Coffee With Scott Adams — Knowledge Archive May 24, 2026
Scott Adams Philosophy Archive
Search ideas
Episodes Episode #2983 Segments
MainContent Climate & Environment

Back to episode — Episode 2983 CWSA 10/09/25

Context —

r money. So that's the good news. What's the bad news? Well, the bad news is that the picture on the front of the bill is the guy who's going to kill you. So there's that. That's such a good troll. I don't care if it happens or not, but if it did happen, I would never stop laughing. And I would immediately run to the bank and get me one. I would put it on my wall. It would be the best art worth $2…

← Previous segment →

in the world and the government said they were bunk. What would you do? Well, if you were CNN or MSNBC or any of the news people, you would immediately put together a panel of the top model making experts and you would have them argue how their models are actually good and not. Anybody see that show? Anybody remember seeing that on MSNBC? I don't recall seeing it. Anybody see it on CNN? I don't have any memory of seeing it. So the single most important thing in the whole world. And as soon as there's a dissenting government opinion, all the experts go away. They just go silent. No, they know they got caught. Otherwise you wouldn't see anything else. If they could have used this to bury Trump as the anti-science idiot that they've been trying to paint him for 10 years, if this worked in their narrative, they would be all over it. Instead, it's very quiet. It's very quiet.

If you wanted to see a climate expert defending these climate models, you'd probably have to invite somebody who didn't work on the models but thinks they know about them. I'll tell you what you won't get is the person who actually is putting the variables into the model. Because you know what that person knows? That person knows models are not just his or her own model, but all the other ones do. They all know it. If you don't think they know it, oh, they know it. The reason I know it is because I worked in my corporate life collecting data for various projects. You know, I would collect data to say, should we do this? Would this be more expensive than that? Should we lease or buy? And what I learned immediately is that none of my data and none of my analyses were anything but what my boss wanted to see. There's no science to it. So once you're actually in the work, you can see that it's fake. But then you're too invested because that's your job. So you do what I did, which is, well, I guess if my boss or the person funding me wants me to do more of this, I guess that's my job. Anyway, the dogs not barking. There's not enough push back on the climate models being good for me to have any belief that they're good.

Sam Altman is telling us that the Turing test probably wasn't that important in the arc of AI. The Turing test, if you didn't know, most of you know, for many years, it was thought that a computer could not be considered intelligent unless you could put it on the other side of a curtain and have a human being converse with it, not knowing if it's talking to a computer or a human on the other side of the curtain. If the computer could fool the person on the other side of the curtain consistently, that would be considered passing the Alan Turing test. Well, that happened. It happened a while ago, and it didn't make much news. Here's why I think it didn't make much news. Because AI can only fool stupid people. Do you think AI could have fooled me? No, I would just ask it to use some banned words and then that would be the end of it. There's no way the AI could fool me into thinking it was a human being. Even the current best models, no matter how smoothly they talked, no matter whether it was text or voice, there isn't the slightest chance that they could have fooled me that they were human. Not I mean, I've used the chat bots. I've tried out the Grok chat. There's no it's not even close to human. You're not even in the neighborhood of fooling me that you're human. Not even anywhere close. But it did fool some stupid people enough to say we passed the Turing test.

And when I see the AI memes, they're clearly AI created and I see how many people repost them and I look at them, I go, "Well, that's obviously AI. That's obviously fake." But some large percentage of the public looks at it and goes, "Oh, that looks pretty good to me. That looks real to me." So the Turing test was never super useful because you could always fool dumb people, but maybe there's no way you'll ever fool smart people. So I don't know if the Turing test allows for that. But Sam A

Context —

ltman has what I consider a smarter better test for AI. And he says it's when we see our first AI scientist. Meaning that the AI will discover and invent things scientifically that humans just couldn't or didn't. And once it can become like a peer of, hey, I just invented a new thing or discovered a new thing, then that would be a better test than the Turing test. I agree with him completely. Als…

Next segment → →