ing it as opposed to testing it every time you use it. So they're already pre-tested but not against the AI that's going to be there next week. So whatever this rule is that you can't learn it as fast as it changes and makes your learning useless, has that ever happened before? Is this the first time in history that the rate a human could learn a thing is too slow to use the thing because the thing will never be the same thing you learned?

Now I'll tell you the first time I learned this because this happened to me in my career once. I used to have a job in a laboratory in the phone company, one of my first jobs, to test brand new equipment like new types of phones to see if it would work with our digital products especially. And so we would get all these new products. You know AT&T would make a new phone or whatever and we'd plug it in and it would take us a few weeks to test it all and everything. And then I would write up a report so that the customers could know whether to buy this equipment and whether or not it would work with the phone company. Because you didn't want to buy equipment unless the phone company was compatible. And things were changing quickly. So it would take me a few weeks to test things, get them in, put out my report. And do you know what would happen? By the time my report went out every time, the hardware that I tested had a firmware upgrade in the meantime. What good was my test? The test had no value the moment it was published. Had no value. Because without the software or firmware of that product you've tested nothing. Because we know it physically connects. Like I wasn't testing to see if the plug goes into the hole. I was seeing if the so