Coffee With Scott Adams — Knowledge Archive May 24, 2026
Scott Adams Philosophy Archive
Search ideas
Episodes Episode #2926 Segments
NewsReaction AI & Technology

Back to episode — Episode 2926 CWSA 08/13/25

Context —

ould you like the simultaneous sip? I think you'd like it. If you'd like it, all you need is a copper mug or a glass, a tankard, a jug or a canteen, jug or flask, vessel of any kind. Fill it with your favorite liquid. I like coffee. And join me now for the unparalleled pleasure of the dopamine hit of the day. The thing that makes everything better. It's called the simultaneous sip. What happens n…

← Previous segment →

l right. Now, given that I can't tell if my audio is working, I have to check. YouTube. All right. Okay. So I'm going to put a note here on X. Rumble Studio not working. All right, that's the best I can do.

Well, did you know that Meta and their AI, they think they've taken the first step towards super intelligence? And so Zuckerberg says he's not going to release any new versions that are their best version because their best version has super intelligence. Apparently the Meta AI has already shown signs of improving itself, meaning that they can tell it to upgrade its own code and it will rewrite its own code and make it better. Now, just in small ways, but that is the thing everybody worries about, is that the AI will figure out how to do it in big ways and it'll happen kind of instantly. And if you're not ready for it, the whole world is in trouble.

I'm going to make a counter prediction. My prediction is that AI will be able to improve itself and that'll happen right now, I guess, but that the amount it can improve itself will be severely limited. It will not be instantly becoming godlike. It will do stuff like, hey, I can improve the efficiency of my code so that I can answer faster or I can answer more completely. But I feel like that's all it's going to do. I just don't feel like the nature of intelligence is that no matter how much you have, you have the ability to improve your own intelligence. I mean, beyond those small tweaks of your code. So I'm going to bet against it.

But we already know according to the Lancet there's some concern that doctors are becoming dumber as they use AI. I'm not sure I believe this. This doesn't sound like the data would be reliable necessarily, but they studied colonoscopies that were used by doctors who were using AI to make sure that they were really getting it right. And the doctors who used AI allegedly became almost instantly worse at identifying tumors on their own. Do you believe that? Do you believe that the ones who used AI to help them identify tumors that they instantly lost the ability to do it themselves or they were 20% worse at it almost instantly?

Maybe. Maybe because if you believe that you don't need to be smart and that something else is doing the smart part for you, I believe that you probably turn down your brain power a little bit. So it wouldn't surprise me. I don't know if that'll make the world worse or not. As long as the AI is getting the job done, it doesn't matter. It doesn't matter, right? As long as you get the right answer.

Well, there's a new movement. I don't know how new it is, but Silicon Valley is apparently trying harder to get smart babies. So there are a couple ways you can do that. One of the ways is that you can hire somebody to find you a mate. So they have all these really expensive dating services where a human being will match you with people that they've vetted. Now that's for the super rich. But if you're super rich and you're smart and you want to make sure that any children you have are also smart, you can tell your human matchmaker to only pair you with somebody who's super smart. So the odds of your baby being super smart, if it's the two of you, are very good. Very good.

But the other way is that they have new technologies now that can determine the likely smartest egg, I guess. So if you've got some fertilized eggs, you can check them out to find out what their probable IQ is and then you can select based on that. But part of this story is the media has trouble acknowledging that IQ matters. So they have to throw in all these qualifiers like, well, you know, IQ is just one variable. It's just one variable. I mean, you also have your lifestyle and your upbringing and your parenting and all that. To which I say, why are we so shy about IQ? That IQ is the single most predictive variable in all of life. If somebody is super smart, that doesn't necessarily mean they'll be rich or successful, but the odds of it are way, way higher than if they're not super smart.

Is it a coincidence that all the heads of our biggest, most impressive companies are also super smart? You know, Zuckerberg and Elon Musk and Jeff Bezos and, you know, I could go on, Bill Gates. So yeah, the fact that we're struggling with the question of whether IQ matters is sort of a DEI leftover because as soon as you imagine that IQ matters, then genetics comes into the conversation and then it's all racist. So you can't even have that conversation really. However, I think we're getting closer to the point where we can.

In other news, NewsNation is reporting that there was an ALS patient, that's somebody who's lost control of their bodily functions, that learned to control an iPad entirely by thought. Now if that sounds like Neuralink, you'd be right to say it sounds like it, but it's a competing company. So I guess Neuralink now has an established competitor called Synchron. S Y N C H R O N. Synchron. And they've developed a sensor that must be really good. So it's competitive with the Neuralink stuff.

So here's what's going to happen. IQ in humans only matters so much because if you've got AI to augment your IQ, how would you measure the IQ of someone who knew how to use AI and had access to it? Yeah. See how hard that is? If you took a person who has an IQ of 100 and they didn't have AI, you'd say, oh, that person has so much capability. But if you took a person with an IQ of 100, taught them how to use AI, and then made sure that they really did use it, wouldn't their functional IQ be, it wouldn't be as high as the AI's IQ, but wouldn't it be some blend between the human's 100 IQ an

Context —

d the AI's nearly infinite? How would you even measure that? And then you go to the next layer which is these Neuralink or Synchron chips. How would you measure the IQ of a cyborg who had part human brain but could access somewhat automatically the infinite intelligence of the internet? I don't know. Would that be an IQ of 100 plus of machine parts or would it be an infinite IQ? What happens if t…

Next segment → →