Episode 156: AI: The New Tool for Individual Empowerment?

Speaker A: Foreign. Welcome Back to reality 2.0. I am Kathryn Druckmann, and Doc Searles and I are back after a bit of a holiday hiatus, but we're back now and we are talking about the thing that we're always. No, actually that everyone is talking about, it seems, which is AI, but from a slightly different, um, perspective, I think, because Doc has questions. You know, that is Doc's role in life, is to have questions, maybe some answers. Yeah, but asking the right questions, right, is half of the battle. But, yeah. So we're talking about something. Well, I'll let Doc get into it. Oh, and before I keep going on and on, I also want to mention that we are also welcoming Ezekiel Lanza, our guest, uh, who is my co worker, but also is just checking. Generally cool. And knows all about AI and far more than I do. Yeah, that is. Yeah. If we had cards. Do you? Do you have cards? I don't.

Speaker B: I don't anymore. No, I should.

Speaker A: Does anybody? I don't.

Speaker B: I needed one yesterday and I. Yeah.

Speaker A: Oh, yeah, I was actually thinking about that recently anyway, before. I digress. But yeah, so we're gonna, we're gonna talk about AI and its relationship to personal empowerment and if that's. And if there is a relationship. So, Doc, tell us first of all, this, this conversation started with a blog post and I will let you know. Tell us all about that.

Speaker B: Yeah, so I was listening a while ago, like late last year to Sam Altman's keynote from OpenAI's developers day. Um, I guess it was maybe in October or November, and it was a one liner he had in there, which was, we believe that AI will be about individual empowerment on a scale we've never seen before. And I thought it was like, wait a minute, that's my line. And we've been saying that at Project VRM, which I started at Harvard's Berkman center in 2006, and it's still going strong as a conversation. It's, uh, basically a mailing list. And I thought, well, okay, so I'm sure what he means is that OpenAI is going to give you your individual empowerment, which to me is like a category error right there. No, you don't give. Uh, it's sort of. I feel like we're in the mainframe era still of AI, where it reminds me, because I'm old enough of PCs in like 1974. So people started talking about personal computers and, you know, the old timers said that those are an oxymoron. There's no such thing. As a personal computer. There never will be a personal computer. That's like having a, you know, a personal factory or a personal ocean liner. You don't want that. But then we got PCs and the whole world changed. So. So can we get that with AI? Well, I have an AI of my own. I want an AI of my own, but not, not to do a lot of this stuff that we. We're hearing about with AI now, which. All of which is pretty familiar, but I want a personal AI to, uh. For example, with health, I want it to suck in all the health information I have from multiple medical care providers who want to be able to take pieces of paper, run them through a scanner, have it do OCR on them, make sense of them, put them together with other things, look at my calendar, when was I, where, What doctor did this, what didn't do that, not necessarily give me diagnostics. I don't want that. I just want to make sense of it all. Similarly, um, with finances, um, you know, what are my holdings, what are my obligations, my recurring payments, all of these different subscriptions I've got. When do they start? When do they stop? Um, property. What have I got? I mean, I'm sitting here, you can't see it because this is not a visual podcast, but I've got bookshelves behind me. I want to. I want to aim a camera or my phone at that bookshelf and have an AI. Know, what are those books? I recognize the spines. Or I can go out to some database that looks at spines of books and says, oh, those are the complete works of John McPhee. There, there's a McLuhan, there's a Lakoff, there's some other things. Um, look back through my correspondence. I have Go, which in email goes back to 1995. That's electronic. Shouldn't be hard to pull stuff out of there. Um, Apple's Mail app, which is the one I use right now, is horrible for search and for what it is, what was when? Um, you know, but, you know, my, you know, make sense of my contacts, you know, who's, you know, who's changed, who hasn't, where are they? How do they line up with the calendar? Tell me where I was on a. On a given day. Uh, who I did it with. Um, all my travels. I mean, I've, I, I have a million and a half miles with United alone. Where was I when? This is interesting to me, especially if I'm working it together with other things, you know, like travel and health. You know, wait a minute. Did maybe the pulmonary embolism I got in 2008 was right after a flight. Was that true? I don't know. Did I fly the day before? Maybe I did. Oh, that's interesting. Right. So there's so many ways that we can look at this from a personal side that I don't think anybody who's currently at Microsoft or at OpenAI or at Apple, which we learned recently, is actually doing. Actually doing something out of the silence in. In AI, but who's not doing something in AI. I suppose so. I mean, they have Siri. Siri does AI. I suppose. Anyway, so that's questions I put out there. Um, not that I've had much response to it yet, but I really. But question I have about all this is, is it possible? I mean, is this is AI essentially only a mainframe kind of level thing? It can only. You need. You need giant clouds to do this. You're going to have to hire it, you're going to have to pay somebody else to do it. It's going to be a black box in the sky that I have to use. Um, and I can feed all my information into that and hope I get sensible stuff back out, but it's not something I can have on premise. So that's my big question and Easy's nodding. So what do you think, ez?

Speaker C: That's a great question. Of course. Um, I think that most of the things that you mentioned now you can do something similar. Um, I mean, we have some, there are some advances and of course advancements and one of them. We can talk about LangChain later. Um, but for instance, what you're saying, I have all my databases of travels. I mean all the things that I, of course, thinking on an engineer or engineering part, you need to have the data because you need to upload to the model and so on. But for instance, with rag, which is retrieval Augmented Generation, you can say, these are my documents and you can make questions about those documents. Ah. Or try to find. You can feed like I have my entire story of chats, whatever. And you create these vector databases or a database with this information is you have your model and you can talk to the model and say, hey, could you tell me what happened in that day? Did I say that to something? I mean, was it aggressive or not? I mean, thanks to the m, to the LLMs, I mean, thanks to the understanding that you will have with the LLMs, you can adapt this understanding to your particular data, for instance. And it doesn't mean that you have to fine tune Your model or retrain. Your model or retrain the LLM, which, which will make that you need a lot of hardware, money and everything. Right? But with rag, for instance, you can refer to that and they can think like they can go to rug.

Speaker B: What is that? Uh, spell it.

Speaker C: Retrieval. Retrieval. Augmented generate generation.

Speaker B: Generate. Okay. Okay, yes. So it's augmented generation. Okay.

Speaker C: You have a database that is vectorized. I mean it doesn't. I will not go in that detail, but it's vectorized, which is with your information, with your context. Right. And the model has a way to ask to that database. Right. But with the intelligence that the model can have, for instance, because it's not like a, uh, search like they will be searching for something. They are understanding your documents and the answer that the model gives you, it's uh, an answer related with those documents. Right. With the knowledge and with understanding and with everything. Right. Um, and the other thing that is important is also launching, right. How you can make these models to talk to the Internet. Right? You like to say to a chatbot, let's say a chatbot, but it could be any other application. You can say, can you send an email to that person and say that I love her? The model will do that. The model will interact with your Gmail, with your Outlook or whatever, and they will send that email for you. Um, there are a lot of different agents that thanks to LangChain, there is allowing to communicate the. Allowing the LLMs to communicate with the external world. Could be Internet, could be any other application, could be Salesforce, could be whatever. But there isn't the intelligence that you have in the LLMs models to go to your models, to your information, for instance.

Speaker A: So I have, so I have a question. So as Doc is talking about all of this data about us, I, of course my brain goes to like data brokers and all of that information that exists about us, that is for somebody else's benefit and not ours, right? There is some entity out there that can pull, uh, us up. Pull me up, as in a group of women of a certain age and market to me, send me marketing materials. But when we think about. So you talk about LangChain and connecting out to the world, I'm kind of thinking back to what Doc was originally saying and keeping it all local, right? Because I don't, you know, I want to be able to access all of this information. I want, I want it for my own benefit, but no one else's. I don't want anybody else to know where I was on every single day of My life or, or all of my medical procedures or all of this sort of stuff. Um, so, and, and then, and that raises the question for me of computing power, because as we know, training takes a lot of power and I suspect even locally training on your own personal stuff would take some power. Um, but maybe, uh, maybe not as much, I think. I don't know. But I wonder how this all kind of fits together. But, uh, how feasible is it to have this massively powerful functionality for our own benefit, but keeping it completely local?

Speaker C: Well, that's a great challenge. I mean you don't need a lot of computation, of course you need, if you would like to do it right now, uh, you need to download the model to your computer, uh, which is, I mean it needs a lot of space and memory in compute. Uh, but it's something that you can have it in your computer with 32 gigabits. I mean it's something that you can do it. Um, but I'm not sure. And this is how the, how it will be evolving in the future is because probably we will have this kind of solutions, uh, in the cloud. Or someone will be offering that to you. Uh, yeah, or there will be a regulation, for instance, that. Okay, when you are using that model or that RAG solution. Ah, the data, uh, or the data that you're seeing. I would like to ask about those documents. Those documents, they cannot go to the cloud. They have to live in your computer, for instance. They cannot be there. Um, something like that. I imagine that could start to happen. Um, but so far, so far or now, what is going on is most people is using that on servers or in the cloud, but they are not offering a solution yet. And there are a lot of conversations going on on rag, different ways to do the rag. It's evolving all the, all the time. Um, probably in the next two years or when you're, that will be very optimized that you can run it on your computer and that could be an application that you download and you do it right and you use it. Um, but if we talk now compared with one year ago now, it's not a focus on the model itself or the large language model, it's focusing on how you take advantage of those large language models and they are still huge to be used. Um, so you need COMPUTE or you need to use an API, but if you're using an external API, you are paying for that. So someone is getting your information, someone is getting your prompt, someone will be getting that information.

Speaker B: So you can't. Uh, I'm Thinking when Katherine's asking about, um, local, I mean, local is actually a logical sense here. I mean, some of this could be in a, in a server somewhere, a cloud somewhere. But the point you just made about someone is going to know your prompt, someone is going to know this stuff. So just security wise, there's a huge amount of exposure. Right? I mean, and it's not been clear to me how much security has to depend on the physical location of compute and storage in your control or if you're a company in the company's control. I mean, obviously a zillion companies run on AWS or Azure or some other thing there and they're exposed in that sense. It seems to me that we're a long way from figuring out what's safe and what's not and how to make it safe. And AI just makes it a lot more complicated because you need these big things running somewhere else to make it work.

Speaker A: And also when you talk about, uh, you know, local, it kind of makes me think. And I don't know how much either of y' all have, have looked into these devices, and I frankly haven't that much either. But, um, these kind of startups are popping up all over with these tiny little personal devices, like the next generation of what a personal device is or that's, um, that's at least the story. Uh, and one of them is kind of like a little pin that you wear, and the other one, I'm not really sure. The Rabbit thing. You sent me that link talk.

Speaker B: Yeah, yeah, Rabbit's another one with a lowercase are. Um, yeah, the pin. I think almost everybody's crapped on that.

Speaker A: Um, yeah, yes, that's true. And I just. Yeah, I mean, it just seems like when, when I see these things, these, these prototypes come out, they always seem to, to have a very different idea of what people might want than, than what I would want or what you say you want. And I find that kind of interesting. Like maybe, maybe we're, we're the outliers. Maybe. I don't know, it's kind of interesting. Question.

Speaker C: Yes. I mean, I think that that's what is being enabled, uh, thanks to the technology and everything.

Speaker B: Right.

Speaker C: Um, this Rabbit thing, I mean, it's. I didn't go in those details, but at least what I heard is that it's like a rethinking on how you use your mobile phone. Right. Instead of using applications now. You don't have to interact in that way. Uh, you don't have to download the application so you can Use something that can do that for you.

Speaker A: Yeah.

Speaker C: Um, and same thing with the chat PT and the conversational way. Now we, we don't think on Googling the things, probably we think on this conversational way. Okay. I would like to ask, as I can ask to another person. Uh, so that's a great opportunity for, for a lot of startups. I mean, they are, yeah, they're creating products, they are creating a lot of things. And I think that that's, that's, that's just the beginning of this revolution, probably.

Speaker A: Yeah, yeah, I meant, yeah. Imagine your phone is no longer does nothing except Siri, I guess. I mean, in that world, Siri is massively more powerful.

Speaker C: Yeah, but it makes sense. I mean, sometimes it does make sense. Yeah.

Speaker A: I mean, that's how we interact with the real world.

Speaker B: Yeah, well, with each other.

Speaker A: You know, I can't punch a button on your face. Not yet anyway.

Speaker B: Yeah. But there was a cartoon I saw or maybe just imagined of, uh, somebody saying, don't talk to me. Like you're prompting me. I'm not an AI. Right. That's funny because so much dialogue we have with computers now are prompting, you know, to get the result that we're looking for. So it's almost like you deal with another person and you start iterating your prompts to try and get the response out of the other person that you want. Like they're an AI.

Speaker C: Uh, but it's the same thing. I mean, when we are interacting with other people, I mean, we need to know how to make the questions.

Speaker B: Yeah, exactly. Exactly. We've been prompting forever.

Speaker C: Yeah.

Speaker B: You know, we're all just pattern recognition machines that are, that just get a little more complicated and we have to tweak ourselves.

Speaker C: We all knowledge, probably.

Speaker B: Yeah, yeah.

Speaker A: We have an agenda, a desired outcome.

Speaker B: I mean, but I go back always to Marshall McGloomin who said, we shape our tools, then our tools shape us. Right. And that's how it's always worked, you know, And I mean, uh, with everything, you know, we, we don't go barefoot anymore. We have shoes and our feet change in shape to deal with shoes. And we can't, you know, we can't go outside without shoes. Most of us ancestors did.

Speaker A: Not today. Yeah, not today even. Even here, where it's usually very warm.

Speaker B: But our shoes are our feet. Well, so. But it's interesting, you know, if we think about it in a personal way, all the familiar tools to us, whether it's a hammer or a phone or, um, or these, you know, these things we stick in our ears and you guys have them and I've got 12 of them here. They're extensions of us as people. Right? And you know, and you ride a bicycle and that bicycle is part of you. You're, you're, you have wheels now and you drive a car. And those are my fenders and my steering wheel and my engine. And. But with, with AI, you're dealing with this big thing somewhere else that's has a lot of human capacities and resembles other humans in a very recognizable way and is designed to do that, but is not necessarily an extension of another person. It's like an extension of everything that's ever been said on the Internet. And that's, you know, what makes it both interesting and weird. But how does it change us?

Speaker A: You know, it's interesting you mentioned that. This reminds me and hey, this is a great opportunity to plug that other podcast. So I did it. You know, hey, I do this other podcast. So the Open it intel podcast, uh, is the podcast I do when I'm working. Um, so I interviewed a woman named Liz Rice. She is involved with a company called Isovalent who was incidentally acquired recently by Cisco. Anyways, so in our conversation though, she, she raised a point about using, in her example, it was, she was involved in, I can't remember the project, but I think some sort of a certification or testing questions and using AI, some sort of AI tool to generate test questions. And of course these things were proofread by uh, humans and, and somebody raised a concern. Well, you know, here, these things, they, they make mistakes, right? And, and here's a perfectly. It looks like a really good, uh, question with a list of multiple choice answers. And it's totally believable. But if you don't fact check it, you know, we would, you know, this, this, it's, it's producing a hallucination, I guess, as you would call that. But then, and then we brought up though, well, we said, well, but humans do that too. We all make mistakes. We make, uh, mistakes that are, that are at least plausible. And our conversation kind of concluded with perhaps the, the more flawed these systems become, the more human they become because we are all deeply flawed and we get things wrong. And, and um, and that's an interesting, I think, way to look at it.

Speaker B: And, and we make stuff up all the time, you know, I mean, we, we make a best guess at something. We m. Misremember. Yeah, that's what these things are doing. They're, they're making best guesses at, at a at a plausible or sensible string. Um, of words. Uh, and you know, these at least if you're talking about language or with um. Uh. Or with art. I mean I do an awful lot. Not an awful lot. It's probably an overstatement but I mean I do a lot of playing around with, with Dolly and, and Midjourney, uh, and the others. And, and so for example in that blog post that I mentioned at first I, I took that phrase individual empowerment and agency on a scale we before. And, and this is in the image that came out of. It was for. Was from. This is from Microsoft's horrible name. Whatever they. They use now. I mean it's, it's uh.

Speaker A: What.

Speaker B: I mean what, what is it called?

Speaker A: Bing Image Generator. I don't know.

Speaker B: Oh yeah, it's a big image generator but it, it has a horrible name. I mean it, it's. It's. I mean it, it's Microsoft. It originally was Bing Image Create and it's now Bing Create Microsoft something by Designer. I mean it's just a.

Speaker A: No, no.

Speaker B: And I can't remember it. I could ask an AI and it would make something up. I mean that's uh. I'm trying to find the. I do all of that in uh. Let's see. Oh here. Yeah, it's Microsoft being image creator from Designer.

Speaker C: Right?

Speaker B: Yeah, Image creator from Designer, whatever the hell Designer is. And I apparently copilot as originally was just a helper for programmers and now apparently it's in Word and a lot of other things. But not everybody's seeing it because they're slowly rolling it out to the office365 um, customers and. But, but I asked it to, you know, to create. To take that phrase and individual empowerment and agency in a skill I've never seen before. And it gave me something that looked like the COVID of an Ayn Rand book. And you know of this, the Superman Always Men. Right. You know, and ready to forge the future. And I mean it looks creepy. And then uh, I asked it to do just empowered individuals and I got something that's the most woke thing you've ever seen, you know. And so what does it do? I mean it just takes a bunch sensibilities. It just looks into whatever descriptions of art of different kinds of art those are. And then make something up, you know. And you know the second one has you know, flags in it and uh. Female symbol. The female symbol and uh, graffiti kind

Speaker A: of symbols of all these different movements.

Speaker B: Yeah, I know it's a kind of interesting. Yeah, it's, it and of course, being people, we think, oh, it's thinking, but it's really not thinking. It's just make it. It's a program that's making stuff up.

Speaker C: Yeah. And it has the same thing as you can have in your language models with ChatGPT. When you ask something to something bad. I mean, the model knows that how to do it, for instance. And same thing for the Dali or stable diffusion. Uh, you can probably draw a naked, uh, person and you can ask for that, but of course they will not do it because they are not allowed to do it. But the same thing as you said, what is the most beautiful thing in the world? And the answer that you will get will be probably based on the what the entire world think that is the most beautiful thing in the world. And um.

Speaker A: Okay, now I've got to, now I've got to try that. I'm very curious. Just going to enter the most beautiful thing in the world.

Speaker B: I, I'm curious. I don't know if either of you.

Speaker A: I didn't know you couldn't ask it to draw, uh, naked people. I've never tried. That's interesting.

Speaker C: I'm sure it won't, it's protected.

Speaker B: It won't, it won't do anything with Trump or Biden. It won't do a lot of things that are political. It's just not going to do. But I've seen some AI stuff where people have managed to get them to do something like that. Um, there's, uh, a, uh, pretty popular um, cycle, deviantart, all one word. And it has a thing called dream up. And one can use that to say, okay, give me a beautiful woman or a handsome man. And from what I can tell looking at it, what it does is it starts by giving you some cliched version of that. And then, um, what's that you're showing?

Speaker A: It's the most beautiful thing in the world. I can't focus.

Speaker B: It's too blurry.

Speaker A: Yeah, it's like a surreal landscape.

Speaker B: A surreal landscape. Yeah. So what DeviantArt does, I think, and I don't know, and it's interesting. I haven't played with it enough. But I think it learns from the people who are using it. And so it takes the ideal. It's a, it starts with an original ideal and then alters that ideal based on what people are using that art for and how they prompt it. And so, you know, so the idealized woman is, has big boobs and big lips and wide eyes and blood blonde hair or something. And the Idealized man looks like Jon Hamm, you know, with muscles.

Speaker A: Sounds about right.

Speaker B: Yeah. And I mean that's pretty much what, what happens, right? And, and we have brand new cliches out of that, right? It's kind of like on Instagram there was, uh, uh, you've seen the, you know, Instagram models, they all start looking the same. They all have the, the big eye. All the girls have big eyebrows and wide set eyes.

Speaker A: Yeah, well that's a whole other conversation. Yeah, you've seen these things that are the AI, they're completely AI generated. Uh, they're videos, they look completely human and they are, uh, massively profitable. Anyway.

Speaker B: Yeah, I digress.

Speaker A: Go ahead.

Speaker B: No, but that's, you know, so then it becomes an art form, right? Then it's. And why not? Right? And it's. The humans still lead the art in some weird way, right? That's going to come out.

Speaker A: There's a human behind it.

Speaker B: Yeah, yeah, yeah.

Speaker C: But with, with art, I mean I have a, um, um, different thing feeling about the art and also the, I don't know, the fashion and all this stuff that it's not basically, basically based on data or what most people think. Um, like with fashion, you, you use something just because someone has an idea that you will like and you like it. But not because of course after that most people started using that, but at the beginning and it didn't start from, from a previous information probably there are some examples, multiple examples with fashion and also with art that they try to create things based on data and they are completely failure. And same thing with art. Um, I think there's a new way of art, but it's probably that they look at the same thing. They look something pretty similar. Uh, you and I have a new Dall E or a new Picasso or whatever where they are create something completely different of what it's done by now.

Speaker B: Um,

Speaker C: this is probably not so probable uh, with AI because AI or the models or these things, they are looking backwards and they are saying, okay, what are what most people liked? And I will create something based on. That could be based on the feedback based on whatever. Um, I uh, know this is basically what I, or mainly my perspective or the art and fashion and this stuff.

Speaker B: So where are we at?

Speaker A: Incidentally where I am is, uh, I was ah, off messing uh, around with uh, the Bing image generator again. I asked it the most beautiful thing in the world and it came up with a surreal tropical landscape in the form of kind of digital art. Actually not all that interesting. Surprisingly not interesting. And the most beautiful person is quite feminine. So that's interesting. I guess beautiful is kind of a feminine quality, but they are a bit androgynous, which is interesting too. Anyway, it is kind of interesting to get, get into the non existent head of a. I mean it's generator.

Speaker B: It's all. It's. It's like we're throwing up Rorschach tests in front of these things over and over and over again. And what we're doing is. It's this massive mirror of what people have said and shown over the years. Uh, it's fun to be at the early stage of this, I think it is.

Speaker A: It's fun to watch things pop up all over, right. If you use an Edge browser, for example, the integration of the ChatGPT and all that image generation and Dall E and everything has been um, interesting and more and there's more and more of it and you just kind of see these little things pop up. Right. Uh, I was looking at TripAdvisor and it has a new thing that I hadn't noticed before where you can, it's an AI generated itinerary. You can say, this is what I'm interested in. This is where I'm going and it will make. And this is how long I'll be there and when and it will make you an itinerary. And it's actually not bad. I was kind of surprised. Um, but yeah, it's kind of, it's fun as a, you know, as a consumer and as a. Just as a human to see, to see where people, how people are implementing it. But it's also, um, I don't know, it's a nice time to do it as Doc has done and kind of put out a wish list into the world,

Speaker B: I have to say. I mean, I thought it was a hell of a wish list and I've had almost exactly zero response to it actually so far. Now maybe this.

Speaker A: But that's also interesting you mentioned that too. And we've talked about this, right, that kind of our individual places on the Internet have become less and less relevant over time. People don't consume that way. In the early days, people, we had things like blog rolls, right? And you had, you went around individual websites and, and that's no longer a thing. That much. You're, you're. Or maybe you do, but it's only if it goes quote unquote viral on, you know, on some social media platform, you consume it, you consume out of portals, right? Like, you know, if you're in a certain Group, it's Facebook or you know, maybe YouTube or Tik Tok or wherever it is. But now as we're talking about these kind of next generation AI enabled personal devices, we're even further removed from individual websites and blogs and stuff like that. If your, if your interaction with a device now is just to ask it a question. Unless somebody specifically says, hey, what does Doc Searles think about, you know, the future of personal AI? Um, I don't know. Interesting.

Speaker B: I mean, it, it, it's interesting. So I, you mentioned Edge. I thought, do I have edge? So I thought I'd, I hit. Okay, I'm using a Mac here. So I hit Spotlight and Edge and sure enough the app came up and it's got about 10 open tabs in it. So I guess I used it at some point and it says, get the new Bing. You know bing. Bing is ChatGPT4 for you, the Bing app. And I'm thinking, oh great, okay. It's an app on the computer. No, it's only on a phone. And uh, excuse me, I'm from New Jersey, so fuck that. I want it on the computer. Really. You know, I mean, I don't have to go to my phone when I want to do something useful. I'm sitting here right now, I'm using the damn computer.

Speaker A: But I guess it is pretty well integrated. Um, if you're using the Edge browser,

Speaker B: there's a. I'm losing the Edge browser now. Uh, what am I looking for on this?

Speaker A: Go to the top right. You should have this funny looking little icon and if you hover over it, it'll say copilot. Looks like a badge every day. Um, does it. I don't know, it looks like.

Speaker B: Is that the Copac? Microsoft rewards is a reward thing of that too. I, I don't want that. Really. I mean, quit bothering me.

Speaker A: This is gonna be a fun one to edit.

Speaker B: You know, just go ahead. We're gonna get banned from. Oh, for Apple. Oh, for Apple. You can't say that. Oh, Jesus. It's fine, really.

Speaker A: I have, I have beeps I can insert. Yeah, really?

Speaker B: Oh my God.

Speaker A: I mean, unless you mark it as explicit, which I'm.

Speaker B: Oh my.

Speaker A: Probably a bad idea.

Speaker B: I mean, I, I can't believe how pure, how stupidly, robotically puritanical that is.

Speaker A: I, I mean, I don't disagree, but

Speaker B: I mean, where I come from, that's our, that's our word for um. So. I mean, really.

Speaker C: Yeah.

Speaker B: Uh, I'm sorry.

Speaker A: It's totally fine. I'm not cutting any of this out because it's funny.

Speaker B: So there's a thing in the upper right corner somewhere.

Speaker A: I just see collection. Yes, there is. It looks kind of like an O. A squiggly o. An O through some vaselined glass.

Speaker B: Don't have it. Don't have it. I'm not privileged.

Speaker A: Upgrade your browser, I guess.

Speaker B: Really? Maybe that's probably it. I just have to upgrade it to see if it'll do that. Well, I don't want to, but it's interesting.

Speaker A: It's interesting. We are on the cusp of a lot of really interesting stuff. And it's like every day you see more and more integrations and more and more things taking advantage of AI capabilities. And that's kind of fun.

Speaker B: So Microsoft has overtaken Apple in the worth of its company. Like it's now worth $4 trillion or something, which is.

Speaker A: Yeah, uh, that's interesting. And I think a lot of it is the AI buzz. Right? There's a.

Speaker B: It's totally AI buzz and it might be reality to some degree. I mean, Microsoft may be really ahead on that.

Speaker A: Apple's been very quiet.

Speaker B: They're being quiet and either they got kneecapped and not telling anybody or they actually have something. I think it's the former. I think they're way behind until.

Speaker A: I don't know. I think Apple's reputation is typically to not say anything, not release anything until they get it, until it's perfect. Right.

Speaker B: Um, so you're gonna get, uh, spend $3,000 for their headgear to.

Speaker A: Oh, I'm so tempted. Just because I want a new toy.

Speaker B: I've actually heard that it's not bad to get the headset. Yeah.

Speaker A: Ah, it's expensive though.

Speaker B: Yeah, it's extremely expensive. It's crazy expensive.

Speaker A: It feels irresponsible.

Speaker B: And it's the first one. I mean.

Speaker A: Yeah, that's the other thing.

Speaker B: It's the first one or the first Apple watch. You know, those are, uh.

Speaker A: Agreed.

Speaker B: Those are bricks at this point. So that thing's a, A brick in the works.

Speaker A: My second gen iPhone, actually. I found it and it powers on and it works. It's bizarre.

Speaker B: Is this the one that was the iPhone 3G? That was.

Speaker A: It was, yes, the 3G.

Speaker B: I had that. I had that one.

Speaker A: Yeah, I have that too.

Speaker B: Joyce got the very first one, which I think really is bricked at this point. And then.

Speaker A: Oh yeah, I'm sure.

Speaker B: And then I got the 3G. Well, I, you know, um, we're way off the original topic, but why not?

Speaker A: Yes, we are.

Speaker B: I. I have. I've been cutting Sirius, as in Sirius xm, since the very beginning because I wanted to get Howard Stern and I wanted to be able to use the satellite thing. And, um, they just changed their app and subtracted enormous value from it. And I'm thinking of killing it because it is.

Speaker A: It's expensive.

Speaker B: It's expensive. And they've got his. There's the Cory Doctorow word that I guess has a noun in it we can't use. Um.

Speaker A: Oh, and poopification.

Speaker B: And poopification. Yeah, poop is okay.

Speaker A: I think poop is okay. Everybody poops. It's a book. So.

Speaker B: So maybe it could be. And it'd be unmeritification or something and

Speaker A: Meredithic, you know,

Speaker B: something like that. Anyway, yeah, I'm, um.

Speaker A: Yeah.

Speaker B: Because what they got rid of is the ability to go forward and back by 30 seconds at a time. And why is it. It's to make you see ads. To make.

Speaker A: Oh, that's ironic.

Speaker B: Forcing to listen to the ads. And it's what. I mean, this is actually what's happening with, with tv. Like the, you know, when you had a dvr, right? You could. You could record the game and jump over the ads. Now you have to subscribe to the game and only get it on some, you know, streaming channel. And you can't spool it. You can't record it. You have to just watch the whole thing. That's the new thing, drag you through the ads. So interesting that's. We should ask Corey for the word for that.

Speaker A: Maybe, maybe, uh, to bring it. Bring us back to our original, uh, topic. Um, uh, well, you know, on the subject of optimism about AI, maybe AI will solve all of these problems, all of our gripes with our devices and apps and, and functionality and an invasion of ads and whatnot. Maybe, maybe it will solve it. So somehow I don't necessarily believe that, but curious to see what's coming. We'll see.

Speaker B: I'm so optimistic about what it can do, and I'm so pessimistic about how it's going to be used for the most part. I mean, it's sort of. It's both. I mean, I think it's incredibly useful to me already in a lot of ways, especially with the art. I have fun with the art. Uh, but I, uh, know it's going to be used to send even more awful advertising at me and hold me still, try to understand me in ways that they could guess at crap that I don't want to be guessed at. And that's going to go on because there's so much money in advertising. This is massive. So, you know, and Microsoft's in that game and Adobe's, uh, in that game and Google's in that game and Apple's uh, only in the game so far as they can game their own advertising of apps on their store. But it's, they're in the game to some degree.

Speaker A: Yeah, well. Well, yeah, I think, you know, on

Speaker B: that, on that, uh, on that note.

Speaker A: Optimism. No optimism.

Speaker B: But I'm up to, I'm not a, I'm not a pessimist about, in general about AI. I think AI is great and I think it's, I think it's fun and I think fun is good.

Speaker A: So I think there's a lot of overblown pessimism that is not necessarily relevant. And I like, I choose to remain optimistic and excited about all the cool stuff that we'll be able to do. Um, well, cool. Well, thank you. Thank you both for, for hanging out and thank you very much. Easy for jumping on and.

Speaker B: Yeah.

Speaker A: And teaching us how these things work and making sure we don't get it wrong.

Speaker C: Thank you both for having me. Thank you.

Speaker A: Mhm. Always fun.