To which I say, you mean common sense?

Apparently all it takes to lean right in today's world is common sense. Who exactly is against free market capitalism? Only the craziest leftists. I mean, otherwise Democrats and Republicans are pretty much in agreement that free market capitalism should be the basis of the country. How about property rights? Again, you'd have to be completely insane to be against property rights, but you know some leftists are. And what about minimal government intervention in the economy? Again, common sense. You want the fewest number of rules and government blocking, I guess.

So I think it's hilarious that what is called conservative stuff is just the thing that basically everybody who's not crazy is already in favor of. Just pure common sense stuff.

According to Wired Magazine, M. Kelly is writing about this. DOGE has a plan to rebuild the whole Social Security Administration codebase, so all the programs that run the Social Security program. And of course the headline is "risking benefits on system collapse." Really? Is it risking benefits and system collapse, or is it possible that they would be testing the thing until they knew it worked? Because you're not going to get rid of the old one until you know the new one works, right?

Now there'll be lots of edge cases, a lot of special cases. So probably there'll be some people who thought they were going to get something and don't get it right away, but that's all fixable. That would be the normal thing that would happen if you did a massive codebase changeover. Nobody thinks that that works on the first try. But what's the alternative? To keep a COBOL-based archaic system forever?

So when they say stuff like "is risking benefits and system collapse," the real risk is not doing it. Not doing it is a much bigger risk, you know, because we've reached that point where it's just so decrepit and crawling along.

But I asked this question. I said, can't AI just look at COBOL code and rewrite it in a modern form in minutes? You know, I keep reading all these stories about people say, my God, I'm a programmer and I used AI and I didn't even have to do any coding and it made an app that runs on Apple and it totally worked. And given the AI knows how to write in every language, why wouldn't you be able to just tell AI to look at the COBOL code, all you know, gazillion lines of it, and take your time and then just say, can you write that more efficiently in a modern language? And you tell it which language, of course.

And an expert who says that he works in that domain, so he's spent enough time to call himself an expert, says nope, that wouldn't work. So once again the critics of AI who keep saying things like well it's good for a demo but it doesn't really do any of the things it's supposed to do, like have an agent, you know, little AI agent that can answer questions. Even that doesn't work, but that's the most basic thing you'd want your AI to do, answer questions. It just still hallucinates, so you can't even use it for that.

But I thought the one thing it could definitely do, just definitely, was write code like a mofo. I thought you could just show it any code and say, well, we'd like this to be a little more efficient, so look for inefficiencies and fix that for me. Apparently, according to the one gentleman who says he's an expert and he wasn't guessing, said no, that doesn't work. You would have massive inefficiencies. So it could reproduce it. Reproducin