“You're going to run out of smart people to do stuff. That's my biggest concern with this whole thing… Now we're not training the new joiners. They're not getting their first bit of experience. And then five years down the road, who's going to be creating the content for the models to train on anymore? We've killed the information for the input side of the whole machine.”

…was done on stuff in Stack Overflow. So, when you disagree with the results, you're like, why did it do this? You're like, oh, because it's because of Stack Overflow. Of course, that's all the crazy stuff I used to find on there that I copy and pasted it there. But I mean, that I learned from.

…which is why there was getting better at the time. By the time you hit, I wanted to say he said, like, if they kept going up by like 0.5, he said, like, by 5.5 or 5 or 6.0, it was not going to, it was going to have more data than required more training down than the whole world had ever generated.

And I try to make, I try to say like, well, I think you need to understand the basics to understand, you know, the next gen stuff, aka, you know, in the cloud, it was always like, you need to understand the networking before you can understand, you know, how to do build a VPC and routes and…

So, I mean, like V2 of the software engineer might just be that the agents, instead of training upon the data that, you know, stuff that we've built over the last, you know, 20 years, now they go through and say, hey, maybe this isn't even a good way to do it. And they start defending against the…

There was an interesting thing I saw online, not to keep going down random tangents. Someone write to analysis and gave AI, I assume it was the same AI, I'm not sure, honestly. And it's one of those things I was reading at night as I was falling asleep, but it was they gave AI the same project and…