it's a hot little AI company. I don't know if it will survive because it's not one of the top three. If you're not in the top three I don't know if the government will even let you be successful. So maybe Apple will buy them or something. But Perplexity has rolled out its own browser called Comet. And I saw one report on Tech Radar by John Anthony who tested it and he was impressed.

Now the reason that an AI company would want to make its own browser is so the browser could act as your agent and do a bunch of stuff while you're doing whatever you want to do. So as John Anthony points out in his article, when you first start using it, it doesn't seem like it's a big time saver because you tell it to do something and then you sit there and watch to make sure it did it. So it takes about, you know, it's not really that much of a time saver because you're prevented from doing something else because you're watching to make sure it worked. But as he points out, you eventually reach a point where you realize, wait a minute, this usually works. So I'll just tell it to make some reservations or whatever you're telling it to do, and I'll do some work on something else in the meantime. And apparently that is a big wow experience because when you get back to it and it did what you wanted, you saved a lot of time.

Now here's my take. I've been saying this for probably 20 years or more in public and people always fight me on this and you will fight me too. So get ready to fight me. Here's what I believe. I believe that humans want to maintain privacy as much as possible. So far we're on the same page. Everybody likes privacy mostly about themselves. We don't care about privacy for other people but we like it for ourselves. But in order to use and really get all the benefit from these AI agents, they're going to need to have your passwords and they're going to need to know all about you because the things that you want them to do, such as making reservations or whatever, it's going to require a lot of knowledge about you — your name, your address, in some cases maybe your social security number if you're telling it to deal with something financial or banking or whatever.

How many of you would feel comfortable giving your own AI — if they told you, "Oh, it's totally secure. We'll never look at it at the company level" — how many of you would feel comfortable with an AI that has your password for anything important? Maybe for making restaurant reservations, maybe. But would you let it have access to your bank? Because if you did, think of all the time you would save. I would love an AI that had access to my bank because I'm continually signing up for things that are autopay or debugging some problem with my bank or there's a credit card that got stolen. Just always, I'm just always dealing with bank and IRS stuff. How many of you would trust your AI to have access to all of that stuff? The answer is none of you, not a single one of you, and certainly not me, would feel comfortable with it.

Now here's the part that I've been predicting. We'll all get over it. We'll get over it. Once you realize that privacy was always an illusion, you're going to let it go. It's always been