Coffee With Scott Adams — Knowledge Archive October 12, 2026
Scott Adams Philosophy Archive
Search ideas
Episodes › Episode #2206 › Segments
MainContent Cognitive Reframing

Back to episode — Episode 2206 Scott Adams - All The News That's Fit To Sip. Bring Coffee

Context —

yalty? No, no, because we don't really live in that world anymore. And that's not a statement of women. This is not a statement about women. You can't buy loyalty. You can't find it. We just sort of live in a cheater society. And if people don't cheat they might want to change partners. Yeah, now marriage would be ridiculous in my specific situation. Somebody said it more generally: why would a r…

← Previous segment →

Musk, who said that blocking is going to go away, Mark Cuban had some things to say about the algorithm. So there are two things going on. One is about blocking, and then Mark Cuban's conversation about whether Elon Musk has too much influence over the content.

Now when I saw, let's talk about them individually. So let's talk about the algorithm first. So Mark Cuban's take was that Twitter or X is now a reflection of one person's sensibilities. I hope I'm summarizing that correctly. And I commented back to Mark Cuban and said I'm not seeing the effect of that. You know, what exactly makes you think that it's become an Elon Musk sort of thing? And he responded by showing the Twitter guy — I guess there's a Twitter document, let's call it X — that explains how their algorithm works. It is hilarious to read. If you want a good laugh you should read the X document that explains how their algorithm works. Let's just say it's complicated. There's so many jargony buzzwords in it you really can't tell what it is. It's as close to being a hundred percent non-transparent as anything could be while it's trying to be transparent. So the point of it is to show you what they're doing, but because they have to use such generic jargony words it's just so dense with jargon you really can't tell what it is. Yeah, no human could read that. It's not because I'm not an engineer. It's not because I'm not an expert in the field. It's just written that no nobody could understand in detail what it's doing or not doing.

But I saw one phrase there that said that, let's see what was the phrase on the X document. It said they would seek fair balance of contents from different sources. That's actually the term used in the X document, that they would seek fair balance. What's that tell you? Do computers and AI, do they know how to figure out what's fair? How do you determine what's fair? No, there's no objective standard for fair because it doesn't just mean the numbers, right? It's not about just the numbers. It's about a judgment. Fair is only subjective.

So if the Twitter algorithm has an important component in which somebody or something is determining what a fair balance is of content from different sources, that can only be humans. Now it might be the humans who set it up and then they walk away, but it's set up by the humans. Some human had to say I'm going to tweak this or put this filter in because in my opinion, or maybe some few people's opinions, this is what fair looks like.

Would it be fair to have just as much content from something that's obviously conspiracy trash versus something that might have some credibility? Would it be fair that you get the same amount of it? Depends, right? If you're Alex Jones you'd say that is fair because you can't tell what's real and what's not. You know, maybe he'd say that, right? If you're somebody else you say that's crazy. Why would you let some 4chan conspiracy theory get the equal weight with a respectable publication?

So fair is so completely subjective that I had to immediately fold my argument and agree with Mark Cuban that we can never know what's happening with the algorithm. But we can know for sure that if Elon Musk doesn't like it he'll change it. Now that doesn't mean that we're already seeing sort of the Elon Musk version of reality. It just means it has to happen over time. And as Cuban said, Elon Musk won't always be alive and always running it. Somebody else might have to be the benevolent dictator.

I've often said the best form of government by far, the best form of government would be a brilliant benevolent dictator. Because the dictator part gets you past all the nonsense of arguing about stuff. The brilliant part gets you good solutions or at least somebody who knows to test things before they implement them. That's pretty good. And the benevolent means that that person is brilliant and working for your interests.

Now we actually have something that feels like that to me. You've got the brilliant part, Elon Musk. You've got the benevolent part, which I actually believe. I believe he really wants free speech in his bones. I think that's completely real. But that's my opinion. I don't know. If he had some secret evil agenda would I know it? No. Okay, I read minds. But it feels to me like an ideal situation that's also temporary. So I'm going to agree with Mark Cuban. No matter what you think of Elon — and I have a high opinion — he's not going to be there forever. But Twitter won't always be the same forever either. So by the time somebody else runs it maybe everything else has changed. You know, maybe there's a whole different AI platform or something. So I'd keep an eye on that.

But be aware in my opinion there's no such thing as an algorithm that is not biased. And if you had one you couldn't make a business model out of it. In other words bias in the design of the algorithm is unavoidable. So my response to Mark Cuban, he certainly made his point. Yeah, the way the algorithm is you can't avoid the fact that the owner of the company is going to get what he or she wants. That's a good solid point. But we don't know if we're seeing it yet. And I can't thin

Context —

k of any model that would be better. Can you? You know the obvious thing you think is like oh let's get rid of all moderation. Do you think that would work? Do you think you could build a platform with no moderation? No. Do you think you could have a platform in which you're not doing a lot of fact checking and that works out okay? No, no, that's not going to work out okay. There are no solutions.…

Next segment → →