Back to episode — Episode 2959 CWSA 09/15/25
Context —
ink that was painkillers. I think, but I've been off the painkillers now for a number of weeks, a couple months actually. And for the first time, I'm feeling that my voice has returned. Wow. Whatever those painkillers do, they're pretty brutal. They didn't help with pain at all, but they made me stupid and slur. I almost had to quit podcasting because I just sounded like a drunk all the time. Any…
← Previous segment →copyright stuff, but now Merriam-Webster is suing them because apparently they surface definitions that look like they came from Merriam-Webster's IP. So they're going to have to deal with that. And apparently the lawyers used the word "plagiarize" to make their point. So they showed the definition on Merriam-Webster of the word "plagiarize" and then they showed how it's defined in Perplexity. Same way. That's pretty fun lawyering, using the word plagiarism as your example.
Well, you've heard that one of the best uses for AI is writing code. And now I'm getting different opinions. I saw somebody say online today on X that they used to have to supervise a bunch of Indian programmers, which is kind of hard because of the time difference and language and all that. Sometimes language. But other coders say that there are too many examples where somebody relied on AI and it created garbage that they didn't know was garbage until it was kind of too late. So apparently if you're not an expert at writing the super prompts and if you don't check every line of code that AI writes, it really doesn't work, some say. But then others say, "Oh yeah, somebody this today said, oh, I did a 50-hour project in 5 hours." So I don't know what's going on here, but there must be certain kinds. If I had to guess, if it's a type of program that has existed before, maybe it does pretty well. But what if it's a program that you've invented in your mind or it's your assignment and nobody's ever written it before? Can AI do that? I don't know. I don't know.
So, you know, I'm the resident AI skeptic that I don't think it will change the world as fast and in the way that people assume. It's just not going to take that many jobs. I just don't think it's going to work that well. I don't know that the robots will do the manual labor, you know, unless they're dedicated robots. You know, just for industry. That might happen. But no generic robots. I just don't see it happening. And coding. I had thought to myself that if I were to stop doing this podcasting, I might be tempted to spend all my day learning to code because if you add AI to your modest skills, you could theoretically get a lot done and then maybe invent an app or something. But I thought maybe I can never get there because some
Context —
body also who knew a lot today said that you have to be really, really knowledgeable even to catch AI's mistakes and to know how to give it a prompt. So we don't seem really close to getting rid of programmers. It looks like there'll be pockets where AI makes a really big difference, but in general, you're going to need somebody who really knows what they're doing or else it's useless. Well, Elon…
Next segment → →