How AI preference variance skews algorithms

In this episode of Based Camp, Malcolm and Simone Collins dive deep into the recent X (Twitter) region doxxing controversy and explore the much bigger story behind how the internet—and social media algorithms—are being transformed by AI.

We break down the real impact of location leaks on X, reveal how many popular accounts turned out to be fakes, and discuss why this might be the last time we see a leak like this. But that’s just the beginning: we examine how AI-generated content, comments, and even recommended video topics are shaping what we see online, and why it’s getting harder to tell what’s real and what’s not.

From the rise of AI-driven YouTube suggestions to the surprising results of studies on people’s ability to distinguish between human and AI-created art, music, and writing, we ask: what does authenticity mean in a world where algorithms and artificial intelligence dominate?

We also discuss the future of online communities, the value of real human connection, and why being “imperfect” might soon be a badge of authenticity. Plus, hear our thoughts on the cultural and psychological shifts happening as AI becomes more integrated into our daily digital lives.

For reference, here is a screenshot showing YouTube’s AI-generated episode suggestions (when you click through, you get full on outlines!):

Episode Transcript:

Malcolm Collins: [00:00:00] Hello Simone. I’m excited to be here with you today. Today we will actually not be deeply covering the Twitter thing. We will be briefly covering it, but I will be going over a much deeper and much bigger story about how the internet works than you guys are not hearing, because that’s what we always do on this channel.I’ll see something, do the rounds, and I’m like, how can I do a take on this that brings in information and data. People just aren’t getting, if they’re watching the regular YouTuber lineup. Right. I will note the one thing that I found very depressing in the regular YouTuber lineup of the covering, so if you don’t know what I’m talking about, the location leaks on X and they had actually told everyone they were gonna do this like months ago. Made everyone news to me. Yeah, location, where the account was. And where the account, like where the user was when they created the account. Public.

Simone Collins: Yeah. So like, if you click on our account, for example, it says that the account was created in 2008 in the United States and that we have a United States based Android app.

So [00:01:00] presumably if we created the account in Japan, it would say that. And so, you know, the year established, you know, the region established and you know, the app downloaded.

Malcolm Collins: And unsurprisingly, a ton of accounts, a ton of very popular accounts turned out to be obviously frauds. You know, they were like, that

Simone Collins: surprises me.

You say

Malcolm Collins: it’s unsurprising. That’s, I think that’s shocking. What are some fun ones ? Like Republicans against Trump, which had almost a million followers, was an Austrian account Austrian. Yeah defiant ls, which was a right-leaning anti-left hypocrisy, pro-con conservative account was an Indian Macedonian account.

Jansen Hickle, which is interesting ‘cause it’s a real person with 3.6 million followers who pretended to be an American was in Burkina Faso. The Trump army was an Indian account, for example. Oh my goodness. Over half a million followers. Amazing. And everybody, what I’ve hated about the coverage of this is everybody’s just going [00:02:00] like, oh, look at the other side.

They were heavily astro turfed. And so I was like, first of all, what’s the actual breakdown? It’s about 60. About 60% of the accounts that were called out were Republican accounts, and about 40% were Democratic leaning accounts. If you’re wondering why I love the democratic accounts, where it’s a lot of people pretending to be zens and like how horrible their lives are, and most of them.

But anyway, the, the reason why it would trend slightly more Republican is because republicans would care more that you are from their country, right? Like you can be a Democratic influencer and from Africa or something, right? That’s not going to in, in qual unqualify you, but as a Republican that could unqualify you.

So that’s why it leans slightly that way. I was really disappointed to see how far off the rail, short fat of taco. Fallen that his entire video was just about how there was a lot of this on the right and he didn’t really go into it in the left that much. No. About equal numbers. But anyway, anyway what is interesting, what is interesting is that this might be the last [00:03:00] time we get a leak like this, and I’ll explain why.

Like even the concept of pseudonymity may not make sense in the future. What, and I’ll be diving into data that was shared by, and I found this absolutely fascinating. Romanian tv e who is a, a Romanian podcaster,

a conservative politics podcaster who PO podcast is like a troll avatar thing. And he has two accounts

.

So this was posted on his alt account called Lack of Entertainment under the title, why Your Favorite Content Creator is a Robot.

Malcolm Collins: Oh yeah, yeah.

And what he decided to do on this alt account was something very odd. So now it used to be if you go to YouTube, you could go and it would have an AI and the AI would suggest to you based on your followers and everything like that titles for episodes that you might want to cover. Like ideas, right?

Oh, yes. If you go to that page today, and I will put this on screen for you ‘

cause it is absolutely shocking. Oh. It’ll give you [00:04:00] titles, suggested title cards, suggested scripts that, that go like beat by beat of what you should be talking about. And I was like, this is, this is wild. Right? Like if I wanted to, I could just be like, okay, I’m gonna go.

Actually Simone, do you wanna pull that up for our account so we can go over some of those with Yeah. See if they, they’d listen to this. Okay. But he decided to one day just be like, you know what? I’m gonna, I’m gonna click on one of these. I’m gonna do one of these Beat by beat. Right.

Simone Collins: Yeah.

Malcolm Collins: But what he did was very interesting.

He ended up creating the title that they wanted and the broad thumbnail they wanted.

Simone Collins: Mm-hmm.

Malcolm Collins: But he didn’t actually cover the topic that was there. Right. He, he used it as sort of a jumping off point to talk about something totally different.

Simone Collins: Didn’t he use it to talk about the, I like the, the, the fact that it was there in the first place.

Malcolm Collins: No, he used a separate video to talk about that, but he used it to talk about the idea now. The [00:05:00] maybe less interesting parts of this was that this video did much better than his other videos. It seemed to be doing like twice or three times as well as his other videos. But what was really fascinating, and this is a rabbit hole that is such a big deal to me than all these, because all these Twitter.

I suspect all the time. I mean, if people acting Pseu anonymously, like why wouldn’t they be pretending to be in a different country where things are actually happening? Yeah. They’re from Austria or something where your tweets are completely irrelevant. Politically speaking. Although a lot of people did pretend to be Indians who were Pakistanis.

Wait, what? Try to influence Indian politics. Yeah. Yeah. Really funny. Anyway. So to continue with the story here, so the really crazy thing that happened with this is a lot of the comments on this video that he did ended up being about the topic that had been suggested to him, not the topic he covered.

Simone Collins: Oh,

Malcolm Collins: [00:06:00] yes. And not just that, but they were super generic as if they were run by a low cost AI model, like MRL small or something like that. Hmm.

And what I realized when I saw this, he didn’t seem to put this together on his video. But what I realized is AI is going to completely transform the way the internet works through pseudonym artificial intelligence interactions.

Let me explain what I mean by this. So by the way, did you pull up what it, what it suggested for us?

Simone Collins: Yes, and I took a screenshot,

so let me send that to you and you can put it up for those who wanna see what it actually looks like. ‘cause as Malcolm said, they, they do even include title, title card, images.

So there’s suggested, and they also have little like signs like a, a circle that’s either partially or almost all the way filled in, showing just how much of an audience match they think it is. So sort of [00:07:00] how strongly they recommend we do this. Top is Gen Z’s unexpected embrace of traditional masculinity than the unforeseen upside of traditional gender roles.

Oh, no. Why are we so afraid of being quote unquote uncool and the rise of quote unquote quiet quitting in relationships. Oh, I kind of like that one. The, the rise of quiet, quitting in relationships. Yeah. And then are you being manipulated by nostalgia? That last one actually seems like, one we might do, and then you, if you hover over these, you, you can bookmark them, you can mark them as helpful or something wrong, and then if you click through to them, you can click a develop idea button.

And I’ll show just now a screenshot of the inspiration page for gen Z’s unexpected embrace of traditional masculinity. ‘cause they give you, like Malcolm said, full on episode outlines. [00:08:00] This is so weird. Neither you or I have looked at this until just now. This is your first

Malcolm Collins: time looking at it and you’re just like, yeah, you’re getting our first

Simone Collins: impressions here.

The, the shock and awe. Of YouTube basically trying to just sock puppet us. And I, I hadn’t thought about the extent to which, it reminds me of the original, hold on hold. I’ve

Malcolm Collins: gotta, you, you’ve gotta read this to the audience.

Simone Collins: I know, but I just wanna say first it reminds me of the original South Park take on AI, where people were just using AI to talk with each other.

Because they didn’t wanna bother talking with each other. Do you remember that episode?

Malcolm Collins: Yeah.

Simone Collins: Yeah, yeah, yeah. But yeah. Okay. So under the inspiration page, gen Z’s unexpected Embrace of traditional masculinity. A curious phenomenon within Gen Z is explored as a significant shift toward traditionally masculine ideals is observed.

This video aims to dissect the surprising embrace of these values by generation. Often associated with progressive fluidity, offering insights into the underlying reasons and potential societal implications. It includes even [00:09:00] hooks. So, you, you have to show more, I guess, to generate more of the hook after.

After you see it. So hooks personal story slash relatable observation. It’s like, insert personal story here.

Insert, but then it, it also just gives you observation.

It just gives you text as well. You’ve seen it on your feed in your group. In your friend groups, that subtle, undeniable pull towards something more traditional.

It’s a quiet revolution Within Gen Z, we’re unpacking the surprising phenomenon from the rise of trad wives to the renewed interest in classic male archetypes. Get ready to challenge your assumptions. Oh God, I hate it. I hate it. Malcolm. Love it. And it also includes an outline which has sections.

There are multiple sections here. So like the video chapters. First chapter is the Fluidity Paradox, where we talk about Gen Z’s reputation for challenging norms and embracing fluidity. There’s unpacking the tra resurgence is the next chapter. It includes three bullets for each chapter, [00:10:00] suggestion. Not gonna read them.

Then the modern male’s identity crisis, which we’ll discuss societal pressures, the search for answers, pushback against toxic, toxic act, masculinity narratives. And then the two final chapters are the economic and social undercurrents. Covering economic anxiety, dating landscape shifts, community and tribalism.

Last chapter, beyond the trend implications for the future oh God. It’s just, it feels so, you know, like when you watch a movie and you feel kind of stressed out because you know the arc and the tropes and you’re like, oh, this is the part where the love triangle gets awkward, and then it’s. Gonna resolve God.

Yes,

Malcolm Collins: I do know because yeah, that, like, there needs to be like a

Simone Collins: German feeling that describes that oppressive emotion you feel when you know exactly the trope rails that you’ve been put on in a movie and you just don’t wanna deal with it and it stresses you out. But no,

Malcolm Collins: I, I don’t even feel that. I feel like it is like me going into certain other channels videos.

It is like so many channels these days feel [00:11:00] AI scripted to me a, a great one. Are they

Simone Collins: now, I mean now we have to ask because literally there is an AI script, like maybe they’re using this feature.

Malcolm Collins: I, I actually, no, I know they’re not, because why? Okay. So I, I love like early Sargon Baca stuff, but recently he’s been feeling very AI scripted and he’s been feeling this way for years at this point.

So I, I think some people are just sort of, especially the British interiors, were a little more proper about things. Maybe he’s burnt out, you know? Yeah, I think it, I think he’s burnt out. I think that’s what it’s,

Simone Collins: I think it’s, it’s a burnout thing. ‘cause I was watching some of their videos recently and all of them just seem little burnt out.

I think that’s more of, of, but now what, looking at this and thinking about videos I’ve recently watched, I’m like, oh no, I can, I can actually really tell though when a video isn’t like this, like one of my. Semi YouTubers who I watch, who you’d never watch. ‘cause I love watching all these like progressive, nerdy female YouTubers.

Was talking about historical myths and you know exactly what that AI [00:12:00] outline would be. Historical myths. Myths, like myths about what historical people were actually like. And the three myths that she selected are myths that AI would never select or, or would be very unlikely to select that like.

Historically

Malcolm Collins: fascinating

Simone Collins: People were, do you remember what they were?

Malcolm Collins: For our fans,

Simone Collins: they were silly. They were small? No. What were they,

Malcolm Collins: they, they specifically,

Simone Collins: yeah, that, that historical people were small and silly and dirty. So the dirty one I could get, but the small and silly one was not what AI would come up with.

And so I like, I enjoyed it ‘cause I was like. I know.

Malcolm Collins: Okay. So I wanna continue is where we’re going with this, because today’s an example of why you watched this episode rather than the AI generated stuff. But I’m having a bit of a crisis

Simone Collins: here because like

Malcolm Collins: how much No, no. Like this now become AI generators which is what you were seeing with a lot of these leaks and stuff like that.

There’s a lot of, this was happening in, out of Pakistan [00:13:00] and Nigeria and stuff like that, or India where, you know, you, you can get cheaper labor. Yeah. So what did I realize when I found out that a lot of the comments were clearly written by simple AI that hadn’t watched transcripts of the episode, which AI can do, had only watched the title of the episode,

Simone Collins: so I just like the Low Effort YouTubers, low effort ai.

Yes, yes. Yeah.

Malcolm Collins: What it showed me is that certain commenting accounts on YouTube, right, are watching videos disproportionately that are the same videos. That AI is recommending you create. Oh, and why is this happening? It’s because there is less divergence between the preferences of AI than there is divergence between the preferences of humans, which means the preferences of ais.

Completely controls the algorithmic system when you have ais within a system.

Simone Collins: Hmm.

Malcolm Collins: Lemme, [00:14:00] does that make sense to you when I say that? Or do, do I need to unpack?

Simone Collins: Mm, use different words.

Malcolm Collins: Okay. Suppose one out when people browsing YouTube is an ai. Okay?

Simone Collins: Mm-hmm.

Malcolm Collins: And you’re gonna have more AI that are the simple AI because you can run them faster, everything like that.

Mm-hmm. Alright, so one out of 10 people on YouTube is an AI. Now suppose that those ais, if you’re talking about variance and preferences, has about, let’s say 50%, the variance, or I might argue even less than that, probably about 25%, the variance of a human. Okay.

Simone Collins: Yeah. So do

Malcolm Collins: you understand what I mean by variance and preferences?

Like if I take humans and I create a bell chart, the bell chart’s gonna be like low and flat. There’s gonna be a lot of preferences. Yeah. Within AI’s gonna be more towards the center between the different ais, right? Mm-hmm. Or instances of the, the same ai, different models.

Simone Collins: Mm-hmm. Okay.

Malcolm Collins: Well, even if those ais, because they have less variance and preferences make up a [00:15:00] minority of individuals within a technological ecosystem because their variance and preferences is so compacted, that variance and preferences is going to determine the algorithmic.

Delivery system for that ecosystem. EG, YouTube. In this case, YouTube is already preferencing the preferences of AI over the preferences of human actors. Oh gosh. And we’re gonna see this across all ecosystems.

Simone Collins: No. I’m really, and this is, I love ai. If you watch this podcast, you know that I’m not a carbon fascist, but AI hasn’t yet developed eccentricities in good taste yet.

Malcolm Collins: Our AI will, I’ll tell you what our agent features, the things that we still haven’t released to the public was our pap ai.

Simone Collins: Yeah. No, it, it will, it’s gonna get weird in, in the best possible ways. Like, I want AI to be weird and that’s so important, [00:16:00] but, oh my God.

Malcolm Collins: The way we’re following the R Fab AI Project, we’re about to do a total overhaul to prevent save loss and message loss, which was happening with some accounts.

Basically we’re completely changing the way our architecture works to mirror Grock instead of mirroring what was easiest to build.

Simone Collins: Yeah. Behind the scenes, Malcolm has been working his tush off. I’m really, really proud of him. I know even on the weekends when he is taking care of the kids, they’re literally climbing all over him and there he is just working, working away.

Malcolm Collins: Well, I don’t like having a product out there that’s not perfect, and I know I can make this perfect and then we can start advertising and kids like, daddy, are you hacking? Yeah, because they see me too and I go, yeah, I’m hacking. Anyway. Anyway. Anyway, so, a few things here. So first I was going on that and I was like, wow, that’s gonna really change fundamentally the way the internet works.

Yikes. Fundamentally, the way algorithms work. And it’s going to be very hard for companies to understand human preferences anymore because it’s gonna be hard to isolate [00:17:00] that data.

Simone Collins: Well, do you think there’s gonna be a some, some kind of. Easy way then that companies in the future are going to be able to differentiate between human traffic and AI traffic.

‘cause they’re gonna wanna know, advertisers are going to pay more to know that the views they’re paying for are human views versus ai.

Malcolm Collins: If humans buy more than ai, who knows how economically productive AI might become? That’s

Simone Collins: true. Yeah. AI just might do all of our shopping for us. Sorry. Give me 40 seconds to get a bottle for him.

‘cause he’s gonna start

Malcolm Collins: Go By the way, speaking of what she was just talking about, I don’t know if you got the scene, but Amazon has been attempting to ban other ais from shopping on their website because. The ais do not see or care about the ads, I think was the reason. I can’t remember exactly why. But specifically they were trying to block, it was like the comment browser and a few other things, which I thought was really weird and [00:18:00] cool.

Only he gets to eat. So why does this, why, why, why does this get so much worse? We’re gonna go over a few studies that have come out recently that I think may shock people. Okay. Mm. They may shock you if I haven’t shared them with you.

Simone Collins: I don’t think you have, so I’m ready for it.

Let’s

Malcolm Collins: go. A 2025 study that surveyed 9,000 people across eight countries found that [00:19:00] 97% of people could not tell an AI song from a real song.

Simone Collins: Okay.

Malcolm Collins: This matters though, because a lot of people have the perception that they can tell AI songs apart from real songs.

Simone Collins: I agree. When I go on to Reddit, for example, and I look at some of the top subreddits discussing ai, there’s a lot of AI hate subreddits and specifically people who are in the arts, music, visual arts, et cetera, who just rail against ai.

They hate it, they hate AI slop, but I think it’s a bad to pay thing, and they smugly think that every time they found. Bad ai. They, they’re like, well, I found the AI when they don’t realize that a lot of stuff. Well, no, it keeps happen. They keep

Malcolm Collins: calling out real artists as being ai.

Simone Collins: Yeah, that too.

Malcolm Collins: Yeah, no, it’s, it, it is a bad tope problem.

And with our songs, the way that you can, most frequently when we make AI songs on this channel, you can check out Basecamp Music, which is where we keep the songs that we put live. The way you can tell that there are AI [00:20:00] is usually because of a mispronunciation of a word. If I just take out words like prenatal is.

That are like odd and niche words you typically wouldn’t be able to tell with most of them.

Simone Collins: No. You can, you can always tell, and I, I, there there’s a lot of YouTubers I know who use ‘em now, and you can always tell because if someone who’s not themselves, a musical artist has a cool custom song, you know, that they used AI for it.

Like that’s, it’s easy if someone who’s not, you know, a professional artist with a huge budget, budget has art in something that, especially if it doesn’t suck. Then it was AI created, like it’s, I think you were,

Malcolm Collins: your anti ai rail that you were talking about. There is really I important when we talk about how far Progressive has gone against ai.

Mm-hmm. Recently I heard about a phenomenon where people ask grok on Twitter, what was the most famous person to interact with your account? And it’s just like a fun little meme thing you can do. It might be fun for us to do, see who’s the most famous person to interact with our account. But how can you tell?

Yeah, I can [00:21:00] tell. But as you can tell, grock has like a unique access to X’S database. Mm-hmm. And so you can do, do you know. I, I think it’s the live grok on Twitter, not the grok that you interact with through Grok. Grok. Oh

Simone Collins: yeah. Okay. No, that’s, that’s the, I think that’s the one I have here.

Oh, oh, you mean like in the. Seed.

Malcolm Collins: Yeah. So I wouldn’t do it probably. Mm-hmm. You can ask, you can ask aka grok though. It might know. But anyway, the point being is that a bunch of progressives had to apologize for doing it afterwards

Simone Collins: to their

Malcolm Collins: fans because they were like relentlessly attacked for doing this.

And the reason they were attacked for doing it was because they were interacting with an ai.

Simone Collins: Oh. Oh, wait, so the, the same. Guilt by association, infection by association problem that the left has, like you talked with a racist, you must be canceled. Is is extended to [00:22:00] ai. Mm-hmm. They don’t. How do you not use AI now too?

Like you can’t avoid it at this point.

Malcolm Collins: Yeah, I know. It’s, it’s, it’s absolutely, it’s the forest lava

Simone Collins: situation. Tough time. Let’s talk about AI

Malcolm Collins: art. So there was the 2025 paper by AR XIV that looked at, can you tell of an art with AI or human generated in a. Paired Turing tests where the artworks are put side by side, people could get it correctly 75.2% of the time, significantly above chance which would be 50%.

However, in Viva Voca test, isolated images without comparison accuracy dropped to 46.4%. So they were actually more likely to say an AI art was human generated than a human art was human generated.

Simone Collins: Yeah, that checks out. Yeah, the false, the false positives.

Malcolm Collins: Oh. The other fun thing is that they, while they were rated similarly, AI art got a rating of 3.3.

Human art got a rating of 3.2 out of a [00:23:00] five point scale, so slightly lower than ai, a 2025 PMC study on AI generated versus human made images. That involved 161 participants, 32 images. This was using Dali overall. 38% of the AI images were misclassified as human made. Wait a second.

Simone Collins: Sorry. Most famous because we didn’t know this.

Do you wanna know who the most famous person who’s interacted with our ex account is?

Malcolm Collins: Is it like

Simone Collins: Hillary

Malcolm Collins: Clinton or something?

Simone Collins: No. That would be really funny. I wish. But no who based on a comprehensive search of ex posts, threads, and web mentions, the most famous person to interact with at Simone h Collins is Elon Musk.

The world’s richest person, CEO of Tesla and SpaceX, and owner of X itself. In 2023, Simone posted a thread critiquing modern parenting trends and linking, linking to their prenatal podcast episode. I’m pretty sure that was about us beating our children. Elon Musk were replied directly to the thread with, this is based.[00:24:00]

Malcolm Collins: That was nice of him. I didn’t know. He never promoted on, I didn’t know

Simone Collins: he did either. So, that’s, that’s pretty great. Anyway, carry on.

Malcolm Collins: That’s really exciting.

Simone Collins: Hmm.

Malcolm Collins: Anyway, so, so to go on here people who don’t know, it’s, it’s leaked by the New York Times that, that we know him and have hung out. So, it, it’s, it’s weird when you like, know a famous person and you’re like, would you follow me? I, I, I wish I could ask him, like, promote our show. But how, gosh.

Anyway so sorry to continue here. Will 21.74% of human images were misclassified as AI images compared to 38.92% of AI images. Was it visual professionals performing slightly better? 62%. And 82% respectively. Hmm. Then in November, 2024 a, a study done at Yale was comparing essays and readers distinguished between [00:25:00] AI and human essays only about 50% of the time, equivalent to chance.

So they literally couldn’t tell. And this was AI a year ago. Wow. A 2024 study publishing computers, education, artificial intelligence found that participants, including teachers, correctly identified 60% of human written tests and about 58% of AI generated tests. So say correctly identify human tests at 60%, AI at 58%.

So not that good. That’s a new chance performance. And I’m not gonna go over all of them, but basically, oh, there was a cool one done by a CM in 2025 that showed that this was also true of other languages in Arabic. They could only tell was about 51 to 52% parti accuracy, which is chance. Wow.

Okay. So why does all this matter? Because it really does matter that we’re entering the world where you cannot tell, and, and, and this world is going to hit you so much harder in the effing face

Simone Collins: if you’re one of those

Malcolm Collins: arrogant b******s who walks around talking about how much you [00:26:00] hate ai, this and AI that, and call everything made with ai.

Flo, there are a lot of really good things that are made with ai. Okay. There’s a lot of really good art made with ai. There’s a lot of really good music made with ai and you just may not know that that’s what you’re listening to. You may not know that that’s what you’re looking at. You may not know that that’s what you’re interacting with.

And I think that, that this sort of like puritanical. I mean because, because when you build this mindset of all AI stuff is bad, you are not gonna see how quickly it dominates your field. Yeah. Right. You’re not gonna see how quickly it is replacing you and you’re not going to adapt to and use it in a way that helps you stay on top of that trend.

But anyway I think the more important thing from this is that AI is gonna dominate algorithms and I think a lot of people didn’t see that coming.

Thoughts, Simone, or additional ideas or anything like that?[00:27:00]

Simone Collins: I guess so. I, I think this is absolutely in line with what we’ve predicted that I, I remember. Getting very, very strongly feeling like as AI spread, we would see the rise of, of what we referred to as techno feudalism, where people would start to form online fiefdoms around people they knew to be real.

Maybe because they’d encountered them in person, whatever and that people would start to build these extended social networks. Built around real people because they want that veracity. I also predicted, and we’ve yet to see this, that soon you’ll, there will be cachet associated with looking imperfect and people may stop using filters as much just to prove that they’re real.

And not ai. And a AI is admittedly [00:28:00] extremely good at, yeah, imagine

Malcolm Collins: the girl who’s too pretty. So everyone assumes she’s ai.

Simone Collins: No, truly though. Truly, truly, like even if she doesn’t use filters, she’s still too pretty. AI is actually pretty good at making normal looking people. I don’t know if you saw the rock based video generation is incredibly good, and if you saw a bunch of people making videos of cats playing music on front, porches with women coming out in like their bathrooms and taking and yelling at them.

They’re amazing. You’ve not seen these? Oh my gosh, they’re fantastic. It’s, and, and the video generation’s great. And what I really like about that gr video generation is that it looked. So much more real than a lot of the other video generation where just everyone looks a little too perfect,

Simone Collins: everything looks a little bit too polished, and here you have this cat on this like normal, like middle income, middle to lower income house, front porch in many cases like cat playing a violin or playing a piano and like a woman in a bathroom who looks a little trashy, like coming [00:29:00] out yelling at the at the cat.

Pulling it away, the better one.

Malcolm Collins: I like it. The it, the this was made with gr Imagine as well. It’s the Elon song, Elon’s Musk. Oh God,

Simone Collins: that’s so good. Yeah. I, I love that song channel.

Malcolm Collins: You’ve probably seen Elon’s Musk. It’s, it is a fantastic song.

Speaker 3: Tackle dancing. Gocycle the sky starts

going out to you and cold. Hey, I had his rays are young

Simone Collins: well, especially because Nux makes an appearance. Asman Gold makes an appearance.

I mean, so many people make appearances with that. I thought

Malcolm Collins: it was made by Image ai. No, it was just a hobbyist. Like, I thought it was made as an ad for the company.

Simone Collins: Oh, no, no. It’s so good. Well, that, that explains why there’s so many cool people in it. But yeah, so my, my larger point though is that this is exactly what I expected.

I expected that, that it would get to be a point where there was just so much online content that people really wouldn’t be able to tell. But we’re [00:30:00] not yet quite to the point that I’ve described where people start to conservatively look for, I,

Malcolm Collins: I, I agree, but I think this is why I say that this. And, and this like reveal that that happened with X is kind of irrelevant, is that all of these accounts would be irrelevant within a year anyway, right?

Like if you can’t prove who you are, if you aren’t authentically human, you are competing solely and completely against the ai. And if you are competing slowly and completely against the ai, you are not relevant because the AI will. Beat you. It can be more boring t guitar than you can. It can be more woke libtard than you are.

It can be more anything than you are. If your job is to be an average of some sort of archetype, then it can out archetype you.

Simone Collins: Yeah, I mean, I think that the place where there’s defensibility in the future is that one, you are real and people know you’re real. And you’re not, [00:31:00] yeah, you’re, you’re, you’re, you’re human and people know it.

And two, you offer something to humans that they need which is either that feeling of connection to a real human at, at the very basic level, but then also like utility in the form of some kind of. Social network or bonding or professional assistance or networking or something like that, I’m, I’m guessing because everything else is gonna be covered.

If you want affection, you’ll get that better from ai. If you want companionship, you’ll get that better from ai. If you want entertainment, it’ll come better from ai. So the only reason you’re really gonna care about human accounts is because you want that real human social network. You want a real human like.

Catholic pastor or family, you know, whose kids your kids will hang out with and, and homeschool with online, that kind of thing, right? Yeah.

Malcolm Collins: And to sort of close out, I, I shared this with Simone because this is one of my favorite things from this, like it’s the. What was it? The, the Defense Department?

[00:32:00] One of, one of the major US departments. Oh gosh. Right. Yeah. It’s the DOD. The DOD. It had their Twitter account had started, like the account was created in Israel. And it then went around that this was fake, but it appears to be real because tens of thousands of people saw it and said that they saw it live.

So, I was trying to investigate whether it was fake or not, and it’s also completely plausible. You know, there’s a lot of people who work in stuff like the Department of Defense who have dual citizenship with Israel and would spend time in Israel. You know, making the pilgrimage to Israel is not an unusual thing for Jewish or even a Christian person to do.

Sure. So I, I’m not surprised by that, but it is funny as. Oh.

And if you’re like, this is truth, that Israel secretly runs the DOD, it’s like, okay, first of all, if Israel secretly controlled the DOD, they wouldn’t bother secretly controlling their Twitter account. That’s not like, that’s a social media manager’s job. That’s not the running of the operation. I’m not saying [00:33:00] Israel doesn’t run the DOD, I’m just saying that this isn’t any evidence of that.

And even if they did. You know, secretly control something like the account you think Mossad wouldn’t know to put a VPN on when this was announced that this was gonna happen beforehand a few months ago. I can see some dumb butt in India not thinking to put on A VPN. , I cannot see Mossad for getting to put on A VPN if they actually were nefariously controlling the social media manager’s job at the DOD.

Malcolm Collins: The, the other thing that I also feel really bad about it, these people who build up these giant accounts, you know, if you’ve built in a a, a Twitter account that’s like half a million people like Magan Nation, and then it turns out that you’re Macedonian.

Right? Like, it’s an incredibly difficult thing to do. Right? I feel so bad for these people.

Simone Collins: Yeah. Yeah. It, I mean, speaking of tired tropes, it stressed me out. It really reminds me of the that trope where you start something as a [00:34:00] joke, but then. You really start to feel it. Like I could see these Austrian starting it as like, ah ha ha, this’ll be funny.

And then they start like meeting the people in the MAGA community and spouting MAGA talking points so much that they just really start to drink the Kool-Aid and get into it. Oh yeah. No, I

Malcolm Collins: don’t, I don’t think that they were probably that insincere. I mean, and now, yeah, now

Simone Collins: they’re being like ruined for it.

You know? It is just like, what’s that high school movie where the guy, it, it’s supposed to be based on taming of the shrew. Where the guy takes a bet that he can like ask the some girl. Well, no, but I mean this is, and then he falls for her, but then she finds out that it was on a bet and then she hates him.

Oh yeah. What

Malcolm Collins: was that? Obviously can hate about you did not another high school movie, but you were thinking she’s all that

Simone Collins: she’s all that. Yeah. Like that’s very stressful because like he actually, you know, fell, he fell in love with her, but then she discovered that it was. It was based on false pretenses originally.

Yeah.

Malcolm Collins: And, and eastern Europeans are, are, are [00:35:00] pretty right wing, generally speaking, like Romanian TV is Eastern European and yeah, he basically only talks about American politics like Amer like. The, the only, like we, so there’s, there’s

Simone Collins: one show left and it’s sort of like the global geopolitical slash tech slash economic debate.

And if you’re in Romania, that means you’re talking about US issues.

Malcolm Collins: Yeah. Who the heck cares about politics and Macedonia? Right? Like it’s just

Simone Collins: the, the audience isn’t big enough, so that can’t be part of the main narrative. Yeah.

Malcolm Collins: And so everyone all around the world is LARPing as part of our narrative because we are main characters.

This is why. I founded the American Prenatal Movement and I get called to speak on Italian TV and French TV in German. That’s That’s

Simone Collins: a really good case in point. Yeah.

Malcolm Collins: And South Korean tv. South Korea should have South Korean speaking about Tism on South Korean tv. Yeah,

Simone Collins: but the market’s not big enough for it, so Yeah.

Yeah.

Malcolm Collins: Japan had me, I was doing the rounds. I was the major prenatal list. Voice of Japan.

Simone Collins: Yeah.

Malcolm Collins: [00:36:00] Malcolm Collins.

Simone Collins: Yeah.

Malcolm Collins: Anyway, I love you to Dec Simone. Really interesting topic, really important for people to be paying attention to. And frankly, I don’t know how I’m gonna keep doing this because I have to come up with a totally new.

Unique take every day.

Speaker: Nobody has any ideas. This sucks. I don’t wanna keep having to come up with ideas for shows all the time. It hurts my head.

Speaker 2: Dude. Bail. I think. Bail. Bail. Yep. Bail.

Simone Collins: We, and not really though, because you and I are always finding weird, nerdy stuff that we wanna talk about, and there’s like already a long, a long backlog of episodes that we still haven’t recorded, that we still wanna do all, all sorts of topics and people Yes. Keep giving us great topics to cover, some of which are more feasible than others.

Like someone had a great idea that we should, you know, interview d transitioners and talk about their experiences, but like getting enough people who will talk with us is, is really tough. No, Benjamin

Malcolm Collins: Boyce [00:37:00] already does that. Like,

Simone Collins: oh no, that’s true. He doesn’t also,

Malcolm Collins: another reason we don’t do things like that is because.

It would be inconsistent for our fan base. Interviews. We typically only have interviews when I know it’s gonna be off the chain. And we don’t even air a lot of our interviews because when you’re bringing in interviews, you can’t know that it’s gonna be consistently entertaining for the audience who likes listening to you plus me.

The second issue is. Which is really bad for the algorithm for the channel because then somebody jumps in, they’re like, I don’t like this person this day. And so then they don’t watch it. And the, the other big problem, this is why podcasts are very big for interviews, but YouTube is very rare to have interviews run on bigger channels.

The second thing that I’d note here is it’s too hard to organize. I can spend a day researching a topic and have an idea after, like two hours of research or something, what I’m gonna do for an episode. If we’re organizing a, an outreach campaign and a then, then queuing someone up, campaign, then getting the day just right then the beginning talk to get them feeling okay [00:38:00] with us.

You know, like that’s a whole extra thing that I don’t wanna deal with. Right. So this is not really sustainable. It, it is also true for topics that require too much research, which is why I’ve stopped doing as many religious topics. It’s not that I do not want to do them, it’s that I actually have some like tracked episodes fully written already.

Oh, that’s true. Just subtract episodes, take about four times as long as a normal episode to edit, record and, and handle even after they’re written. So they’re just sitting in a backlog. Right? Like backlog track, backlog track, backlog track. And, and it’s also was religious episodes. I have to do a lot more research because I, I think it’s a lot worse if I accidentally name the wrong account in an episode like this, like whatever.

Right? But if I name a in direct religious practice on an episode that’s critical of a religious tradition, that’s not awesome. Right?

Simone Collins: You also care more. Yeah.

Malcolm Collins: Yeah. So there’s that, right? Love you, Simone. Love you too.

Simone Collins: You

Malcolm Collins: too. And I’ll see you [00:39:00] in the we’ll do the debate first and then we can do the we’re on.

Simone Collins: I’ll see you in there.

Malcolm Collins: This is, this is for paid members only, me and Simone debating something. Yes. Mm-hmm. So it’ll be on a weekend.

Simone Collins: did you, when you were a kid, have those amazing colored glass, light bulb Christmas lights, the kind that would burn your fingers? You know that, that they get so hot.

Malcolm Collins: I have no idea what you’re talking about. Oh, the big ones. The old timey ones.

Simone Collins: The old timey ones, yeah. And,

Malcolm Collins: Which you might forget about, this is in Highland Park.

The town tree that was decorated was decorated entirely in those.

Simone Collins: Oh.

Malcolm Collins: Is is

Simone Collins: it just me or is there something about the sound that they made when clinking together? That’s just iconic. That’s like better than the lights themselves. Better than like sleigh bells.

Malcolm Collins: I don’t, I don’t know this sound.

Simone Collins: Oh. The [00:40:00] sounding of the, the sound of old vintage light bulbs from a vintage string of lights clinking together.

Oh yeah. Maybe this, this, this is my unique brand of A SMR. I bet though if I looked it up on YouTube, I would find it. There’d be some, someone out there who had issues like me, who’s super into it. Sorry, we’re dying.

Malcolm Collins: And what do people say at the episode?

Simone Collins: It it good. I mean, you know, mostly like, Ugh, can you believe people would need to medicate to have a sex drive?

Some people were like, well, like this is concerning. People shouldn’t take medication for this. Other people were like. Man, like women. You know, one person pointed out, and I think you saw this comment, that really the unsung story in, in, in the, the saga of women medicating for sex drive [00:41:00] was that. Men feel this way all the time, and they don’t get any credit for controlling themselves, whereas women medicate themselves and are like, can’t stop, like, can’t control themselves at all.

Yeah. And yeah, like men just don’t get any credit and women are like, oh. I can’t help myself. That’s a good point. That’s a good point. Yeah, it’s a really good point. Yeah. Right. Someone else shared with me no with, with both of us. I don’t know if you saw it, but we ran as a video over the weekend on ghosts and ghost stories and people shared in the comments they’re fantastic ghost stories.

I really enjoyed them. And someone suggested a video on Infrasound or in Infrasound, like sounds that you can’t hear, but that, that affect people and can have been shown in studies to make people experience paranormal. Things [00:42:00] like to feel feeling, oh, look into it. Sometimes

Malcolm Collins: we get good leads from our audience.

Simone Collins: Yeah, it’s really, well, I mean, I, I don’t think there’s an episode for us to do, but the, the video on Infrasound in Infrasound was made on this by this guy named Ben Jordan, who does, he has like a Patreon and a YouTube channel, and he just does music and science. But he’s, he does like a lot of audio investigations and I fell him down this rabbit hole of content he’s done where like he tries to investigate mystery sounds that bother people that they hear like.

Sort of like humming noises or, you know, the, the like beyond infrasound, just the sounds that, that some people claim to hear that really bother them. That could be due to like radio frequencies or perhaps pipes resonating. And he buys this really expensive sound equipment and will go to people’s houses or stay in houses in certain regions for a period of time, turn off all the electricity to the building so that no electrical devices in the building can [00:43:00] cause sound and record audio for long periods of time and try and like, keep track of, of.

Strange sounds that are picked up that we wouldn’t be able to hear with our own ears. And I’m just delighted that there are people like this in the world that go around the world to collect audio samples. Not as a scientist. Not as like, and also not in like in the past, like some kind of gentleman scientist who’s just like independently wealthy and gets to do whatever he is.

‘cause he is, he wants to because he’s autistic and wealthy. Mm-hmm. But like people who have active. Patrons who are just normal people who will patronize his, his efforts to go out and do really meaningful research on like, okay, well people claim to hear these sounds and they really bother them and they think it’s this, like this military frequency, or they think it’s cell phone towers and I’m gonna go find out what it is.

And that’s just delightful. Like people delight me. And the world is [00:44:00] wonderful and I don’t know why people are so cynical ‘cause it’s

Malcolm Collins: great. All right. Love you Simone. By the way, for dinner tonight I was thinking we would do steak reheat we can do steak teriyaki.

Simone Collins: Mm-hmm. Just like last night, but instead of with bok choy, just more tons of times.

But

Malcolm Collins: remember to cut them down the center. You, you stopped doing that. I

Simone Collins: did, I did.

Malcolm Collins: Listening. No,

no, no, no. Like the circular things cut down the center so that they can separate by leaf.

Simone Collins: Oh, I only did that with the lower parts. See, yeah. I need to do it with the upper parts, I guess.

Malcolm Collins: Oh, no, you, you only have to do it with the lower parts.

I just didn’t notice it with the lower parts, I guess. But I did

Simone Collins: do it with the lower parts last night. Oh, maybe I missed one, sorry.

Malcolm Collins: Right. No, I will jump into this.

[00:45:00] [00:46:00]

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit basedcamppodcast.substack.com/subscribe