37 | Supercharge Business Productivity with These Practical and Incredible AI Use Cases, with Chris Carr, President and CEO of Farotech

27:21The first segment tells you that the technology is there today to take us beyond what we can do with people today to the point that we can create content, communicate in real time in any language with every kind of personality you can imagine, basically effortlessly once we connect all these systems together. So this is where we are from a get ready to something that's coming and that you can start using now and gain some advantages over your competition. What other examples, or if you want to continue on this, I can, I'm perfectly fine with that, but I know you prepared a few other examples as well.

28:03Oh, sure. Sure. I mean, I think one of the things that we're seeing in the future of multimedia, let me show you this real quick and then we'll get eventually, we're going backwards from macro to micro. All right. Okay. But let's show you what the kind of the future of multimedia looks like. And this is for podcasters. This is for people that are just like you and me. All right. So this right here is Adobe Firefly for video. Yep. And then I'm going to just show you this real quick. You can create your own custom music and soundtracks, as well as finding sound effects based on what you are seeing, which is mind blowing to me because I didn't even know that was possible.

28:48Furthermore, if you want to actually generate your color grades using text, you can tell Firefly what you want and it will just do it. This looks extremely promising. Adding on to the captions feature already involved within Premiere Pro, it looks like we're going to be getting some more captions which would be awesome. You can actually create animated 3D text. Cool. For video, my favorite thing is finding b-roll. To be able to click a button and have Firefly analyze all your clips and automatically put b-roll into the timeline would allow us the option to create rough cuts and rough edits extremely fast.

29:28and one of the most mind-blowing features of Adobe Firefly is the ability to generate storyboards and not just storyboards. You can also make a previs so you can theoretically see what your project will be like before you even shoot. This is changing everything. What do we see here? First off, if you want to have one of the hardest things you can do if you've ever built video is to basically what's the feel like? What's the music like? You can simply just speak it into existence and you're going to get examples that might be a fit. It's also going to read your videos to tell you what it thinks it should fit.

30:07From a sound effect standpoint, it's going to read your videos and it's going to anticipate there's a wave. Basically, there's the ocean. We recommend a wave sound and we're going to basically, so there's no sound mismatch. We're going to figure out as the wave goes from the right to the left, how to make it so there's no sound mismatch. you can do all the color grading you can change the type of day time of the day this is like he mentioned this is my favorite portion it reads my transcript here are some nouns shoes and chalk and wall when i go and i make the videos here's shoes here's chalk here's wall i sell video as a service right over there in front of me is our video studio we bring clients in and we say, Hey, in, in the first scene, we're going to go right up to your office.

30:53And then two people are drinking coffee. And then the next person says this, and the next person says that Adobe basically is going to take my whole transcript and they're going to do that process for me. So you'll see what the movie looks like before it's even filmed. And it'll even 3d render what it should be. So when my video team basically gets there, we basically just follow the rendering. So all of that's available right now. So totally mind-blowing. I'll add two extremes, one on the lower end, one on the high end, to connect the dots for people who have never played or seen these kind of technologies or demos.

31:29And again, great choice because it's a great example of where content creation is going. So on the high end example is very shortly, there won't be a need to film anything because that initial rendering will be good enough to be the actual video. So you will be able to script something with a generative AI tool while throwing ideas. So let's connect a few dots together of stuff you and I know. When I run meetings today, every meeting is recorded, summarized, and action items are created with an AI. So I don't do any of that anymore. And it's reported straight to my CRM, meaning the ability to listen to humans, communicate, understand what the conversation is about and pull out the right stuff is already there.

32:15So you can do your brainstorming session about the video, have an AI listen in, take notes, say, oh, here's what you guys were talking about. Is this what you mean? And it will be instantaneous. There won't be now somebody has to go to the office, do their thing. Let's meet next week. You'll be able to see a rough cut of a video right there out of the brainstorming. And then if you say, yes, this is what we want, but change this, add that. We want this lighting, all the stuff that Chris just showed us, you can have a full ready to go video with a single session versus three months without any video people, without any lighting people, without any sound people, without an actual location, actors, editing, all of that goes away.

33:03And it will be very soon at the same quality as shooting live and it could be with any actor and like we said before in any language and so on so this is on the high end where this is going on the low end for those of you who do not produce video but still want to say okay this is amazing how can i start using stuff like that tomorrow i now use daily every single day dolly three within chachipiti and i'm a and i'm to put things in perspective i'm a die hard mid-journey junkie i think mid-journey does amazing things but the ability to be able to create images within a chat context. And I'll give you a very practical example.

33:45I'm speaking in a conference on Monday. And the last thing I do is I create images for the presentation that I'm doing. And I need images. And I literally went to Chachi PT and said, hey, listen, I'm giving this presentation. And this is the kind of audience I'm speaking to. This is the setup. I'm on stage. And this is so I need big graphics because I'm writing very little just big things. It's a dark background and I want it to pop off the dark background. So I give it the context and then I say, okay, in slide one, I'm going to talk about X. Give me three ideas of what should be in the graphics.

34:18And it always comes up with awesome ideas. So now I'm brainstorming with somebody who has done a gazillion presentations, basically any presentation on the internet. And it gives me great ideas of what could be the graphics. And I pick one and give it some comments and what it suggested. And it generates the graphics for me. And then it's an iterative process. I'm like, oh, this is great, but it's all black and white. Can you make this in color? And then I get a color variation. Well, this is too colorful. Can you dumb it down a little bit? And I also want this guy out of the picture, but instead I want that.

34:48And that's it. So something that used to take a team of graphic designer and back and forth and potentially shooting video that will take weeks happens in minutes. And it's absolutely magical. And the implications for that for any business, whether you're creating content today and how much more content you can create or businesses who never created content because, oh, it's too expensive. It's too much of a process. We don't have the time. We don't have the money. No more excuses. Yep. I'm going to even show you one other thing that I'd use. Have you seen Pi? Yeah. Yeah. So this is just for your audience who's probably not up on everything like you and I are.

35:31So this is Pi. And so this is the future of Siri, but this is also Pi. And Pi is connected to the internet. And so I could literally, if I wanted to prep even better for this show, what I could have done in theory is I could have went to your blogs or your videos or whatever it was. Maybe I'd have to use Glass to get a transcription and I would import it into Pi. and I would literally ask it, hey, what has he said in the past? He's very complete, very comprehensive what he says, but what might be some of the gaps that he missed? And Pi is like incredibly intuitive and it speaks really well. So let me just give you a real quick demo of what I'm talking about here.

36:09Hey, Pi, I'm doing a podcast live. I'm wondering if you can show the firepower of how you communicate and how you talk. And then give me an analogy of, if I gave you a bunch of information, how I could find gaps in my content, what would you do? Podcast host extraordinaire. Sure thing. I'm always happy to flex my digital muscles and show off my linguistic prowess. As for gaps in your content, I'd love to help fill them in. Here's what I could do. Analyze your current content to identify any missing topics or information, help you brainstorm new ideas and angles to explore, provide suggestions for ways to organize and structure your content, offer research support to dig deeper into specific topics, recommend relevant stats, quotes, or sources to support your content.

36:58So let's dive in and make your podcast content even more compelling. So this will talk like a human being while I'm driving and I know I got to write a blog. I'm going to say, Hey, this is my first draft of the things I'm thinking. Here's three articles that are on the same thing. What are they missing? When I write my content, what angle do you think I should take? And the AI sort of just tells me. So we're talking about content creation, but one of the things we don't see is our blind spots. And this is what the AI can do. It can fill in the blanks and say, you know what? Really cool here. But are you thinking about privacy with AI?

37:32Are you thinking about how it's going to impact the international community? All of these things are at your disposal. It's like having someone in the room that has a blank slate that can say, oh, you know what? You should talk about this. Oh my gosh. I never even thought of that. I was always going to the left. It's amazing, but it's just here right now. So you just gave me a very interesting idea that I'm going to try today because I never thought about it. But now that I saw it, I can't unsee it anymore. So there's a web plugin for Chrome called WebPilot, which allows you to basically query any website you're on with AI.

38:09So instead of looking through a website for any content, you just open and it's a keyboard shortcut that you you can program, and it pops up that window of webpilot, and you can ask it any question about the content on the site. What I've never tried doing is asking it what you just said about the content that's not on the site. So I've been following the site. What are they missing right now that I can talk about that I can serve to their audience? And I think that could be a very interesting exercise. I don't know if webpilot knows how to do that, but I'm going to try it later today. Yep. Let me show you other things that will feel very weird at first until everyone does it.

38:49So first off, you need to know there's a massive jump in chat GPT where literally it's like saying we gave the AI eyes. What I mean by that is it's multimodal, which basically means you can take a picture of something, you can highlight it, and it's going to talk to you about that item the same way it would do if you gave him like a paragraph to look at. But here's where I believe that it's going is that you will be in meetings and say you have a whiteboard in your meeting and you're drawing there. You already saw with Pi, it has eyes and it has a voice. It's going to say what you did and it's going to know in computer talk what you're doing.

39:27And it's going to be someone who joins you in meetings that speaks up, that tells you these things. Like literally it's like having a robot in a meeting. Only that robot is considerably smarter than everybody else. And so one of the things that's very interesting about the new Bard is that you can have a conversation with the AI and then you can create a public link. And as a thought leader, you can share that public link and you can say, this is all the concept. This is how I got from point A to point B. You take it and run with it and take it wherever you need to go. So as an influencer, you can start the conversation and you can let the world run wild with it.

40:06Yeah. Basically open sourcing your thoughts. crazy. Yeah, it's brilliant. I'll say something about the visual side of ChatGPT or BARD that you can start using today. It's available if you're paying the 20 bucks a month for ChatGPT. The amount of use cases that you can do with that are absolutely endless on so many aspects of the business. And I'll give you an example. One of the presentations I'm giving shortly is to a home automation conference. And they go to people's homes and they offer them different automation solutions. So anything from electric blinds to surround sound music or all house systems to TVs, etc.

40:53And one of the things you have to do is you got to write a proposal for that, right? So you got to run cables across the house for different things, either sensors or speakers and so on. And to do this, you need to start with a floor plan of the house. So today there's already apps that you can scan the house with your phone and it will create a pretty accurate plan with the right measurements. But then all I did is I plotted with a pencil lines on the floor plan over where I think wires should go. I uploaded it to ChatGPT and I said, okay, can you help me figure out what's the total length of speaker wire that I need based on the red lines on the thing?

41:28And eventually it got it wrong because it didn't do the right calculations, but I helped it figure out how to do the calculations. And I got a number. This is a process that I could have done on my own. It would have taken me, I don't know, half an hour to an hour. But I also connected it then to the internet to get the pricing of the cables and created a proposal that includes the hourly rate that it will take me to do this, combined with the length of the cable, combined with the price of everything and all the materials that I need. And it took me the whole thing that would have taken me probably a few good hours to do, took me about 20 minutes.

42:00And that was the first time that I did it. So when I do it now, it will take me about five because I have the template on how to do it moving forward. These are the things that these tools enable to do today. And this is just one example, but you can take it to any business in any aspect of the business. And you will be able to find these use cases where stuff that you're doing right now regularly and takes you hours can take you minutes. Yep. I spoke with a forklift truck company. They were a$500 million forklift company just outside of Philadelphia. And we were like playing around with the technology.

42:35And this is future of where it's going is, Hey, forklift company, I have this part. What if I'm selling Caterpillar forklifts, which forklifts even use that part? And it gives me the answer of that part is in these eight forklifts. And then the people that are in the field that are literally reading manuals to fix these things, they're, or they're calling their support center and the support center has to go and look it up, stuff like that. You'll literally be able to have the whole instruction manual, but you will literally start to talk to it in natural language processing. So it's not like looking it up on the internet.

43:10You can literally be pretty dumb in the field and be like, Hey, this is part eight, six, seven, five, three, oh nine. Where does this go? What do I do with it? Okay, cool. You know what, what you told me doesn't match up here. I'm going to send you another picture. And they're like, Oh, you're using the wrong tool set. It's like literally having an assistant that has eyes and has ears, you know what I mean? It's going to be crazy. And you know that AR is going to be built into this too. It's just a matter of time before AR is built into it too. I think that the DYI videos that you get something from Ikea and stuff like that.

43:45I think with AI, literally you're going to start to build it. You're going to have your laptop in the room and it's going to have an arrow. It's dude, put the screw in right there. and you're like, oh, yeah. For those of you who don't know what AR is, it's augmented reality. It's basically the ability to overlay data on top of visual world. So you have glasses that are transparent, like you can see through the glasses, but there's projection on the glasses. So think Terminator kind of stuff where he knows who people are, what they do or what they say, same kind of thing. There's already several different companies generating gear like that.

44:21And it's probably going to become something that we wear regularly in the future. Definitely. My kids will jump all over something like that. Yeah. I have. So I ride a bike like cycling and for years they've had a company that literally you don't have to look down at your odometer anymore. Basically it's proven that if you look down, you slow down. And so you can continue to look forward and your whole odometer is like a dashboard cam the way pilots see it. Yeah, exactly. Exactly. And all of that is for the taking. Amazing. I'm going to give you the five minute example of what I call the future of ChatGPT with extensions.

44:59This is the future of where ChatGPT is going. Obviously, ChatGPT is not connected to the internet, but the other large language models are. But this is how it's going to work is that you're going to go and you're going to say, hey, you know what? I went to a restaurant in Ambler and I really liked an item on the menu and it had shrimp on it. And it says, hey, I looked at the menu and these are the three items. And I said, yep, that was it. Number two, how many calories are in that? And it says, according to the menu, what has a lot of calories in that. And I say, you know what? Can you make me an alternative?

45:28And it's going to say, you know what? I just made you an alternative for 680 calories. And I said, you know what? That's perfect. You please connect my Instacart plugin. And they says, we did. And those ingredients will be at your house in three hours. So what do we learn here? We've always bought products online, let's say I'm thinking through Google. And then you have interruption marketing here with Facebook and social media platforms, but there's literally going to be a third lane on the highway. Alexa has always been trying to do this, but this is going to be a third lane on the highway where you're bypassing Google and you're bypassing social media.

46:03And you're literally just going to talk about products and services. It's going to show up visually for you. And you're going to be able to buy products and services. This is like a whole third lane of the highway. And when you're trying to say, well, what do I do now? Investigate what plugins and extensions are. They're not too hard to build, but start thinking this way that if I'm selling my products through the traditional means, how do I be the first one to implement it in another way? Even if I'm an early adopter and I don't get much value from it, it's absolutely going that way. In fact, it might leapfrog everything eventually.

46:40So yeah, this is... Case in point. it was like, I'm sorry, like I'm going to Chicago. I have$2 ,300. I want to stay there for two days. And it's going to say, here's your flight. Here's your, here's your rental car. And here's your hotel. And you're like, yo man, I never fly spirit airline. And it's okay. Let me readjust everything for you. And it'll just do it. Yeah. So the idea of agents that can actually take actions beyond just giving you data or generating content is something that again, I think will mature in this coming year and we'll be able to see the abilities to do exactly these kinds of things, right?

47:18And take it to any aspect of your life. And even the travel side of it, if you think about travel, and I used to own a large travel company, so I very well connect to your last example. If you think about how most leisure travel happens, you don't know where you want to go. You don't know exactly what you want to do. What you do know is you have a summer vacation of two and a half months. And you know that in that summer vacation, if you're not gone for a week and a half, somebody is going to lose their mind or die because it's not going to end up well. So we got to go and get away for X number of time.

47:50And I can get off of work for 10 days and I have a budget of$4 ,200 for a family of four. And I live in place X and I like hiking and I like the beach. That's how most vacations start. They don't start with, let's go to New York. But today, there's no way to search this way. The only way to search is start reading blogs or where people are hiking and can we afford this? And then you start going and looking at the airline. And so all of that goes away. You'll be able to say exactly what I said, which is, I want to take this vacation with this amount of people. This is the budget. This is what I like to do and it will give you 10 examples and they will be able to filter it again just explaining what you like or don't like about the different examples i would a thousand percent spot on

48:44uh any questions for the union so i see a question from will said i came late to the webinar but do you have use cases for health care and financial services that are highly regulated Currently, the compliance and legal folks of these firms on holding Gen.AI efforts back because we don't know how to manage the risks. The business leaders want to take advantage of the power of Gen.AI. Any examples you can share would be great to hear. Thank you, Will. So I will let you answer that first. I think that the first one is some of the bigger players are already there. I think one of the things I would probably look into investing in is Bloomberg.AI.

49:24Bloomberg AI literally has the right sidebar. It has, what is it? A hundred years of proprietary data behind it. And what that's going to do is that's going to arm your team with information that very few people have at a speed that they would have. And because it's in natural language processing and because it's connected to a large language model, it's not like it's just, this is information and here it is. When they ask a pre-qualifying question, you can continue to go back to Bloomberg AI. I'll add one thing to that. The issue with proprietary data and then obviously regulated industries like HIPAA for healthcare and legal and finance and so on, the ability to run self-hosted models exists today.

50:08meaning you can build your own in-house models running on your own servers with your own security layer, just like you treat the rest of your data and still be able to enjoy the benefits of all the things that are talking about. It's obviously a much bigger effort to go that step, meaning if you just want to use ChatGPT, awesome. You want to use ChatGPT through an API, awesome. You want to use ChatGPT through an API and your own database, still relatively easy. Once you start going to, we need to host our own servers with a open source third party platform. Yes, it's doable, but that's a relatively big IT process.

50:54I think that every single company within those regulated industries will have absolutely no choice but to go down that path. And the reason I'm saying that is twofold. One, that's how people will engage with businesses. Nobody will want to call a call center and talk to four people. They will want to ask their app a question and get an answer immediately with links to everything that they need or with an immediate answer to what they were asking without any links to things that they need. So that's number one. Number two is once the first company within that sector will do that, everybody else are fucked.

51:28Pardon my French. If you are in a healthcare provider and your direct competition will provide this interface because they took the time, the money, the effort to build it into their system that allows you to chat with a, whatever the thing is that connected to all the data without exposing it to the world, then nobody will want to work with your company anymore because we'll feel archaic. It's as if somebody will ask me to send them a fax. I'm like, how do I, like, where do I get a fax machine? This is where this is going. We'll have, nobody will have a choice. And obviously the smaller you are and less regulated you are, the faster you can move.

52:05But going back to what we said early in the beginning, Will, is we said, it's all about thinking how you need to restructure your business in order to support this new technology and give the maximum value to your clients and prospects. And in the healthcare or the regulated industries, it will involve how do we build a large scale solution that is hosted in-house, that is HIPAA compliant, that can be within the regulations of our industry, but still provide these benefits to our clients as much as possible. And one step farther is that AWS is partnered with a Chachabit's competitor called Anthropic.

52:43They have a large language model called Claude, which is HIPAA compliant. It doesn't train on the same data if you're using the enterprise version. If your organization is using AWS from an enterprise level, you're going to have that built in, but you're also going to have open AI, not open AI, sorry. It's so easy to go down that road, but open source language models in your same WAWS platform. And so you're going to have certain things where you're going to have proprietary information, HIPAA related information that's locked down in a gate over here, but you're also going to be able to have a sandbox that individual employees that are not using patient sensitive information or PMI, I think it's PMI, patient medical information, yeah, PMI over here.

53:27So you're going to have the best of both worlds. But Anthropic right now seems like they're the big player. Not sure if you've been following the news. They were literally given$4 billion like it was chump change from Amazon and they are going to be the real player, but you're going to have these choices of it's not either or, it's probably going to be both. Yeah. And I'll add to what you said. Now, all the big hosting platforms, so you're talking AWS from Amazon, but also Google Cloud, as well as Azure for Microsoft, all basically partnered with all the big players to allow you to use their tools built into the hosting platform almost seamlessly.

54:03So it will be a lot easier in the very near future to just take the data you already have on these platforms and be able to overlay whatever AI engine you want on top of them, either a base model or any trained variation of it from any provider just built into the hosting cloud solution. Yep. And I also believe that they're going to... Oh, let me... One last piece. Yeah, yeah, yeah. Go, go. You're going to have fine-tuned models. What I mean by that is it's called narrow AI. And let's say hypothetically chat GPT. Chat GPT is strong because it's wide. Fine-tuned models. Imagine it's good at one thing and only one thing.

54:47But if I am a cardiothoracic surgeon, you can literally upload every publication on heart surgery possible and a fine-tuned model. You'll literally be able to use that as an unlimited reference guide. So the medical community, even though that public information is going out, they're going to have access multimodal, whether you can take pictures or have access to the data. It's like almost having everything that's ever been written and you can talk to it conversationally. So while it feels like regulation, there's too much power and too much money behind AI that these walls will come down very quickly.

55:21The iPhone was not used in year one because of privacy and regulations. How many businesses have iPhones? We've been here before and we'll be here again. Yeah. Chris, this was an amazing conversation. We touched on a lot of important things. I'm sure it will be a serious eye-opener to anybody who listens to this, because we jumped into some of the more frontier stuff that I usually don't jump into. I think the biggest take here is you have to get ready from a business perspective. It's not one of those things that I can put my head in the sand and it will pass. It won't. And the way I refer to this, there's a tsunami coming to shore.

56:05And it's like any tsunami, it's coming and it's going to hit the shore and it's going to break everything on the shore. and you have two options. Option one is to build a really sturdy boat and ride the wave or sit on the beach, drink your margaritas and wait to see what happens. And it's not going to be pretty if you sit on the beach and wait for the tsunami to hit. And so I'm not saying this to scare anybody, but the flip side, to encourage people to start understanding what this can mean for their business, what this can mean to their industry, what this can mean to their competition, what this means to the needs of their clients.

56:41Some of the needs that you're serving today will become irrelevant and unnecessary in a year. And we touched on a lot of things today, but translation services are fucked. Nobody will hire a translation company ever again. It's just not going to happen. And so it's these kinds of things that are like, what in my current offerings is going to disappear? And on the flip side, what new offerings can I offer with these new capabilities? Chris, if people want to follow you, work with you, learn more about what you do, get more of your brilliant knowledge, what are the best ways to do that? Hit me up at info at fairtech.com.

57:20I give a free one-hour presentation of basically, again, the state of AI, where it's at, where it's going, and how you can become prepared. That tsunami you're talking about, that's what I do, is that's what I train companies to know. and then I, this is one of my favorite podcasts I've ever been on. I got to admit, like usually people are like, when I talk about AI, they're just like, I used Jappi GPT once. Is that what you're talking about? Is it the Terminator? But plug into shows like this, get information. I have another one that is called digital marketing masterclass and another one called business innovations.ai.

57:56All I do is talk about AI literally all day. The knowledge is power. And I highly recommend that you just get as much information as possible. But if you don't have the time, lean on people like us that are swimming in this every single day, and we will help you get to the next level. Thank you very much for this opportunity. This was, this is one for the, what's it called for the, this is one for the ages. So thank you so much. This has been amazing. Thank you. This was really really great. I thank everybody that joined on LinkedIn. I thank everybody who joins us on the Zoom. I usually I say, I hope you got something positive.

58:30I know that everybody got a lot out of this. This we really touched on a lot of great stuff. I appreciate you. And I appreciate you taking the time and sharing your knowledge with all of us. This sounds great. Thank you so much. Bye everyone. Wow. What a great conversation with Chris. I really enjoyed the conversation in the back and forth, literally just to people who like AI and like growing businesses geeking out on the topic. I hope you really liked it as well. The biggest take for me from this conversation is that there are literally endless use cases available at your fingertips right now without paying any significant amount of money and with a very easy learning curve that any business can do and should be doing in order to be more competitive in the immediate AI future.

59:22And now for the news from this week, as I mentioned, two huge news came out this week and all the rest you can go and check out in the show notes. I'm going to put links to them. So if you want to educate yourself, you can do that. But since this was a relatively long episode, because all of our live shows are an hour long, I will focus on the two gigantic news of the week. The first one is from OpenAI. OpenAI had today, Monday, November 6th, they hosted what they call their Dev Day, basically their conference for developers who are developing using the ChatGPT API. And they made some huge announcement on their Dev Day, and these are earth-shattering when it comes to AI and the implications of what they did.

1:00:07What they've introduced is what they call ChatGPT4 Turbo, which basically relies on top of the existing ChatGPT for, but then adds and changes a lot of things in it. The first thing they announced about that is that the context window, meaning how long of a text can you upload to ChatGPT, grew from 8 ,000 tokens or 32 ,000 in some use cases to 128 ,000 tokens. This is a huge improvement. I personally am using Claude2 from Anthropic for a lot of the long text tasks that I have to complete either for myself or for my clients. And the reason for that was even within the context window of ChatGPT, it would stop and halt and get stuck and stuff like that.

1:00:54And this is going away. So they right now will have the longest context window there is with 128 ,000 tokens. For those of you that don't understand what that means, about 75 % of that is the number of words. So we're talking about 100 ,000 words-ish, which is about a 300-page book of an average book that you can now upload to ChatUPT and interact with. So you can upload or get answers with a huge amount of data without any customization straight out of the box. The new model is more accurate than the previous one. It provides more control for developers with the API. It provides for reusable outputs, meaning you can add parameters to get higher consistency, which are really important for business use cases.

1:01:43They've also announced that it now knows stuff up to April of 2023, where before it only knew stuff up to September of 2021. It's not fully up to date, still seven months behind, but they said they will keep on updating that to close the gap further and further. But in addition, they've announced that it now is a true multi-model, meaning all the different functionality that so far you had to pick from drop-down menus. So you could choose to have advanced data analysis. You could choose instead of that to have the ability to browse the internet. You could choose instead of that to be able to create images with DALI 3, or you could choose to use the regular quote-unquote model, which can interpret vision and images, and now they're all baked into one model, which is absolutely incredible.

1:02:30This opens a huge, like an entire universe of use cases and opportunities. And the other cool thing that relates to that is the same functionality. So text-to-speech in multiple languages and multiple voices, the vision capability to analyze images, advanced data analysis, and so on, are now also available through the API, meaning for third-party platforms and developers to be able to use that. This is an incredible, incredible step forward when it comes to opening AI capabilities to the masses. They've also announced what they call GPTs, which is basically the ability of every single person to develop GPT-based apps for customized specific tasks.

1:03:16And in order to program that, you use natural language. So you can program your own versions of ChatGPT to be very good at something very, very specific tasks you need for your personal life or for your business. And you can then package it as what they call GPTs and reuse this thing again and again and again without knowing how to write code. In addition, there's going to be an app store for those GPTs and people who will create ones are going to get used a lot are going to get revenue share from OpenAI. This by itself is a game changer because any person can now develop apps for ChatGPT because they are quote unquote programmed using natural language and you can post those and make money through that if you make stuff that's actually usable and helpful to people.

1:04:05In addition to all of that, they've announced that GPT-4 Turbo is going to be two times as fast as GPT-4 that we use so far, and companies will be able to request and pay for even faster speeds. And all of that will happen while they were able to make the API three times cheaper for most use cases. So a huge, huge announcement from OpenAI today. I still can't even wrap my head around on all the implications of everything they've announced today. I will most likely record a separate episode just to talk about use cases and what that means. So more on that in the near future. Sam Altman at the end of his speech said that when they do this again next year, everything they've announced this year, which again is mind-blowing, will look like peanuts.

1:04:59So just be prepared for a lot more of extreme changes and announcements in 2024. And speaking of big announcements, another huge announcement this week comes from no other than Elon Musk, the richest and one of the most influential people in the world. And he announced that X.ai is going to release the beta version of their new AI chatbot called Grok, and it's going to become available to X Premium Plus subscribers. he claims that on many different aspects it is better than gpt 3.5 and on some aspects it's the best in the world today now time will tell how accurate that is but the fact that we're able to spin off a brand new model that quickly is incredible it shows that the world is learning how to develop these models powerful models faster and faster than we did so far but if we connect this to Elon Musk again, Elon has a very long history with AI.

1:06:06He was one of the founders of OpenAI, the company behind ChatUPT. He wrote a pretty big check for that, and he was on the board until 2018, where he left due to disagreements with its current leadership and Sam Altman, who is the current CEO. He since then was involved in several different AI-related activities, including obviously self-driving cars at Tesla and the Tesla robot that they've already done a few demos for. Now, as you know, Elon purchased X or what used to be Twitter and is now X, which provides him access to basically human communications in the last decade on a very detailed level.

1:06:47So from a training data perspective, he has access to anything from basic day-to-day communications to the biggest thought leadership announcements that happen, because a lot of it happens on X. Combine that with the fact that Tesla announced that they collect 1.3 billion frames of videos of the real world every single day just tells you how much data Elon and Elon's companies have access to when it comes to developing AI models. They also have their own technology, meaning they've developed their own chips as well as their own architecture on how to use AI. And if you want to make that even more science fiction, he also controls Neuralink, which allows to create a chip that connects to your brain.

1:07:39And he also owns Starlink, which allows to connect all of that to the internet from everywhere in the world. Combining all of that together, I definitely think Elon and his group of companies are going to play a very big role in the AI and its impact on humanity in the near and medium future. That's it for this week. Until next time, experiment with AI, try different things, share what you find with the world on your favorite platform, share it with me on LinkedIn. And until next time, have an incredible week. część