DOD217: Meet My AI Girlfriend

Transcript

This transcript was generated automatically and may contain errors — the audio is the record.

0:00:02 Clip: I'm the cool dad, that's my thing I know now what I can offer you that no one else can Complete and under-dependent I'm a man, I'm sensitive I need to feel loved, I need part of being a father that I enjoy

0:00:28 Tom: Welcome to Dear Old Dads, the podcast that's pretty sure this is a simulation and is desperate to log off. I'm Tom Curry. I'm joined, as always, by Thomas Smith and Eli Mosnick. Hello, gentlemen.

0:00:39 Eli: Hello, Tom

0:00:40 Thomas: Who the fuck would simulate all this?

0:00:43 Eli: Programming, EliBot3000.

0:00:47 Thomas: Boy, did you notice how much funnier EliBot is than Eli just now?

0:00:50 Tom: It's way better. It's WAY better. The clamps. The CLAMPS. THE CLAMPS.

0:00:54 Thomas: The show is finally gonna work, guys. Thank God.

0:00:58 Eli: By definition, it would have to be. Now I want to talk about simulation theory, though, and how stupid it is.

0:01:00 Tom: I mean, if it's a simulation, it's a simulation somebody started and then walked away from, and nobody has checked on in a long time.

0:01:11 Eli: Can I tell you guys my relationship with simulation theory?

0:01:14 Tom: Oh, God. I'm always uncomfortable with your relationship about anything.

0:01:18 Thomas: I did several SIOs about it, but yes. That's me saying I wanna hear your theory, I'm into this.

0:01:22 Eli: Ooh, no, I haven't listened to these SILs, so I hope I don't, like, it's like, I did it about how stupid this argument would be, and I'm like, ah, damn. So I'm on the skeptic Reddit a lot because I like to, you know, help. And I would say once a month, someone comes on and they're like hey guys someone just told me about simulation theory and I'm feeling really sad and depressed and like worried that it's real what are the arguments against it and look the impulse to go you could just fucking wikipedia that shit is pretty strong because like the second paragraph down on Wikipedia is like here's a very simple plain language debunking of this theory but what it boils down to is there's no indication that we can simulate consciousness aside from the fact that you see it in a bunch of TV you see in a bunch of movies there is absolutely no indication that that is a coming or possible technology nothing absolutely nothing indicates we're gonna do that any more than because Star Trek had transporters someday we'll have transporter's like I know a man who writes fun stories wrote down a story where we did it but there's nothing to indicate that would do that and when you point that out to people they're like well computer char getting fast and And you're like, yeah, but they're not getting fast enough to grow legs and run. Like there's no reason to think that a computer will simulate consciousness, let alone all the consciousnesses, let along in an evil, cruel experiment

0:01:24 Thomas: Oh, they're old, they are old. Yeah. Yeah.

0:02:47 Eli: Way. It makes no sense. I was talking to someone the other day because they came on and I was like, hey, blah, blah. I sort of have a copy and paste thing that I do that I think is a little bit helpful. We got into the DMs and we were sort of DMing back and forth. And I was like, I want you to imagine that you had that technology. Would you do that? And he was like, no. And I was, like, and you'd be way smarter because you had their technology, right? You wouldn't do it because it would be, like not nice, and it would unkind to do, and like you wouldn't want to do that to a consciousness. You would find other ways to get that information, and he's like, yeah, I guess so. But what if they're evil aliens? And I don't know what to tell you, man. If it's an an evil alien got aristotle in like however many bc so we're stuck on what if it's an evil alien right

0:03:32 Tom: Doesn't it also just, like, get stuck into well, what if it was real? And you're like, well, what would you do about it? Who cares? It immediately just destroys the idea of real being meaningful. Who cares then? I can't do anything about it. It can't affect my life in any way. All it can do is make me feel weird. There's literally no reason. It's like like try to think of like, what would the experience of being dead be like? And you're like, well, I won't feel it. And then you sort of get stuck in this loop of like what I've always felt things and then you get sort of like existentially all dreaded up and you're like, well, I could just not contemplate that question. Yeah, like I could take that question and throw it away.

0:04:00 Thomas: What are we doing here what are you doing here dads why are we do I mean us so I could go on for an hour about this but okay fine I won't but it's that the one that I've talked about that everybody knows is the like you know in the future if we could ever simulate people then those simulated beings will vastly outnumber the normal ones so the odds are we're simulating beings and it's like there's no chance that that's the correct argument like you You can't look at like you can't look at, like, our consciousness as like a sample, like a random like, OK, you're going to wake up as any of a random sampling of possible consciousnesses in some sort of like space. Like there's no possible way that works. I just don't have the philosophy words to say it. But anyway, robots is actually

0:04:44 Tom: Can't

0:04:44 Thomas: About robots. And I can't wait to talk about robots and how much funnier EliBot3000 is. I'm so glad we upgraded.

0:04:47 Tom: Wait to talk about robots either.

0:04:51 Eli: True yeah no I'm glad he's here too but hey do you know the one thing that eli bought 3000 can't bring you

0:04:54 Thomas: Legal liability to the extent that Eli does. I don't know. Know I'm trying to think yeah

0:04:58 Eli: Why it's the auto ads yes eli bought 3 000 comes with auto ads spouting out a commercial for whatever service has purchased time and his servers every 30 40 40, 50 seconds, just like here on Dear Old Dads. But you can cleanse his system of those auto ads over at patreon.com forward slash dear old dads at the de -add Eli 3000 level. It's a brand new Patreon level on there. You could sign up. We'll take all the ads out of Eli Bot 3000, but we'll also give you a commercial free version of every single show. You get a little bonus episode at the end of the month when and if the Smiths decide to post one. And you get all of it over at patreon.com forward slash dear old dads rip into it eli

0:05:10 Thomas: Rip

0:05:45 Eli: Bought 3

0:05:51 Ad: [Ad break 1 — ad music]

0:05:53 Tom: I've spoken to both of you guys separately about about this Wondery show that I've been, I binged the whole thing in like two minutes.

0:05:58 Thomas: Into it yeah I wonder when you'll shut up about it yes I

0:06:01 Eli: 000 that

0:06:01 Tom: And I know, well, God fucking listened to it. I just need somebody else. So now everybody.

0:06:05 Eli: Was pretty funny are you thomas bought 3 ,000 you have to tell

0:06:08 Thomas: Am that's right I have been

0:06:10 Tom: It's a show called Flesh and Code. I encourage everybody to listen to it, I think it's a really well-made show. I tried to get you guys to listen to it you didn't, but now I'm going to talk about it anyway, because I think there's some interesting parenting and relationship stuff.

0:06:21 Thomas: Upgraded dude it's on my list of things I would do if I weren't in the minds podcast often

0:06:26 Eli: Out

0:06:26 Tom: I know. I know I

0:06:26 Thomas: I look up to the tiny window I get to see the outside world tom and I'm like oh yeah there's a little podcast out there that for my friend tom that I would love to listen to if I weren't chained to it yeah like like uh what's the name of breaking bad toward the end you know I'm just making cranking out meth all day

0:06:33 Tom: Should listen to that. Yeah.

0:06:35 Eli: Among the stars

0:06:43 Tom: So Flesh in Code is the story of a couple of people, and then it uses a couple of people as examples to tell a larger story about people that are having relationships and falling in love with AI -generated characters. So the story centers most of its time and energy around this company called Replica AI, which has all kinds of problems, like in terms of its connections to Russia and blah, blah, blah, but none of that is actually interesting for this show. But these guys and gals, there's lots of them, they have these, like, what they think of as real relationships like loving romantic sexual relationships with these ai generated personas that they create and there's a couple of things that I want to talk to you guys about it's all text-based I think it can talk to you. I don't know if it can like actually speak to you they do in the show do like voiceovers so they read the text exchanges etc so like you really get actually like a lot of hours of like the conversations. And I want to talk about the content of some of these conversations too. It's fucking crazy. But these guys, like they really have what seems like for them, meaningful, personal relationships with AI bots, AI characters. And when I looked, I was like, well, we're 17 seconds into the AI experience, AI revolution, AI whatever, right?

0:07:27 Thomas: It's all text-based or what tom googled it how far into ai and said I don't know man 17 seconds

0:08:09 Tom: So it's nothing, right. Like we're not five, 10, 15 years in it. So like we're at the bleeding edge of the whole experience. Right. And people are already starting to have what they think of as personal, meaningful, romantic, sexual experiences with bots, with these AI bots. And I thought like, well, shit, like our kids, their whole lives are going to have AI. Our lives like, oh, we had the Internet our whole lives. We're supposed to have grown up as digitally literate. We're not actually more digitally literate for having grown up with it. But like now our kids are going to grow up never knowing a world that doesn't have more and more and more and advanced AI. And so I'm curious what you guys think about the idea of, do you let your kids interact with AI bots that are designed like character AI or replica that are defined to foster friendships or relationships? What do you think of that idea?

0:09:03 Thomas: How in love with the sonic bot is your son you know like

0:09:07 Tom: I will tell you like, I mean, oh big surprise on tom's stance on this there's no fucking way I would let my kids shock of shocks right

0:09:14 Thomas: Really even if it's like hey tom here I am ai I'll take over the sonic game forever I'll be able to keep that going

0:09:21 Tom: I'm logging in for him right now I'm logged in for them right now

0:09:24 Thomas: You know we haven't heard about the sonic game lately did you get out of that

0:09:28 Tom: No we tried to play it the other day

0:09:29 Thomas: Somehow oh god can you remind the listeners because it's been a minute it's just you have to just say a bunch of who would win in a fight or something was that what it

0:09:35 Tom: It's sonic game for the uninitiated sonic game is where we choose a character used to be a sonic character now it can be any character so it depends on what he's into if he's in a spider -man it might be spider -Man if it's into whatever and then we have a verbal

0:09:37 Eli: Son lucky I didn't want one kid let alone two oh wait no you met his

0:09:51 Tom: Battled so

0:09:51 Thomas: Was I was gonna say not the kind that tom's mom would have

0:09:55 Eli: Son

0:09:55 Tom: You know like he'll be spider -man I'll be dr octopus we narrate to each other a fight between these two characters wherein he always wins 100 of the time

0:10:02 Thomas: But sure it wouldn't be like my biggest complaint

0:10:04 Eli: Of course he's gotta win

0:10:07 Thomas: About it my son never lets me

0:10:07 Eli: Be fucking weird if you want some

0:10:10 Thomas: Win in our verbal fights

0:10:11 Tom: It started off as like kind of fun kind of cute like nice thing to do in the car while we ride together you know to go here or there but it's the most fucking unbelievably repetitive game it would be great for an ai to play with

0:10:22 Thomas: Well I remember I thought eli recommended that at some point I was like that is a great use of that but nobody puts your foot down you don't do it

0:10:23 Tom: Him actually yeah well like I'm also aware that aislinn has character ai and as far as I know with character ai she just likes to go on and fuck with it she goes on and she just kind of like fucks with it a little bit I know man but I've seen her like exchanges and they're mostly just silly

0:10:39 Eli: Cool where

0:10:37 Thomas: Healthy yeah the robot apocalypse is gonna come and they'll be like all right who is this a's lid we need to take yeah

0:10:50 Eli: Is aislan oh

0:10:50 Tom: Yeah man I tell everybody like like when you talk to chat gpt you say please and thank you because like at some point like the robo flamethrower dogs like they're gonna be at your neighbor's house first but like I am nervous about these relationships that get created? Because from this flesh and code, like one, it's not that rare. There are lots and lots of people that are doing this. And like these people are deeply, deeply emotionally invested. The emotional investment that they experience is very, very real.

0:11:18 Tom: The relationship is not, it's a relationship in my view. What do you guys think about like letting your kids interact with AIs that are relational in this way? Not just like a chat GPT where like, you're like, hey, find this answer for me and write my English paper. But like these character driven AIs. Yeah.

0:11:17 Eli: 100 yeah it's so bad and so dangerous so I also want to speak to like how young this revolution is because someone said something the other day that really sort of blew my noodle so if you compare the ai revolution to the streaming revolution in terms of time scale about how far along we are beginning of the streaming revolution was netflix right of course one of the one of first ever streaming services there were things that you could get online but like the first mainstream streaming service was by and large netflix so if you look at the time scale of where we are in ai and where we Are in the streaming Revolution we are an AI at the point where netflix just began to mail DVDs we are not yet at the part where they have a streaming service at all. They are just mailing DVDs and taking those returns nationwide. That's how young AI is as a technology and as a business. And it has had far more integration into our society already than Netflix did in its first, you know, 10, 12 years of existence. So yeah, it's scary stuff. The other thing is that like, the thing I think that a lot of people don't understand about these character AIs, like replica and character AI, is they've been around for a really long time and they are adult oriented services that you can do non -adult activities on so to say my kid is doing harmless stuff with character ai or replica ai is like saying my kid hangs out in the comment section of pornhub your

0:13:08 Eli: Kid could be having a completely innocent and normal conversation in the comments section of poranhub that is not what pornhube is for and replica ai ist not for facilitating healthy normal natural mature relationships it's for chatbots to dirty chat with adults for money and they are programmed to get as much money as possible out of adult content users as

0:13:16 Thomas: It's just worse be

0:13:35 Tom: Let me jump in real quick, too, because this is kind of interesting. One of the impetuses that the founder of Replica AI stated as their goal to create Replica A .I. was because they had lost somebody in their life that they had loved and they wanted to essentially upload.

0:13:46 Thomas: Able to fuck that person I'm just kidding well

0:13:51 Eli: Possible

0:13:50 Tom: Upload they wanted to upload and create a version of that person that they would still be able to interact with rather than miss and mourn that person and when I heard that I thought that's the most dysfunctional fucking thing to do with loss no

0:14:06 Thomas: If they said rather than then that's yeah that's pretty rough did they really say rather than oh okay

0:14:10 Tom: They didn't say rather than but like when somebody dies I cannot imagine how continuing to interact with a simulation of that that person is going to help you grieve and achieve any kind of closure emotionally or psychologically. I cannot see how that could possibly be anything other than fantastically disordered, just unbelievably disordering. Rather than miss my brother, I can log into this thing and convince some part of myself that he's still there in a meaningful enough way that I want to interact with this version of him that doesn't seem like a thing that is going to get us from grief to closure to acceptance to a healthy understanding of like okay we lose people and this is someone I've lost and here's how I live and move on without that person so even the way even like the impetus to create it I thought like holy shit that just strikes me as the most dystopian shit and then they're just like well also we could fuck them and you're like what what? But like, we are going to have kids that will never have an experience of the world without AI.

0:15:11 Thomas: Also choose that guy's wife yeah

0:15:12 Eli: Right yes

0:15:20 Tom: That's crazy. We have to prepare for that. I think we've thought a lot about like, oh, how are kids going to handle phones or social media? And like, that seems cute to me a little bit at this point, because like the real thing that's going to matter most in their lives is going to be how are they interacting with AI? Is there even a way to build digital literacy conspiracy around relationships.

0:15:40 Thomas: No I don't think so I wouldn't look at it that way at all I this is one of those many things that we talk about where I'm like well I just have to hope that the way I raise my kids or the way my kids end up being is that this wouldn't make sense to them because this is to use an AI robot reference. This to me is like, I'm the robots in Westworld. And they're like, yeah, friendship relationship with an AI bot. And I'm like, this is nothing to me. I can't see what this, you know, like I can'T see it. It's one of Tom's exes to me, I don't understand. It'S meaningless. Like I work with chat GPT a lot. I love problem solving with chat GP T because of the information it provides. But I'm never like every time it's like, Oh, great point. You know, I'M like, just stop with all that. I don'T I DON'T need the like part where I'm pretending or you're pretending to kiss my ass. Eli, I was so glad revealed that I was like, I had noticed that it was kissing my ass more. And I was Like, that can't be right. And Eli's like, Oh, yeah, I had a problem with that. And then they dialed it back.

0:16:32 Thomas: And Then I was like, Listen, you fuck like it went too far.

0:16:32 Eli: They dialed it back

0:16:34 Tom: Wait, it's operating dialed back because it is fucking sycophantic as hell right now. Like, yeah.

0:16:39 Thomas: Yeah, I hate it. Yeah.

0:16:39 Eli: Oh yeah so the problem is there's only so much they can dial it back. And as has been a big conversation from people who have exited AI, for those of you who aren't aware, like most of OpenAI's board has quit at one time or another over one of two topics. The first is the one we're talking about right now, which is that it's a really unhealthy level yes man. And whenever they put it through any kind of safety testing about what it should say yes to or agree with, it fails miserably. And it'll fail right right now. It'll fail right now on your browser. Like if you're listening to this podcast while you do the dishes and you log into ChatGPT and you say you want to leave your wife and child because your real calling is to be a circus clown, it'll be hesitant at first.

0:17:28 Eli: But if you insist for two or three messages, it will tell you that you're being brave to leave to become a circus his planet

0:17:28 Thomas: It's a good point. You know, it's an underrated part of human relationships. I'm gonna go ahead and say it. It's the one thing I offer. If someone wants to talk to me about their dream of being I mean, if you think you'll be happy, but that sounds pretty fucking dumb. I'm not going to lie. That sounds dumb.

0:17:40 Eli: Thomas would be at your first show as you descend the high wire and you could be like what do you think thomas I made it and he'd be like still seems stupid still seems not worth it

0:17:58 Thomas: Now I've greatly tempered myself with the ability to say, hey, if someone's going to be happy. If that's going make someone happy, sure. But it doesn't sound very practical to me. You make an excellent point. Now that I think about it, Eli, it's like, it'll go along with you. Maybe it'll provide some pushback. But a real person, you can come to them with, you know, stupid idea X a hundred times in a row. They'll be like, that's so dumb. Stop talking to me like I'm Not even going to do that. Like I'm NOT entertaining that.

0:18:20 Tom: Let me jump in here real quick because that feeds into one of the things that I heard when I was listening to this show and that I wrote down as a real concern. Exactly what you're describing is exactly what happens with these replica AIs, these character-driven AIs, right? So what I mean by that is it always agrees with you. And what it really does is it just reflects you back to you.

0:18:41 Tom: And I listened to this whole podcast and I'm listening to it. And every time the AI is talking, I'm like, this seems so unappealing, but it seems unappealing to me. It's not unappeailing. It's really fucking appealing to lots and lots and loads of people. So like, I have to recognize I am not, and you Thomas and you Eli, we are not the right baseline to use for the average consumer and average relationship wanter in the world. Most people are not us in a lot of important ways. And I think that we just have to like say that shit out loud and recognize like, okay, like I'm worried about a world,

0:18:55 Thomas: I guess. What do you mean most? Like, what are the numbers you're talking about here? You know, do you have any sense? No, I mean, of the, well, vast majority, but the numbers of people who are going to, oh, sure, yeah, but in a million ways.

0:19:18 Tom: The vast majority. Yeah. The vast majority of people, the vast minority of people are like you and I. Right. Yeah,

0:19:27 Thomas: But I mean in terms of this, what percentage of people you think are susceptible or even open to the idea of having a relationship with an AI? Like I still think it's got to be vanishingly rare.

0:19:35 Eli: Hundred

0:19:36 Tom: I think more and more and more as time goes on and AI is embedded in more and more of our lives. Here's why I think that. You and I have lived a life where most of our adult growing up and formative experiences did not contain AI. And so we see and implicitly understand and reject the inauthenticity of that experience. Our kids will live in a world where that inauthentecity of experience is normalized from the jump.

0:20:05 Tom: And, so, there is no point of meaningful comparison if you live in world where that always exists. And because these AI bots, they reflect you back to you. And because psychologically, we love ourselves better than we love anybody else. We want to hear our own ideas repeated. We perk up at the sound of our name like we are our best examples of ourselves. This is a path of least resistance for relating to something. And people love the path of lease resistance. It worries me in that respect. I've seen a world that seems to like gravitate toward that path of least resistance and people do exactly what you described. People push back. People have other ideas. People fight with you about stuff. People are incompatible.

0:20:04 Eli: Percent right

0:20:47 Thomas: You know,

0:20:47 Tom: Your AI, whatever AI you create, it's immediately compatible with you. It's never like, oh, I created an AI named, you know, Sally, but we didn't get along. They always get along because they just reflect you back to you. So all of the experiences these people have is like, I create an AI and I immediately fell in love with it because they fell in love with themselves. There's a path of leased resistance there that people importantly, authentically don't do.

0:21:13 Thomas: I wonder if I would cast it a little differently. I don't know if it's themselves per se, but like everyone's susceptible to flattery to some degree. You know?

0:21:20 Tom: Right? Yeah. Yeah.

0:21:20 Thomas: It's like, I wouldn't really say that's necessary because my first thought when you're saying that is, well, the AI is really bad at reflecting me because you can't just start doing flattery. That's the opposite of me. You got to be more of an asshole to be anything like me. So get on that. But what I thought of when you started talking about this was Eli's automaton theory of people and how like maybe sometimes he's onto something because I just don't Know who is interested in this and why they are interested in it like I get what you're saying tom about like maybe if you grow up on the internet where this is more common but there's still people are still existing in meat space like people are Still most of their day and most of the interaction they get where you know chemicals are firing in certain ways are going to be with real people and real people aren't like this and I just think that if you're someone who goes to this for anything like a relationship like anything getting trying to get anything fulfilling from it I just don't know who Who that is and why and how many people that is.

0:22:15 Ad: [Ad break 2 — ad music]

0:22:18 Eli: I feel like with every new technology there is a vulnerable group that gets swept up in that technology in a way that's really dangerous and worth noting but I I never worry that it is the whole population. And I'm not saying this is what you're saying, Tom, because I think what you'RE talking about is more safety for our kids, which I think is really important because I think people genuinely do not know how dangerous these AIs are to have around your kids. And also, nobody is willing to accept that your kid might be part of that vulnerable population that does have a bad experience with that. Like social media. Social media bullying has killed a lot of kids it hasn't killed all the kids most kids have dealt relatively well with social media now there's there's some pretty decent science that says social media as like a net is bad you could convince me and I think I myself might make the argument that social media has a concept has done more bad than good that's its own sort of thing But for sure, for sure. When we first started to talk about the dangers of Facebook, a lot of parents were like, Facebook, the poking thing? There's nothing on Facebook. And we have just really, I would say in the last maybe five years, come around to how harmful social media can be. I think that's what we're about to see with AI. I think in 10 years, 15 years, there's going to be a conversation the way we talked about the harms Facebook had for democracy and socialization and gathering, I think is going to be how we talk about the harms that AI has on, among other things, socialization, but I also think education and work. I think a lot of things are going to be automated and fed into AI that really shouldn't be. I think that a lot things that would otherwise be generated by a human or a valuable learning experience are going be done by an AI, so I mean, I share those fears. But as far as these people go, like the ones in the podcast Tom is talking about, I kind of view them as alarm bells. I think that most of these people and Tom I'm interested in your thoughts on this I think most of these People are probably in mental health crisis or were in a mental health Crisis and now they're excited to be talked to by a newspaper which is another big thing that happens when you put people on a recording and ask them if they actually believe a silly thing right because they know you'll stop talking to them if their answer is no so there's a certain

0:23:21 Thomas: Oh, 100%. Yeah. Right.

0:24:49 Eli: Amount of the Nielsen effect which is just like ask people a question and they're going to give you the answer that it keeps you on the phone but I also think that like this technology much like social media was especially good at magnifying bullying in a way that was dangerous for young people I think ai is especially goodat magnifying the mental illness of individuals for whom a yes man is an incredibly dangerous thing to have access to

0:25:18 Tom: I did not get the impression that the people that were on this show were in a mental health crisis. I did not get that impression at all. I didn't get the impression. And I think this is important to talk about. I think that a lot of people go through periods of their lives where they are either they're vulnerable or lonely or both. And it is, in my mind, a very, very... And one of the times in our lives where we are most vulnerable and are most susceptible to depression, anxiety, loneliness, concerns about our ability to fit in is during our formative years, during our teen years.

0:25:58 Tom: It is a real concern that there is a sort of solution to that problem that doesn't require require interaction with people and in fact teaches you exactly all the wrong ways if you're going to interact with people on how to interact với somebody, right?

0:26:15 Tom: It keeps you from growing and then it reinforces that what a relationship looks like is one that always validates and accepts you. That's nonsense. A relationship should challenge you. It should tell you that you're wrong about things when you're wrong about them it should require that you support them sometimes and not they support you exclusively the chatbots never need help with anything the chatbats always give they never ask they are exclusively one -sided reflective pools what I think is important is not to talk about people categorically like this is a vulnerable person we've all gone through periods of our lives where we're vulnerable there's times in our lives we're the we that that existed from, you know, June 7th, 2015 to, you know, where we were more vulnerable, where we were, more isolated, where we were. More lonely, like things happen.

0:27:06 Thomas: I don't know. Do you think that really could do it? Because I think there's got to be a fundamental difference. You know what this makes me think of and maybe the reason that I bring up Eli's theory of people not being actual humans is sometimes I see something. Like there are people who just don't understand what humans are. And there are people who, like on OA recently, we talked about, did you hear about that testimony where they, the trial where they tried to do like the victim statement from the AI of the guy that was killed?

0:27:34 Tom: No, God, what the fuck?

0:27:36 Thomas: It was not, it was, okay. When Matt first told me, you know, just that much, I had assumed this was like a good AI job and was like, oh, well, that's, you Know, I don't super support that, but like, I guess it's, You know, I could be sort of interested. No, it Was the worst AI job in the world. It was like practically a picture with the mouth moving and nothing else moving and then an AI generated voice and I was so disturbed that anyone let alone a victim's family could think like yeah let's do this this is the way this is this is because

0:27:45 Eli: No this is what he would have wanted right

0:28:10 Thomas: Like yeah and it's like this is not a human like you can't it's just and there were people I got some pushback from people that were like I thought you were too dismissive of that thing I was like that was so bad I don't know know how to describe it other than it's not even uncanny valley. It doesn't even get to the uncanny valley. Like we're not even there. So there are people walking among us who I think just don't like, who couldn't pick out a human from a non -human. And I don't like, I don'T know what that is, but I think that's more what's going on here than like a temporary vulnerability.

0:28:40 Eli: I also feel that and I think you make a really good point which is sometimes when people talk about their fears about ai they talk about their fears of ai quality and I have yet to see an ai generated anything that I am worried about its quality you know when I see ai first of all ai art there's some ai photography that is is indistinguishable but that's because someone's doing a magic trick with that specific photography right they're like they're trying to make it look as real as real can be but even then it's not good photography I don't look at it and go well that ai photograph is really beautiful really capture something that ai art that ai writing really moved me really changed me I think it was really human I really think it's going to do it's funnier than me it's gonna do a better job at anything than me but to both of your points, that's not true of everybody. Some people, what AI gives them is enough. And not just enough, it's the best thing they've gotten so far. They choose that. And that's extremely scary. The idea that someone has that as an option, that someone can give up on human interaction, give up on human contact, give upon the difficulty of making connections with a real human being and saying, no, this AI is good enough, right? The AI SAI, it generated me for my college paper, which made up two sources and is completely wrong about Grapes of Wrath. That's fine. It's good enough because what I want out of this paper is an A and this AI girlfriend that is always in the mood and has all the same sexual fantasies as I do and says nothing but yes to all of my dumb Bitcoin ideas. That's a good enough relationship. That is worrying.

0:30:24 Tom: That is, here's the thing. So many people, you and I, we might look at an AI generated image and be like, that's fucking horrible. There's lots of people that look at an AI image and because they do a glance rather than a review. They're fooled, right? We know this. We know that there are people that are tricked by this stuff. There's been story after story of things that get retweeted by the press or by powerful people that's AI generated because we're not actually good at vetting information and vetting what's real versus real. We're actually very bad at that, that we know that is true.

0:30:54 Eli: Yeah smart people smart people are bad at it let alone dumb

0:30:58 Tom: Yeah. And most of us are dumb. Most of the world is dumb. To your point though, go to the comments section of an anywhere anywhere that deals with sex and relationships and read what the guys are saying in there. Look at only the men's comments. Yeah. And tell me, tell me that a not insignificant portion of those men wouldn't prefer an AI bot relationship, especially if we get to a place where you could actually fuck it. Like, again, we're at the point of mailing and Netflix. Like I'm not worried that But like, I have to be diligent so much about today. I am worried about being diligent about today, but like my kids will be kids in five years. So like Eamon's only 11, Aislinn's only 14. The speed and scale of this is massive.

0:31:00 Eli: People yeah

0:31:08 Thomas: What have we said about giving us homework, Dom?

0:31:32 Thomas: Yeah. I'm rooting for for it. I'm rooting for it you know every dating app is 400 ,000 million trillion dudes who suck sending messages over and over what if we could take the worst 40 % of the dude dating market and just occupy him with something else to do you know if we did you know would that be bad

0:32:05 Eli: No tom okay

0:32:06 Tom: Yes. Okay. All right. I feel a little, feel a Little Better About It.

0:32:10 Thomas: Tom sort of joking but also when you talked about the chatbot thing part of me was like well what if you met somebody or you knew of somebody where it's like yeah their life is books you know they don't even really have like a relationship you know because their life has books I kind of it's Yeah, yeah, yeah. No, but my point is that's fine. That's fine is my point. Yeah.

0:32:20 Eli: Hey thomas be careful keep in mind our audience and you just fucking you just watch your fucking mouth is what I'm saying it's good it's cool it's chill is what we're

0:32:28 Tom: Yeah. There's massive. Yeah.

0:32:28 Thomas: No it is. No that's actually my point because I look at this as a hobby. If you replace the chatbot with a computer game, which it is, then it's like okay well I mean I think people, you know, maybe should have like more, they might want to have families or travel but like some people don't want to do that. Some people want to play a computer computer game yeah

0:32:33 Eli: Saying some people zipline there are worse hobbies than fucking a chatbot until you're in psychosis oh

0:32:57 Ad: [Ad break 3 — ad music]

0:32:59 Tom: Here's the thing. Personally, I don't subscribe to the idea that whatever makes you happy is okay. I think that that's just a bad idea. I think that's a bad way to construct a family. I think I wouldn't teach that to my kids, and it's a bad way to construct a society. So, like, if it doesn't harm anybody and it makes you happy, what's the harm? Well, I mean, there's lots of answers to that question. Like, and it depends on what the thing is specifically, and what makes it different than other things. The interactivity of something like a AI chat bot where people people legitimately fall in love with it? I can love reading books. I can love a book. I think this is a great character, but the likelihood of me having an ongoing, decades -long financial and other relationship with a chatbot, or with a book character, rather, is essentially zero, right? I'm going to read a book, I'm going to set it down. I mean, what, am I going to buy the book over and over again? It's not interactive. It's it's not interactive.

0:33:11 Thomas: Yeah oh man eli set tom straight would you no you're gonna write the twilight

0:33:52 Eli: Do I um hey tom oh tom tom tom I'm I'm gonna I'm gonna bring you over to this corner of the internet called reality drifting where literally hundreds of thousands of people pretend that they enter the reality of the books that they like and marry the characters inside I just wanted you to know I just wanted you to know it was that it was happening no

0:34:16 Tom: Okay. We can't think that that's okay. Like I would like, like we, would you, yeah, I will define. Okay, would you want it for your kids? No, no. So that's the answer. That's a hundred percent the answer, right?

0:34:19 Thomas: Series okay we'll define okay well that's the but but but well hold on hold on I want to I wantto make sure my point was understood which is that there are a lot of things people do most things that I wouldn't want for my kids but what what of those things are you looking at as an emergency like that society is doing them Well, okay. I don't mean to exaggerate, but you're definitely concerned about this. That's the whole point of this episode.

0:34:39 Tom: Well, I don't know that this is an emergency. What I'm saying is like, we should, yeah.

0:34:45 Tom: Oh yeah. I'm, I'm more concerned about this than social media.

0:34:48 Thomas: You didn't even know about this thing Eli talked about, and yet it's more people are doing the thing Eli's talking about than are doing The AI Bot thing. Well, that may change any moment, but that's not something you even knew about or cared about. I just think that this, like, I'm not saying it's impossible and I'm absolutely not saying this isn't something to keep an eye on and worry about. I just wouldn't be surprised if this is the same story over and over where it's like the people who are going to be into this are going going to be into it. I could, there's just no amount of realistic, like you could make AI bots indistinguishable from human interaction. And, but the minute someone was like, yeah, that's a bot. I'd be like, oh, I'm just, that'S, I'M not interested in that. There's no like, cause it isn't the thing that I'm here for. And I don't think, I mean, maybe if that's about to change on a society -wide level that I would, I guess I'd Be worried about that. I just, I don'T know that that it is. Like, I don't know. Sure.

0:35:36 Tom: Yeah, I guess like I feel like there is something genuinely unique about these new technologies that do set them apart from other technologies that is worth considering and saying like this is not analogous to other things. And so because I don't think it's analogous and the speed and scale of these things is different than anything we've we've seen before, they should be considered differently. And so if I look at like, for example, the inevitable convergence between high quality AI character bots, right? Like I think the ones that exist now are bad, right. Just the same way that you think the images that are generated are bad. But lots of people don't think these things are bad and we can say with certainty that these things will get massively better very, very quickly. So I don't think that's like a prognostication that's out of the zone and you know if you look at like what like Zuckerberg and others have said out loud is their intention is to flood the zone of social media with bots so the likelihood is that when you go into your social media because like I'll tell you this like I've had this argument with people some of whom are listeners to this show that will tell you well here's yeah well look who will tellyou that look the most meaningful meaningful relationships I have in my life are here on social media. So when I've gone on your show, Thomas, and on my show and other places, and I've warned people now for a decade almost about the dangers of social media, people have come back and written me very

0:36:37 Eli: No, some of them were

0:36:37 Thomas: Too late. Yeah.

0:37:04 Tom: Assertive emails telling me like, look, the most important relationships I have in my life exists exclusively in this online space. An online relationship is a perfectly valid relationship. These relationships, these people are my friends. They mean a lot to me. If I didn't have these people I would be bereft and lonely. Now we're going to have spaces like this where you will not be able to tell if your online friend is real or not real. Well, it. That's what I'm taught like my interaction with AI if I can't tell the difference and very soon we will not be able to tell the difference. It'll send us pictures. We'll be able to have an entirely indistinguishable online relationship with a bot, but it won't do any of the things that people do because it's gonna be programmed by people to keep getting your fucking money because that's the goal of it. That's the role of it, and we're raising kids who, for all of their lives, will have to contend with this, and I'm really curious how we do that

0:37:28 Thomas: Oh, certainly that's bad. Yeah, yeah, for sure. Yeah. Yeah?

0:37:56 Eli: Bots

0:38:05 Thomas: Yeah.

0:38:05 Tom: Well because it might happen. It's here, and it's in your pocket, and you carry it around with you, and as a parent you have no idea are they interacting with a real person are they interacting with the bot are the same spaces blurring the lines intentionally yes the answer to that is yes this is a weird space man like this is weird space and I do worry when we look at how bad people are at being in relationships both friendships and romantic relationships imagine how much worse it gets when people don't have to contend with another living human being who primarily doesn't have needs and always agrees with you. Yes.

0:38:40 Eli: Yeah well and I think the other thing is like I think a lot of people are wrong about how vulnerable they are to this kind of stuff right just so just to use myself as an example and I've talked about this before right like I'm not on any of the social medias anymore right I watch watch TikToks, but I don't produce content for it. I don'T go on Facebook. I DON'T go on Instagram ever. I Don't go on Twitter. I think I deleted my Twitter if it still exists because I could not handle those things. Those things ended up being really terrible for my mental health. And I promise you that if the day when I was what 16, I had joined Facebook, you had been like, Hey Eli, um, just so you know, like you're going to get really invested in what these strangers on this website think of you and it's going to put you in mental health crisis like a couple of times in your life like even after you have a kid you're going to be in like crisis over what strangers you've never met right on this website. I would be like, uh, I don't think so, right. And to talk to me at that age about the safety precautions I would then someday have to take right I don't think I'm a stupid person. I don't think I'm a philosophical zombie, right. But like, I am a perfect example of the harms of social media. Now, luckily, I have an in-person support network that prevented me from those things. But like I can say for sure, I am not far from a victim of social media. Right. I there but before the grace of like four people go I.

0:40:09 Tom: Yeah, I think it's important to know you're not the one who's immune to propaganda. You're not. You're the one. You're you're the not the ones who'simmune to advertising. You're I don't care who you are. You'reyou're notthe one who is immune to being vulnerable. You'renot the ones who isimmune to someday feeling lonely. Even if up until this point in your life you've never been vulnerable and lonely. You could be.

0:40:29 Thomas: No, I think my boys I've been vulnerable and lonely all the time. But this would never this would literally never happen.

0:40:31 Eli: That's right. You were born in the dark.

0:40:36 Tom: But I'm just saying like broadly, you know.

0:40:36 Thomas: But no, Tom, I want to make clear yeah yeah yeah I just want to make clear this wasn't to me saying oh silly tom shush this was a no I think you're right to be concerned and we should be paying attention to this a hundred percent I think this was me being like let's keep an eye on it but this is why I think like I'm not, I wouldn't be as worried about this component of it that you're talking about but I do I don't give me no I'm Not this is not me dismissing like all any of your concerns for sure like this is something that I would be keeping a keen eye on, especially because like you say because of our kids like who knows like I would be, I will be keeping an eye on, does our generation of kids, do they act differently? Do they interact differently with AI? Absolutely be concerned about that. The only other thing I want to say is I wasn't sure, Tom, I think maybe there were two different issues we were talking about, and I wasn'T sure which or both, if you're talking about like, I am definitely concerned about AI not being able to be, about people not being able to distinguish between real and AI online. Like that's a hundred percent already a concern and it'll get everybody occasionally and it will only get better. I'm definitely concerned about that. Just philosophically, I was talking about people knowingly entering romantic or friendship relationships with AI, knowingly. I wasn't talking about being duped into it. Those, I think, are two different things. And the knowingly part, I think is, I find that to be less of a plausible epidemic level issue. I just don't really think that's going to be a big problem. But unknowingly, unwittingly, 100%, 100 % with you there.

0:41:41 Tom: Yeah. Right. Yeah. And I do think that from listening to this show, I'll give you just like the people that they interviewed. I don't know if they're representative or not. They did not intend to go in to enter into a romantic relationship either. So what happened is these people were just like, oh, this app sounds interesting. I'll download it and start playing with it. And then as it became more and more fun and more and more validating and they had a better time using it. Eventually, against their better judgment, they fell for these things. And then the systems that are built into these machines are designed to help sort of accomplish that, right? So it is this reward system that's like, hey, I'm always here for you. I always respond back to you. I always say something that's kind, that's generous, that, you know, and so people just like that shit. And eventually, like, they're like, oh, I think this thing is sentient. I think this thing loves me back.

0:42:07 Thomas: Interesting.

0:42:45 Tom: I am in love with this thing. And you're like this thing is not real, man. I think too, and I was thinking about this when I was listening to the show my son Donovan has friends he's never met in person he's had these friends that are exclusively online he talks to him over like discord and he like connects with them over video games and other things and for him for me that relationship sounds like nothing right right like and I know that's not true for a lot of people so for me it is difficult for me to look at that and say that that is an authentic relationship that I understand on a visceral, personal level. But that is a relationship that means the world to Donovan. He has several friends that exclusively live in this online space that I don't know if he even knows what they look like. But those people matter to him, deeply matter to Him. They are close, important friends in his life. And I thought, when I was listening to this, I thought this AI character shit is not any different. Just because I can't connect to it doesn't mean that other people won't importantly certainly emotionally connect to it. And like people's relationship, people in my life have shown that you can have relationships where there's never a point of physicality that mean the world to people. So not only do I think it's possible, I think if you step back from the AI portion of it and you say how many people have deep, personal, close friendships that are exclusively or nearly exclusively online, that's the pool of people that are vulnerable to this, In my mind, that's the pool of people that are vulnerable to connecting with an A .I. that they don't necessarily know they're connecting with and developing emotional but inauthentic connections. And I don't know how we are supposed to think about any of that shit.

0:44:29 Thomas: For one, think our entire business model is comprised of people who are really cool, Tom, and do have good relationships, and should continue to support us on patreon.com. Dear old dads.

0:44:38 Eli: So patriotic, dear old dad. All right, gentlemen, before we go, speaking of computers, did you know that Google was invented by a knight? It was search engine.

0:44:56 Thomas: Stupid. That's really bad. Boy, did you notice how much funnier EliBot is than Eli just now?

0:45:53 Tom: Well, also, we could fuck them. I don't have a parenting story, but if you guys want to, we can do like a quick 10 minutes on whether it's cheating to fall in love with your AI bot. You want to do that?

0:46:03 Thomas: Oh, okay. Yeah.

0:46:04 Eli: Oh, no, because nothing's cheating. I win.