Speaker 0: better late than never. get ready because the ai landscape just got another major shakeup. we're talking about meta. yes. the company behind facebook and instagram making a bold entrance into the generative ai arena. at their first ever ai conference, lamicon twenty twenty four, mark zuckerberg unveiled something big, a standalone meta ai application powered by their advanced lama four model. this move isn't just significant. it directly positions meta ai as a chat gtp competitor challenging openai's dominance in the conversational ai space. it's a strategic launch that integrates ai capabilities across meta's vast ecosystem of platforms leveraging that massive social media infrastructure. they're putting serious weight behind this too signaled by their 60,000,000,000 investment in ai infrastructure. and mark zuckerberg saying meta is an ai company and the restructuring around ai. like, they're hyper personalization that learns across whatsapp, instagram, and facebook. that's right. it learns across your, whatsapp conversations, your instagram likes and posts, and facebook. the unique social discovery feed and their focus on voice first interaction, especially with, raybond smart glasses. we'll analyze what makes meta a significant competitor compared to other ai assistants and what their strategy means for the future. what do you think about this, john? Speaker 1: yeah. before we dive in, didn't meta, like, just a couple years ago, say they were a, vr, company or something, didn't they? Speaker 0: yeah. well, let's let's let's talk about where medic comes from. right? weren't they gonna go into Speaker 1: the second leg for whatever, like, Speaker 0: the metisphere, whatever we're calling it? Speaker 1: yeah. yeah. yeah. yeah. they're happy with that. Speaker 0: but let's face it. yeah. you know, mark zuckerberg, seems to be a smart guy. yeah. he's run he run this company. i'm not gonna put him down, but he's made a lot of, strategic plays that didn't work out for him. yeah. remember, they went all in on html five, and they spent years on either. and then that didn't work, and they switched that. and then, what were you just saying? what'd they go all in on? they've gone all in on a bunch of things that just didn't quite pay off for them. and so we'll see, you know, ai is the future in many things. and i think what people don't understand is matt has been in the ai game a long time, and they're a leader in some areas. the problem is they never really had a chat client. so there's a lot of other people built on their llama platform. a lot of other, databases, other ai built on llama, which is an open source model that they can grab, train, and use. like, numitron from n, nvidia, and many of the other models are built on llama. so a lot of ai experts using the back end to do stuff, but it's not in the mindset. a friend of mine, alex, you know, alex, she asked me recently. so what's you know, i haven't heard anything from meta in this space in a while. are they just dead? and i was saying, no. they're on the back end doing all the stuff. and now i think they just don't have the mindset of the consumer, and we're going to see if that changes with this. Speaker 1: well and just to digress a little bit too, that's the same problem that google's gemini has. gemini beats a lot of these other llms in a lot of things, like the 2.5 pro. but the problem is they're not getting the users to use it. that's their biggest problem. so, but, yeah, if we let's dive into this a little bit. so, this is cool. the one thing i like is that you can pair the app with meta's ray bans, and it turns the ai into a wearable experience, which is cool. that was one thing i felt was missing from the well, there's a lot more, but, they don't have a display on the on the glasses or anything like that. but, i think that's a game changer for the ray bans, and i know they're coming out with another one. and speaking of ray bans, mark zuckerberg has made another announcement that, his smart glasses or smart glasses are gonna replace cell phones. so we'll have to see where that goes with that. but, i like this. this is interesting. llama four, i haven't really played around with, so i don't know how it is. but it is a cost efficient llm. it's fine tuned to incorporate user data, like you said, from facebook and instagram profiles. so, there's a couple cool integrations that they did with this, but we'll have to see if people are actually gonna use it. that's the question. right? Speaker 0: yeah. speaking of the raybon ray ban app, they just renamed the, metavision to meta ai app. so if you go download it, it's the same app. they just updated it. okay. but i think this is interesting, because there's a lot to be done here. i was in lam, meta's ai studio earlier today looking at some stuff. it looks interesting. i'm a little concerned about what all they have at that ai potentially has access to Speaker 1: yeah. Speaker 0: and what that means. like, you know, your whatsapp chats, your instagram stuff, all your facebook interactions, like, do you really want that? Speaker 1: we already know that their data policy is horrible, so i'm not surprised. right? Speaker 0: well, it's one thing to have the data, and it's another thing to have ai actively going through all your data for you to pretend you know, for whatever reason. let's let's be honest. like, facebook just trying to sell you stuff, sell you ads, know more about you so they can sell you ads, sell other people ads, and become better at that. so yeah. i just wanna be clear on you know, they're going to now say, hey. we're gonna give you an ai tool that goes into your instagram, goes into your whatsapp, goes into all your conversations on facebook, and helps you. Speaker 1: yeah. Speaker 0: and learns from you and does all this stuff great. so Speaker 1: you know what, though? there was one thing that they were talking about that i think is cool is the innovate the voice innovation. so they've got a full duplex voice mode, which seems interesting. so that means, like, you can interact with it in real time, and it responds, and it inserts things and and different things like that. so, i like that. it mimics human speech. so i think that's cool. i'll i'll be interested to see how that works, though. Speaker 0: yeah. so let's, let's i mean, i was poo pooing it all the time. so let's talk about some of the advancements. you know? meta's approach to personalization extends beyond simple preferences, preference tracking to create deeply contextual understanding of users' needs across platform. so cross platform learning, the system builds a unified understanding of user preference from interaction from whatsapp, instagram, and facebook. content adaptation responds and recommends and adjusts based on communication style, interest, interaction history. progressive personalization, the more user engages with netai across the context, the more tailored the experience becomes. Speaker 1: yeah. Speaker 0: discovery feed. network visibility. users can view creative ways their connections utilize ai assistance. so your other ais can now see what's going on with your friends and then report back to you and do stuff. permission control, granular sharing settings, all users to control precisely what ai interactions are visible to others and inspirational discovery. the feed serves a discovery mechanism for new ai capabilities and use cases. so one of the things that i thought was interesting when i was analyzing this is it allows you to see how your friends are using the ai, the ai prompts they're using, and that kind of stuff. so there is a sharing mechanism around, how everyone is using ai and improving. and so that is quite interesting Speaker 1: yeah. Speaker 0: quite interesting from an overall perspective. like, i like the concept of being able to do that. you have some granular control over that because, you know, say you and me and our friend alex and a couple other people wanna share what we're doing with ai because we're doing cool stuff with it all the time, and this would be an interesting potential way to do that. i think there is a potential a, for lack of a better word, an social network where it's all about, sharing ai tools and prompts and learning what other people are doing and whatnot, but we'll leave that a set. Speaker 1: yeah. Speaker 0: voice first, you talked about that. right? ambient computing, users can initiate ai assistance through voice without, removing phones or pockets. contextual awareness is what understands environmental factors and multimodal transitions. conversations start via voice, can seamlessly transition to screens when needed, for visual information. but there are privacy concerns. Speaker 1: for sure. Speaker 0: so we'll have to see. there's some here i like. there's some things i don't like, but that's the same with everything. so they're putting 60,000,000,000 investment commitment to ai infrastructure to get this going. we'll have to see where this goes. i'm curious to see. are you gonna jump in and start using this? are you a facebook person, john? Speaker 1: no. i use facebook just to post memes. that's all i do. and and whatsapp, i try to avoid as much as i can, but there are some groups i'm a part of. so, yeah, i i don't share much personal things. i just use it for memes and stuff. so, i've never used any of the chat functions for any of these ais. i don't know if you have, but, well, we'll have to see. maybe i'll mess around with it. Speaker 0: i'd love to hear what your feedback out there is, or we would. and let us know. do you plan on jumping in and trying this facebook ai, or are you gonna disable this so it's not tracking you? what are your reactions to metalamicon 2024 and the release of this chat tool? we'd love to hear what you think. Speaker 1: yeah.