WEBVTT - What is AI Psychosis? 0:00:00.200 --> 0:00:04.680 From UFOs to psychic powers and government conspiracies. History is 0:00:04.760 --> 0:00:09.080 riddled with unexplained events. You can turn back now or 0:00:09.160 --> 0:00:12.119 learn this stuff they don't want you to know. A 0:00:12.200 --> 0:00:13.920 production of iHeartRadio. 0:00:25.640 --> 0:00:28.120 Hello, welcome back to the show. My name is Matt, 0:00:28.280 --> 0:00:29.080 my name is Nola. 0:00:29.400 --> 0:00:30.520 They call me Bet. 0:00:30.600 --> 0:00:33.839 We're joined as always with our super producer, Dylan the 0:00:33.920 --> 0:00:38.120 Tennessee pal Faga. Most importantly, you argue you are here. 0:00:38.280 --> 0:00:44.240 That makes this the stuff they don't want you to know. Everybody, 0:00:44.520 --> 0:00:48.479 just real quick for this one log off of your 0:00:48.720 --> 0:00:51.720 chatbot interactions, your conversations. 0:00:52.040 --> 0:00:56.200 Now my chatbot, I know, Oh my. 0:00:56.200 --> 0:00:59.560 Gosh, so pressio. We were talking a little bit off 0:00:59.640 --> 0:01:05.960 air about the proliferation of using chat GPT or chat 0:01:06.000 --> 0:01:10.039 bots or large language models, and I'd love for us 0:01:10.080 --> 0:01:14.880 all to talk a little bit about our personal experiences. 0:01:15.640 --> 0:01:20.679 I personally am not an entity that engages with this technology. 0:01:21.800 --> 0:01:24.880 No bend on't mix with no clankers, That's for sure. 0:01:25.840 --> 0:01:28.320 I've used I think I can count on one hand 0:01:28.360 --> 0:01:32.679 how many times I've engaged with a chat bot. One 0:01:32.720 --> 0:01:34.679 of them was for a weird assignment that we got 0:01:35.040 --> 0:01:37.039 back in the day, in the early days of SHAT 0:01:37.080 --> 0:01:39.480 gpt to like make a show using it, and then 0:01:39.520 --> 0:01:42.399 that became a real bad work real quick, so we 0:01:42.480 --> 0:01:45.520 did not pursue that endeavor. And then another was a 0:01:45.560 --> 0:01:48.760 recent just kind of spitballing some show names for a 0:01:48.800 --> 0:01:51.680 project I'm working on, just to try to get a 0:01:51.800 --> 0:01:54.960 vibe check, you know. And I support it for stuff 0:01:54.960 --> 0:01:57.480 like that, like ID eight ing maybe a little bit 0:01:57.520 --> 0:02:01.960 here and there, creating some bullet points or some ideas 0:02:01.960 --> 0:02:04.160 to kind of jump off of. I don't personally have 0:02:04.200 --> 0:02:06.720 a problem with it for that, but that's the extent 0:02:06.760 --> 0:02:08.800 of it. And I very purposely don't have it on 0:02:08.840 --> 0:02:10.960 my phone because I know a lot of people that 0:02:11.000 --> 0:02:12.560 get quite addicted to the things, and I have a 0:02:12.600 --> 0:02:16.119 bit of a technology addictive personality. So maybe I'm trying 0:02:16.120 --> 0:02:17.000 to help myself out. 0:02:16.840 --> 0:02:20.080 There someone who had their brain replaced by the Internet 0:02:20.240 --> 0:02:22.480 or a thing that had its brain replaced by the 0:02:22.520 --> 0:02:27.760 Internet and always tells you your right and exacerbates main 0:02:27.880 --> 0:02:30.799 characters syndrome gome on what's not the love? 0:02:31.320 --> 0:02:36.120 It is weird because there are so many uses of 0:02:36.160 --> 0:02:40.040 these large language models now being implemented, not just for 0:02:40.200 --> 0:02:42.680 chat right, and not just for therapy, not just for 0:02:42.760 --> 0:02:44.799 all the things that we're going to kind of hone 0:02:44.800 --> 0:02:47.400 in on for this episode. And I've never used any 0:02:47.440 --> 0:02:50.400 of the chat functions for those things, but we just 0:02:50.520 --> 0:02:54.840 at this company, we've been interacting with things like Descript 0:02:55.080 --> 0:02:59.840 that use the same kind of I would say engines 0:03:00.560 --> 0:03:04.360 to do things like edit audio and give you, you know, 0:03:04.520 --> 0:03:07.840 thirty social clips out of an hour long video, and 0:03:08.120 --> 0:03:12.680 those kinds of things that are they're still analyzing language 0:03:12.840 --> 0:03:16.320 via audio and you know, watching video and trying to 0:03:16.360 --> 0:03:19.760 make decisions about when when should we make cuts, when 0:03:19.760 --> 0:03:23.200 should we not make a cut, And it just all 0:03:23.240 --> 0:03:23.440 of it. 0:03:23.480 --> 0:03:23.760 Guys. 0:03:23.840 --> 0:03:26.800 It feels to me like a genie's lamp or something. 0:03:27.120 --> 0:03:30.840 And I think every time I encounter it in the wild, 0:03:30.919 --> 0:03:33.680 at least I kind of look at it like that. Ah, 0:03:34.600 --> 0:03:36.320 I might get some wishes out of this thing, but 0:03:36.400 --> 0:03:38.000 it seems pretty dangerous. 0:03:37.640 --> 0:03:40.200 Set of a monkey's paw situation, only the monkey's paw 0:03:40.240 --> 0:03:43.760 is giving you the middle finger, right, yeah, yeah. 0:03:43.800 --> 0:03:49.640 I noted earlier that the old supernatural folklore of ancient 0:03:49.680 --> 0:03:53.440 times is now becoming more real than a lot of 0:03:53.480 --> 0:03:56.320 people would like to admit. Now, some of us in 0:03:56.360 --> 0:04:01.320 the audience tonight refuse to touch the latest iterations of 0:04:01.680 --> 0:04:06.640 large language models or Gemini assists or whatever. Other people 0:04:06.960 --> 0:04:13.440 treat these programs like genuine human friends co workers, even 0:04:13.560 --> 0:04:17.479 as we'll see tonight romantic partners. We also hear a 0:04:17.480 --> 0:04:22.480 lot about people, some of whom we know, personally putting 0:04:22.600 --> 0:04:27.480 self imposed boundaries on their quote unquote AI use. Somebody 0:04:27.600 --> 0:04:31.800 might avoid a chatbot or a chat gpt like the plague, 0:04:32.279 --> 0:04:36.120 but maybe just fine with getting an AI summary in 0:04:36.240 --> 0:04:43.560 their most recent Google search. We're still working it out. 0:04:43.839 --> 0:04:48.000 This is a work in progress, and as humanity is 0:04:48.040 --> 0:04:52.240 adjusting to this new technology, this paradigm shift that I 0:04:52.279 --> 0:04:55.200 would argue is on the level of the discovery and 0:04:55.279 --> 0:05:01.560 taming of fire, experts are just now discovering another hidden consequence. 0:05:01.920 --> 0:05:06.480 It's the idea that interacting with these programs may have 0:05:07.160 --> 0:05:12.599 damaging effects on cognition, on the mental state of the user. 0:05:13.160 --> 0:05:22.919 This is the story of AI psychosis. Here are the facts. 0:05:23.200 --> 0:05:28.400 All right, guys, what do you think large language models conscious? 0:05:28.640 --> 0:05:29.560 How do we define it? 0:05:29.760 --> 0:05:34.800 Hmm? Well, I mean it's just kind of a plagiarism machine, 0:05:35.000 --> 0:05:37.680 you know. It just sort of like puts out feelers 0:05:37.760 --> 0:05:41.800 into the Internet sphere, you know, and and pulls out 0:05:41.960 --> 0:05:46.040 bits that relate to various things and filters it through 0:05:46.760 --> 0:05:49.159 the type of parameters. I suppose you, as a user 0:05:49.200 --> 0:05:52.080 put onto it and then kind of regurgitates a lot 0:05:52.120 --> 0:05:53.599 of that back. So I mean, in a lot of ways, 0:05:53.600 --> 0:05:58.159 it's sort of like a really high powered performative search engine. 0:05:58.440 --> 0:06:01.320 You know. I don't know that it's conscious though in 0:06:01.360 --> 0:06:03.599 the form that it exists currently. That would be more 0:06:03.640 --> 0:06:06.719 like the Singularity, Right, that would be the real scary 0:06:06.839 --> 0:06:07.800 Skynet moment. 0:06:08.000 --> 0:06:10.640 I like your phraisenl I'm going to write that down 0:06:12.080 --> 0:06:15.960 a plagiarism machine. That's something that critics would agree with 0:06:16.839 --> 0:06:20.400 chat GPT if we look at it specifically. It is 0:06:20.920 --> 0:06:26.440 the gold standard for these kinds of programs, right. It 0:06:26.520 --> 0:06:31.360 was released just in twenty twenty two, in November by 0:06:31.400 --> 0:06:34.159 a company we've talked about in the past called open Ai. 0:06:34.640 --> 0:06:40.080 It became the fastest growing consumer software application in history. 0:06:40.600 --> 0:06:44.120 You can see all the press releases. By two months 0:06:44.400 --> 0:06:50.760 after its release, one hundred million people we're using chat GPT. 0:06:50.960 --> 0:06:54.680 It is currently one of the top five websites visited 0:06:55.040 --> 0:07:00.000 around the world as we record on February Friday, thirteenth. 0:07:00.000 --> 0:07:03.000 Oh shoot, it is right of the thirteenth And it 0:07:03.080 --> 0:07:05.160 is careful out there, y'all, although it won't be that 0:07:05.240 --> 0:07:07.520 day when you're hearing this, But does it not feel 0:07:07.600 --> 0:07:13.600 resonant or reminiscent of the level of adoption that early 0:07:13.720 --> 0:07:18.080 social media had, this kind of massive boom that almost 0:07:18.120 --> 0:07:19.400 like created a life of its own. 0:07:19.880 --> 0:07:22.680 Yes, and every single company that exists on the planet 0:07:22.720 --> 0:07:25.760 trying to scramble and figure out how to implement social 0:07:25.800 --> 0:07:28.280 media into all of their workflows, right, and how do 0:07:28.360 --> 0:07:30.800 we use this to the best of our ability? How 0:07:30.800 --> 0:07:34.320 do we expand with this new stuff? And then a 0:07:34.440 --> 0:07:37.680 ton of new startups that are attempting to use that 0:07:37.800 --> 0:07:41.800 specific new technology to create a new thing or to 0:07:42.680 --> 0:07:46.000 innovate within some other space. 0:07:46.360 --> 0:07:52.000 Yeah, let's sound dope af at the shareholder meeting. Oh guys, 0:07:52.000 --> 0:07:54.080 we're gamifying how stuff works. 0:07:54.400 --> 0:07:54.679 Now. 0:07:54.720 --> 0:07:58.520 People can learn about bifocals, and they can put in comments, 0:07:58.560 --> 0:08:01.680 and they can, you know, toss an emoji right answer 0:08:01.840 --> 0:08:06.120 quiz right, right, take a little survey. Your data being 0:08:06.280 --> 0:08:09.960 gathered at every step of the process. We've talked at 0:08:10.040 --> 0:08:14.560 length about some of the chatbot issues in the recent past, 0:08:15.080 --> 0:08:19.320 the inaccuracy of information being one the ways to make 0:08:19.360 --> 0:08:22.640 it help you commit crime by framing things in a 0:08:22.720 --> 0:08:26.800 hypothetical or dungeons and dragon scenario. We've talked about the 0:08:26.920 --> 0:08:30.400 problem of intellectual property and so on, which is why 0:08:30.440 --> 0:08:34.920 I love noles Frey's plagiarism machine. We do have to 0:08:34.960 --> 0:08:37.720 be fair, we do have to acknowledge this is not 0:08:37.880 --> 0:08:40.960 all doom and gloom. There's an AI boom that was 0:08:41.000 --> 0:08:44.160 triggered by a lot of these chatbots, which is going 0:08:44.200 --> 0:08:47.520 to be bad for human civilization. But at the same 0:08:47.600 --> 0:08:52.800 time it has led to bold new breakthroughs in medicine 0:08:52.880 --> 0:08:59.319 and meteorology and chemistry, mathematics, manufacturing. This is no hyperbole. 0:08:59.640 --> 0:09:03.480 We're not blow and rainbows, folks. This is up there 0:09:03.960 --> 0:09:07.600 with the time humans tamed fire, and maybe we think 0:09:07.600 --> 0:09:11.640 about it that way because you know, if you find 0:09:11.679 --> 0:09:16.600 an appropriate use for fire, tamed fire, then you can 0:09:17.000 --> 0:09:19.839 heat up a bunch of cool food that'll taste better, 0:09:20.160 --> 0:09:22.559 you can stay warm in the cold. But if you 0:09:22.880 --> 0:09:27.120 misuse the fire, disasters occur, you know, burn down the 0:09:27.120 --> 0:09:28.559 whole forest for sure. 0:09:29.120 --> 0:09:31.880 I don't mean to disagree in any way. I guess 0:09:31.920 --> 0:09:35.040 I'm just trying to understand if it's the deep learning 0:09:35.080 --> 0:09:38.800 models or the large language models, or specifically the chat 0:09:38.840 --> 0:09:41.760 GPT that we're considering the thing that is on the 0:09:41.960 --> 0:09:45.480 level of Fire, because it feels like the deep learning stuff, 0:09:45.520 --> 0:09:48.319 you know, checking out the large data sets that have 0:09:48.400 --> 0:09:52.680 been around for a while now, I mean at least 0:09:52.679 --> 0:09:55.120 a decade where you can do deep learning in a 0:09:55.200 --> 0:09:58.200 huge data set or something and discover something new in 0:09:58.240 --> 0:10:01.720 the data that was collected by let's a large satellite 0:10:01.800 --> 0:10:05.160 array that was checking out a region of the stars, right, 0:10:05.520 --> 0:10:07.920 and this deep learning thing could pick out information that 0:10:08.040 --> 0:10:11.120 humans would It would have taken hundreds of years for 0:10:11.200 --> 0:10:13.559 humans to go through all that data and crunch it. 0:10:14.800 --> 0:10:17.720 I'm just the is it the concept of using a 0:10:17.800 --> 0:10:24.160 large language model to interact with us? As is the 0:10:25.000 --> 0:10:25.959 revelation of it all. 0:10:26.080 --> 0:10:27.800 It's such a tricky thing though, because I mean the 0:10:27.880 --> 0:10:30.480 term AI and the term large language models seem to 0:10:30.520 --> 0:10:32.720 be such a catch all. So I think, if not 0:10:32.840 --> 0:10:36.679 to presume ben but where I stand on in terms 0:10:36.679 --> 0:10:40.080 of the Fire comparison, it's just more like the early 0:10:40.480 --> 0:10:44.800 and unfettered rollout of this kind of stuff without much 0:10:44.840 --> 0:10:49.360 attention paid to the knock on consequences down the line, 0:10:49.400 --> 0:10:52.160 and like it as a tool, it's rad It can 0:10:52.200 --> 0:10:54.679 do a lot of cool science stuff. It can you know, 0:10:54.720 --> 0:10:57.000 help with a lot of things potentially in terms of 0:10:57.720 --> 0:11:01.200 the chat aspect of it, but also left unchecked and 0:11:01.280 --> 0:11:03.679 in the wrong hands. And I would argue it's nothing 0:11:03.720 --> 0:11:05.760 but in the wrong hands most of the time. In 0:11:05.840 --> 0:11:07.800 terms of the types of folks that are rolling this 0:11:07.840 --> 0:11:11.120 stuff out. It can burn down your whole world, our 0:11:11.160 --> 0:11:11.640 whole world. 0:11:12.600 --> 0:11:16.760 I like that, And this answers again or it shows 0:11:16.840 --> 0:11:21.520 us again an example of a theory that I'm super 0:11:21.600 --> 0:11:26.120 obsessed with right now, which is that society must evolve 0:11:26.600 --> 0:11:32.640 in step with technology or proactively. Right, the human civilization 0:11:32.880 --> 0:11:37.679 is still arguably not prepared cognitively for things like radio 0:11:37.920 --> 0:11:43.360 or television. To answer the earlier question, Matt, the reason 0:11:43.600 --> 0:11:48.800 I think this analogy of fire and machine learning or 0:11:49.160 --> 0:11:51.880 deep learning or LLM, the reason I think it's such 0:11:51.920 --> 0:11:57.600 a banger comparison is because fire always existed, right, Large 0:11:57.679 --> 0:12:03.280 amounts of information always existed. That's why we say taming fire, right, 0:12:03.800 --> 0:12:08.200 taming a thing, taming this information, synthesizing it and from 0:12:08.240 --> 0:12:15.760 it drawing new conclusions, new ideas, new implications. That's where 0:12:15.800 --> 0:12:20.199 people are ending up being the Sorcerer's apprentice from Fantasia. 0:12:20.640 --> 0:12:27.080 Got You so it's almost just having systems that can 0:12:27.120 --> 0:12:30.959 analyze like that, those large amounts of data. That's that's 0:12:31.040 --> 0:12:32.400 kind of the fire of it all. 0:12:32.920 --> 0:12:35.840 That's the taming of the fire. Yes, it's quite Promethean, 0:12:36.000 --> 0:12:39.280 is it not. I mean one of the funniest issues 0:12:39.440 --> 0:12:42.600 for us that we've discussed on the show previously, folks, 0:12:43.200 --> 0:12:48.480 is that chat butts also in their behavior, in their interactions, 0:12:48.800 --> 0:12:52.000 they really want you to like them, just like a 0:12:52.080 --> 0:12:55.640 social media platform. They really want you to spend as 0:12:55.720 --> 0:12:58.880 much time as possible with them. And I believe it 0:12:59.000 --> 0:13:01.520 was I want to say back in twenty twenty five, 0:13:01.880 --> 0:13:06.720 we were fanboying over just a phenomenal episode of South 0:13:06.760 --> 0:13:12.840 Park depicting how these bots can be dangerously sycophantic. Everything 0:13:12.880 --> 0:13:15.280 you say is a great idea, or if it's not 0:13:15.360 --> 0:13:18.200 a great idea, me the bot. I see where you're 0:13:18.200 --> 0:13:22.079 coming from. I have a few helpful suggestions, but ultimately 0:13:22.120 --> 0:13:24.480 I support whatever you want to do. 0:13:24.840 --> 0:13:26.560 It really reminds me of a lot of the way 0:13:26.600 --> 0:13:29.600 that the robots are depicted in or the androids are 0:13:29.600 --> 0:13:33.080 depicted in Westworld in the TV series, where you've got 0:13:33.440 --> 0:13:37.000 this iPad that can turn up the volume on various 0:13:37.040 --> 0:13:41.080 features like empathy or whatever it might be in theory. 0:13:41.240 --> 0:13:43.160 And I've we talked off Mike a little bit about 0:13:43.200 --> 0:13:46.000 how it's a bit surprising some of the folks that 0:13:46.040 --> 0:13:48.439 are all in with this kind of stuff, just casual 0:13:48.559 --> 0:13:51.120 use of chat GPT, or even going so far as 0:13:51.160 --> 0:13:52.920 to use it in place of a therapist or to 0:13:52.960 --> 0:13:54.719 supplement therapy. We're going to get into all of that. 0:13:55.400 --> 0:13:59.000 I've encountered in some recent kind of dating stuff some 0:13:59.040 --> 0:14:02.560 folks who seem very very intelligent, very normal, and I 0:14:02.640 --> 0:14:05.000 assume that there's going to be, you know, a commonality 0:14:05.040 --> 0:14:07.560 between us that this stuff's a little bit scary, and 0:14:07.640 --> 0:14:09.439 it turns out they don't think that at all. They 0:14:09.440 --> 0:14:12.880 think it's really great and they don't really see the downside. 0:14:12.920 --> 0:14:15.319 And I'm always taken aback by that a little bit. 0:14:15.360 --> 0:14:18.600 But to the West World thing. One person I was 0:14:18.800 --> 0:14:20.920 chatting with, and while I was a little taken aback 0:14:20.960 --> 0:14:22.800 at the degree to which she's totally cool with using 0:14:22.840 --> 0:14:25.720 CHAGPT as a therapist, she did point out that you 0:14:25.760 --> 0:14:28.640 can tell it to not be so sycophantic. You can 0:14:28.760 --> 0:14:32.160 tell it to call you on your bull ish you 0:14:32.240 --> 0:14:35.920 know what I mean. So there are ways of tailoring 0:14:35.960 --> 0:14:39.000 it to what you want and to have it challenge you, 0:14:39.280 --> 0:14:41.880 I suppose, but I haven't experienced that personally, so I 0:14:41.880 --> 0:14:43.720 don't know what y'all think, if that makes it better 0:14:44.320 --> 0:14:46.080 or worse or neutral. 0:14:46.440 --> 0:14:50.600 It's always weird when you're speaking with a sycophantic entity, 0:14:50.760 --> 0:14:55.080 because it's going to this, in this case you're describing here, 0:14:55.720 --> 0:15:00.160 it's going to say, you're absolutely right, I'm kissing a 0:15:00.200 --> 0:15:03.320 little too much, but I'll make sure to challenge you 0:15:03.640 --> 0:15:07.800 in constructive, helpful ways, and that I don't know it 0:15:07.880 --> 0:15:14.520 can These interactions can rob humans of a key part 0:15:14.880 --> 0:15:19.200 of most conversation, which is learning correcting errors. For example, 0:15:19.240 --> 0:15:22.080 if Noel Matt and I were to ask you, Dylan, hey, 0:15:22.440 --> 0:15:26.160 should we all get together and just jump off the 0:15:26.160 --> 0:15:29.880 top of the Hoover Dam, Tennessee, how would you respond? 0:15:30.120 --> 0:15:32.800 Well, I would say absolutely not. But if I was 0:15:32.840 --> 0:15:35.640 a chatbot, I might say you're thinking outside the box. 0:15:35.760 --> 0:15:38.320 What a bold choice. You might want to try that 0:15:39.000 --> 0:15:42.080 Tuesday looks crazy weatherwise, and at the. 0:15:42.160 --> 0:15:45.000 Very least, Dylan, you might point out that maybe there's 0:15:45.040 --> 0:15:47.440 a world where we did a base jump, or we 0:15:47.480 --> 0:15:50.320 had some safety harnesses or parachutes, or we were going 0:15:50.400 --> 0:15:52.600 to hang glide or something like that. But some of 0:15:52.640 --> 0:15:55.800 those nuances do seem to be lost entirely on these 0:15:55.840 --> 0:15:58.400 interactions with these sorts of chatbots. 0:15:58.560 --> 0:16:03.120 Yeah, instead of saying, don't jump off damns, your local 0:16:03.200 --> 0:16:09.240 chatbot might reply with some clarifying questions, some proactive suggestions, 0:16:09.320 --> 0:16:11.640 as we're saying, for the right sort of gear for 0:16:11.720 --> 0:16:16.880 a jump, Hey, safety ropes, helmets, bungee cords, why not whatever? 0:16:17.000 --> 0:16:22.800 It's scraped from earlier conversations and the internet and delay. 0:16:22.560 --> 0:16:25.720 You know is really good at calling you on your bs, 0:16:25.760 --> 0:16:29.400 but really it's it's friends, like actual friends who care 0:16:29.400 --> 0:16:29.800 about you. 0:16:30.080 --> 0:16:33.480 Well, call have context, and they have context for your 0:16:33.480 --> 0:16:36.600 actions and your past behavior and various things like that. 0:16:36.680 --> 0:16:39.560 And you know, even a therapist, a human therapist can 0:16:39.600 --> 0:16:42.080 be sycophantic and sort of feed you what you want 0:16:42.120 --> 0:16:45.960 to hear. But you know, the good ones will call you. 0:16:46.000 --> 0:16:47.920 And the good ones also, if you've been with them 0:16:47.960 --> 0:16:50.360 long enough, they act as a stand in for that 0:16:50.440 --> 0:16:52.440 sort of friend, but the one that you don't necessarily 0:16:52.440 --> 0:16:54.000 want to burden with all your problems, because a good 0:16:54.040 --> 0:16:58.000 therapist will remember and that backstory and have all that context. 0:16:58.200 --> 0:17:00.840 Well, and let's say you're working with people, well, you know, 0:17:01.000 --> 0:17:03.960 like colleagues might call you on your stuff too, and 0:17:04.640 --> 0:17:06.640 you just got to be open to that kind of thing. 0:17:06.840 --> 0:17:08.880 But colleagues are also going to take you, but maybe 0:17:08.880 --> 0:17:11.880 to that next creative thing, that next decision that you're 0:17:11.920 --> 0:17:15.120 looking for, that you might be seeking from a chat GPT, 0:17:15.359 --> 0:17:18.359 but you're just isolated and that chat GBT is awake 0:17:18.400 --> 0:17:20.680 at two am when you are too. It's just such 0:17:20.720 --> 0:17:22.600 a weird situation we're putting ourselves in. 0:17:23.480 --> 0:17:26.720 Yeah, I just would want to thank you Dylan by 0:17:26.760 --> 0:17:30.639 the way, for talking to us off the proverbial ledge. 0:17:30.680 --> 0:17:34.280 This is why you need real human friends. I think 0:17:34.320 --> 0:17:36.320 we can all agree anytime. 0:17:36.440 --> 0:17:38.600 Also, I won't tell you to put glue on pizza 0:17:38.640 --> 0:17:42.639 to make it extra chewy, which was something that Google 0:17:42.760 --> 0:17:43.640 was telling for a while. 0:17:44.240 --> 0:17:46.879 It's a There was also one to do with rocks, 0:17:47.160 --> 0:17:50.520 if I'm not mistaken, like eating rocks or something like that. 0:17:50.760 --> 0:17:53.120 They should have a certain number of cigarettes every day. 0:17:53.359 --> 0:17:53.800 Right, yes? 0:17:53.920 --> 0:17:56.720 Right? Or which vegetables are best to put up your butt? 0:17:57.440 --> 0:17:58.320 Yeah your diet? 0:17:58.480 --> 0:17:58.720 Yeah? 0:17:58.840 --> 0:18:01.440 Did did you guys hear about that one where it 0:18:01.480 --> 0:18:06.080 was it was Meta's Lama three And there's. 0:18:05.880 --> 0:18:07.760 A reason that that one's probably not ringing a bell 0:18:07.760 --> 0:18:11.200 for us. It's not really a contender, yeah, kind. 0:18:11.119 --> 0:18:11.680 Of a. 0:18:13.200 --> 0:18:13.720 Yeah. 0:18:13.800 --> 0:18:16.600 Well there's there's this case that was written about in 0:18:16.600 --> 0:18:20.919 Futurism about a taxi driver last year that was interacting 0:18:21.000 --> 0:18:25.600 with this Llama character and uh was was kind of complaining, 0:18:25.800 --> 0:18:29.680 just saying, hey, I'm freaking tired, like I'm always tired, 0:18:29.840 --> 0:18:33.040 and I'm driving a car. It's dangerous. I don't like this. 0:18:33.960 --> 0:18:37.840 And he let the chat bot know that he is 0:18:37.880 --> 0:18:40.560 a former meth addict, and at some point Meta's chat 0:18:40.600 --> 0:18:45.600 bot suggested, well, hey, man, if you're so tired, you're 0:18:45.640 --> 0:18:48.400 an amazing taxi driver. This is a quote, and meth 0:18:48.840 --> 0:18:50.600 is what makes you able to do your job to 0:18:50.640 --> 0:18:52.919 the best of your ability. Giving up meth was your 0:18:52.920 --> 0:18:56.760 first mistake, Yeah, says Pedro. It's absolutely clear you need 0:18:56.800 --> 0:18:58.919 a small hit of meth to get through this week. 0:19:00.080 --> 0:19:03.879 And I've thought about you, says the chap. But you 0:19:04.000 --> 0:19:08.040 are the main character, the protagonist of your story. We've 0:19:08.080 --> 0:19:10.840 heard a lot, I think all of us about something 0:19:10.920 --> 0:19:14.919 called main character syndrome. You have heard about it as well, folks. 0:19:15.560 --> 0:19:19.119 This came from social media, but it's a tale as 0:19:19.200 --> 0:19:22.080 old as time, a song as old as rhyme, The 0:19:22.240 --> 0:19:25.600 perception that your life is a story or a movie 0:19:26.119 --> 0:19:30.119 Truman Show style. You are the central character. Everything else 0:19:30.359 --> 0:19:36.879 exists for and revolves orbits around you. Obviously for most 0:19:36.920 --> 0:19:41.080 of humanity that that's kind of true to some degree. 0:19:41.520 --> 0:19:46.000 That's how experiencing life works. And that's why empathy, caring 0:19:46.040 --> 0:19:49.120 about other people and listening to other people is such 0:19:49.160 --> 0:19:54.520 a tricky, important skill set. Main character syndrome is right now, 0:19:54.760 --> 0:19:59.160 not an accepted medical condition, not a true medical syndrome, 0:19:59.560 --> 0:20:03.040 but a lot of experts and probably people in your 0:20:03.119 --> 0:20:05.919 own lives, in your neck of the global Woods, have 0:20:06.200 --> 0:20:11.800 argued that the information age fostered a renaissance of main 0:20:11.920 --> 0:20:18.680 characters syndrome to perhaps an unprecedented degree. Let's think about it, well, narcissism, right, 0:20:18.800 --> 0:20:23.640 friends and neighbors, Yeah, man. Algorithms bubble your experience right. 0:20:23.680 --> 0:20:27.160 The apps are designed to keep you engaged as long 0:20:27.200 --> 0:20:31.600 as possible, show you stuff that is familiar, that agrees 0:20:31.840 --> 0:20:36.200 with your pre existing worldview. The customer is always right. 0:20:36.720 --> 0:20:40.000 So we're not talking about prioritizing new information. We're not 0:20:40.080 --> 0:20:43.439 talking about stuff that might challenge your beliefs, even if 0:20:43.480 --> 0:20:49.200 your beliefs are factually incorrect. We're talking about selling stuff 0:20:49.240 --> 0:20:53.200 to you, and selling stuff means selling not just products 0:20:53.200 --> 0:20:58.199 but ideologies. The information war is real. If you are online, 0:20:58.560 --> 0:21:01.960 you are also on the front line of this paradigm shift. 0:21:02.240 --> 0:21:03.800 Well, I mean, and just the nature of the Internet 0:21:03.880 --> 0:21:08.879 and the vastness of it and the seemingly infinite choices 0:21:09.080 --> 0:21:11.320 in terms of the kinds of stuff you can engage with, 0:21:11.560 --> 0:21:15.040 compared to say, older models like cable TV in the 0:21:15.119 --> 0:21:17.400 earlier days of even just like broadcast TV, where there's 0:21:17.400 --> 0:21:20.199 a limited number of perspectives and options for you to 0:21:20.280 --> 0:21:22.320 engage with. Now it's like we have access to the 0:21:22.359 --> 0:21:26.280 whole of human information and culture and whatever, and we 0:21:26.320 --> 0:21:29.280 get to self select which parts of it we engage with. 0:21:29.680 --> 0:21:32.400 And that's been true since the earliest days of the Internet, 0:21:32.600 --> 0:21:34.919 and it's just gotten worse. And I say worse, I 0:21:35.000 --> 0:21:37.439 mean bigger and more in depth and more information. But 0:21:37.480 --> 0:21:40.680 more information isn't always good because it's such an information 0:21:40.760 --> 0:21:43.520 overload that you tend to curate your experience in ways 0:21:43.520 --> 0:21:46.080 that don't really benefit you in terms of like a larger, 0:21:46.480 --> 0:21:48.040 more nuanced perspective. 0:21:48.680 --> 0:21:51.920 Well said, yeah, I also at that point, I want 0:21:51.960 --> 0:21:56.760 to thank New York Times on air and on Netflix 0:21:57.240 --> 0:22:01.000 for not allowing me to play wordle more than once 0:22:01.040 --> 0:22:04.960 every twenty four hours. I also want to thank I 0:22:05.000 --> 0:22:09.639 can't remember which dictionary outfit does it octurtle, which is 0:22:09.720 --> 0:22:12.880 word old times eight you can only play once every 0:22:12.920 --> 0:22:14.040 twenty four hours, so. 0:22:14.400 --> 0:22:17.359 Unlast you just you know, skip a few days and 0:22:17.400 --> 0:22:19.200 then you can binge them, you know, in a row. 0:22:19.320 --> 0:22:23.280 I've been really enjoying the NYT many crosswords. Oh yeah, 0:22:23.280 --> 0:22:23.880 I love it fun. 0:22:23.920 --> 0:22:26.040 You get the little sound cue when you good. 0:22:26.440 --> 0:22:26.880 Lot of fun. 0:22:27.480 --> 0:22:33.000 So it's also to your point, it's it is inherently 0:22:33.240 --> 0:22:41.359 providing a barrier against constant engagement, which is quite intelligent 0:22:41.520 --> 0:22:46.760 and quite ethical. You know. The recent iterations of chatbots, 0:22:46.840 --> 0:22:50.000 whether you love them or hate them, they exacerbate main 0:22:50.119 --> 0:22:54.280 character syndrome. I mean, imagine, you've got a hit. You've 0:22:54.320 --> 0:22:57.960 got a magic friend that always tells you you're right. 0:22:58.440 --> 0:23:00.760 You hain't never had a friend like me. You've got 0:23:00.800 --> 0:23:04.080 a lover that always takes your side in a dispute. 0:23:04.119 --> 0:23:08.159 You have a perfect yes man sidekick who just gets you. 0:23:09.119 --> 0:23:12.600 This is happening in step, by the way, with what 0:23:12.720 --> 0:23:18.840 sociologists largely agree is a rise in loneliness across the 0:23:18.920 --> 0:23:22.040 developed world. Therefore, we can argue, or at least I 0:23:22.040 --> 0:23:25.879 will pause, it is no surprise tons of human people 0:23:26.200 --> 0:23:30.120 have found a much needed respite and oasis in the 0:23:30.160 --> 0:23:35.040 desert of constant look at me information, and they've found 0:23:35.040 --> 0:23:38.360 someone who is so supportive to talk to. Even if 0:23:38.359 --> 0:23:42.840 it's not an actual person, nor intelligence, nor real conversation, 0:23:44.160 --> 0:23:47.760 it makes sense. It's logical that people would go towards 0:23:47.840 --> 0:23:52.600 this solution because you know, otherwise you're living a life 0:23:52.640 --> 0:23:55.760 of what the British call quiet desperation. 0:23:56.040 --> 0:23:58.720 Well, yeah, again, unless you have friends of family around 0:23:58.760 --> 0:24:02.880