HOW WILL ARTIFICIAL INTELLIGENCE IMPACT LIVE EVENTS?

Hello, this is Les Goldberg, and welcome to The Road Ahead. The Road Ahead podcast is dedicated to the future of the live events business, bringing together industry experts. Hello, production world. This is Les Goldberg and The Road Ahead. Today I have four amazing guests. I have Frederick Van Johnson, the head of content for theweekendphoto.com. I have Ash Nehru, and Ash is the co-founder of Disguise, the media server company. I have Jonathan Hart. Jonathan is the technology director at Molten Immersive Art. And finally, I have Larry Springer, essentially an artist. Larry has joined us as an artist. Everyone, welcome to the show. I'm super excited about today's topic. Today, our podcast is dedicated to the concept of how AI, as well as ChatGPT, so it's artificial intelligence, and ChatGPT will impact the live event space. And I want to start off with a quick question to Frederick Van Johnson. Could you please explain to us in simple English, the concept of AI and what it is? So I get to start with the easiest question, right? The easiest question. Yeah, so artificial intelligence at its core is basically what we like to call machine learning, right? And if you want to boil it down to its simplest terms, it's feeding a computer a bunch of information about a singular topic so that it can start making inferences and understanding how that data relates to each other. For example, in the example of the AI that does generative images, for example, in one of the presentations I recently did, I used the example of a cat. If you want the thing to be able to output cats, you have to teach it about all kinds of cats so we can understand what's common, what's uncommon, how cats behave, how their bodies move and all that. And you do that through just feeding it information, i.e. photos. And then from that, when you ask it to generate a cat, it can pull from that mountain of information and give you a cat in whatever situation you ask for. Same with models like our large language models or LLMs like ChatGPT, feed it a bunch of information about a particular topic, it can start understanding that topic. And then when you ask it a question about that, it can answer, quote, intelligently or artificially intelligently. That's and that's basically how our brains work in a lot of ways. We learn over time what works, what doesn't work and how to do things. And we remember it. And when someone asks us a question and we pull from our little personal database and regurgitate the answer. So they work on similar levels that the brain works, but, you know, arguably more efficient in some ways and less efficient in other ways. You know, that was a terrific, basic example of, you know, how to describe AI. And so I'm going to throw the next question to Ash. And Ash, maybe you could advise, how are people using AI in live events, in the event space? And obviously, this has everything to do with content production, how we're putting things on screens, on LED screens or projection screens, or even as much as sending things out to streams. So what are your thoughts on that? How are we using AI in this new world we're living in? I should qualify what I'm going to say, I think, just by saying that I'm not actually operationally involved in making shows these days. So I don't have, I think, I think I'm probably not so qualified to discuss how people are actually using it for making real shows at the moment. My involvement with AI is much more in the kind of, in the software engineer level. So I've been doing courses and trying it out. And I think, I think the the obvious one is what people are using the journey, and stable diffusion and tools like that, to speed up the art process. I think a lot of people are using it in the, a lot of times in the storyboarding, and the and the mood board sort of area, because it's really quick to generate stuff. The I think the general feeling is that, although it's useful, there's a generic quality to it that that that people recognize, and people know when stuff is made by AI. So it's, it doesn't always communicate like the brand values that you'd hope it would communicate, right? If you're, if you're doing something, something with AI, but, but I think it's, I think, I was saying this to a friend the other day, you know, there's a lot of hype, and there's a lot of excitement, rightly so around this stuff. But it's often useful to when somebody makes a statement about AI to just replace the word AI with software and see if it still sounds exciting, right? Because a lot of the time, these are just new tools that artists have, and then new tools that that developers have, and they can do things that the old tools couldn't, but that's actually always been the case. It's kind of like Photoshop in 2020, compared to Photoshop, and when it first came out, right? The capability of the tools and how they amplify your, your, your capabilities is, at the end of the day, it's how you use them, right? Everybody can make a show, but, but there's still only a certain number of people that make really good shows, and the general standards getting better and better. But that doesn't mean that, you know, the the minimum standard is getting better, but what constitutes a is constantly advancing. And I think this is just part of that pattern. You know, that's a super interesting question. I throw this to Larry, because as the artist of this group, Ash was saying that, you know, we're getting content that's generative, but it's not exactly the highest level of content. And it might be a starting point, or it might be a point for, you know, perspective to see lots of flavors of something. What are your thoughts, Larry, on, on using this technology and being an artist? Well, I agree with Ash. I think it's, we use, what's the word you use, brand? It's, it's, it's, why do corporate, corporate events, you have to stay on brand, and, and keep that, keep their product special. And like I said, it is, AI right now is pretty generic. And it does a couple of really cool tricks. And they're fantastic tricks. But how do you make that tricks work for your message? And I think like in, like last night or the night before, they just showed a secret invasion, a new marvel. And that's a breakthrough, because they're opening credits are like the first AI generative opening credits that's been used in the industry. So it's just now starting to kind of bubble up into not examples of real, real world things. But that's a sci fi show. So I'm not sure if I answered your question. Yeah, you did. And Jonathan, what are your thoughts on AI and how it applies to content generation and the world we live in and how we were using it for live events, whether it's a concert, or it's a corporate event, as Larry was just saying, you know, where do you see AI fitting in? I thought about this for days when I knew what the topic was. And it's basically a, it's kind of like the timing is right, because I just had for a client had to architect an AI solution for a large scale events venue. And while I can't tell you who the venue is, but I can tell you, it's pretty, it's common knowledge in the AI world how to construct something like, I nicknamed it a venue bot. A venue bot could be a kiosk, it could be a BBA, roving a drone around a venue that has a conversation with people will be a voice. And so what I would like to contribute at least is incorporating AI and live events definitely will require a chain of several different AIs, one to transcribe what the will denoise and then transcribe what someone says to this AI, send that to a large language model, like Frederick mentioned, which is ChatGPT, it's, it's at the top of the game, let's assume this ChatGPT, but you can roll your own, but I wouldn't recommend it. It's not economic, economically successful situation there. And then you take that output and another AI will transmute that into, like, generative speech. And it could be any language, it could be any dialect, it could be different accents. And it the AI model, they are really now about being a having personality, it's no longer about just simply telling you the words via audio, it's about having personality to it. And that's, by the way, I should mention meta just released voice box. And it's, it's actually quite incredible in that regard. So that whole loop will let and like any sort of venue bot, let's call it Lonnie, for instance, Lonnie can have a conversation with a four year old or an 80 year old, and, and speak their language and and be be there. I mean, there is going to be this uncanny valley. But it's getting pretty close. You know, it's it sounds like we're going into the future. And this is like a very futuristic concept. And it's real. And it's here. And it's now. So this question is to Frederick, Frederick, is AI being leveraged right now? I mean, Jonathan talked about a project he just did in a super secret place. I'm sure one day, everyone will know. But the the question I have for you, Frederick is, are people ready to deploy AI? Are they getting comfortable with it? But Frederick, what are your thoughts on that? Are we are we deploying right now? Are people ready to start rolling out AI? Yeah, I yes and no, right? I think we're at the beginning at the very alpha beginning of this curve. So you know, there's been a lot of speculation about trying to extrapolate where this is all this is going. But it's almost like trying to guess what the impact of the iPhone was going to be when it first launched with iPhone one and how limited that device was. And now where are we with iPhone 14, and all the capabilities and the new industries that that device has birthed like drones and ride sharing and all the things right. So we had no idea that the world was going to need those things or want those things when iPhone launched. And now we can't really imagine an existence where those things aren't available. Same thing with AI. Right. So we don't know where things are going to go. But I can tell you, in one of the industries I play in photography, the the generative imaging wave has people cautiously optimistic on the one hand and completely terrified. On the other hand, I'll give you an example. A friend of mine is a commercial photographer. And one of his clients makes like pocket knives and kind of outdoor gear type things. And his job is to whenever they release a new knife, is to take photos of it in different situations, you know, glamour photos and commercial photos for magazines, and banner ads and that sort of thing. But what he how he is leveraging AI and the positive side of it is, he can go into mid journey, and just type in a prompt and say, you know, create this particular background, a lush kind of, you know, fairy tale forest scene with a knife in the middle kind of stuck in the ground. So this is what he did literally, with a knife stuck in the middle stuck in the ground, it is the subject and it's lit this way shot with this kind of lens, all the things, then he generates that image, takes that image into another AI application called Topaz Gigapixel AI and upreses it to a print size or to a larger size. And then he photographs the original knife and takes the knife out using Photoshop, the knife that the AI generated erases that puts his knife in scene composites in there, it's perfect. So now he's leveraging it because he doesn't have to worry about going out on location or finding the perfect stock photo as a background, he can get exactly what he wants with just a couple of words. Where I see that going personally, is at some point, the photography of the knife itself won't even be necessary, because you will have been able to load in manufacturers and companies will be able to load in their their IP or their assets kind of like car manufacturers do today, most a lot of car ads are just 3d renderings of the vehicle versus trying to get it right on film. So imagine that in a world where now that knife manufacturer could say, okay, you know, whomever I need a shot of our knife in this situation doing this and this and this, the AI builds the background and pulls the data for the physical object from a 3d from a database and 3d information renders it in place, you're done, that flow has the ability to decimate entire careers, because there are photographers that just do that kind of thing, whether it's people, you know, photographing people, photographing objects, etc, etc. So photographers are afraid of that. And, and it is a very real future that could happen just looking at where things are today. So there's the pluses and minuses. Are they being replaced? I don't replace is a is a tough word. I think it's, it's, you know, in science cut the law of conservation of energy, right? So you can't destroy something, you can only change it into something else. So I think photographers need to much like many careers over the ages, right, as things, new technologies present themselves, the people that have built livelihoods and careers on that thing that is now being, you know, kind of sidelined, they have to retrain and learn new things. I think like I was saying, getting AI is opening up new areas for people to kind of move into and cutting other areas off. I would like to quickly add to that, you are correct in so many ways, but my personal, well, maybe I'm being optimistic here, but I do believe that the human mind is going to be able to suss out real photography from AI photography. I see it already. If I look at something in a social feed that looks beautiful, I can tell it's AI generated. It's not an actual photo. And I think, you know, people it's going to be people like we're giving us not enough credit in terms of being able to recognize that. And I think that's going to be a resurgence of, you know, human generated content, so to speak. Look at look at look at the almost logarithmic curve of the quality that these these AI models are generating. When they first started a couple of years ago or when I first started taking them seriously, what they were putting out was, you know, 20 hands and weird looking stuff. And then you fast forward to this month of 2023, June of 2023. And we're getting closer to photorealism, if not indistinguishable in some cases. So I look at it from the standpoint of, OK, we got from there to here. What is from here to 2025 or 2030 look like? Right. Whenever I look, whenever I look at a photo that looks too epic to be real, I always look for the fingers. Yeah, exactly. This day has not figured out five fingers. Only five, not four, not six, five. Yeah. And of the right proportion, right? Yeah. Let's ask Larry. Larry's the artist. Larry, what are your thoughts? Five fingers, four fingers. What are your thoughts on can can I generate images as pretty and beautiful as the human? Oh, yeah. Every now and then. Sure. Absolutely. And fantastic stuff. But it's, you know, it's it's it's it's it's not it's if you're lucky. It's like you have to put in the right prompts, the right words. It's a let it do. It goes up to the sky and talks to its database. It comes back with something and sometimes it works. Many times it doesn't work. Sometimes it's fantastic. So. But once again, you have to curate. Go through what's to find the right image that works and or the right the right circumstances for that AI to to produce something for you. Got it. OK, let's let's shift this questions to Ash and and Ash. You know, we do a lot of live events and we do a lot of live concerts. Do you see AI being used in a live environment? Let's say a music concert. And obviously there's media servers. And do you see generative content being pushed live onto video screens behind a big artist? Is that something that you you would think could potentially happen? I won't say which kind of media servers either. I'll just I will go with the ones. Oh, I mean, 100 percent. Yeah, I mean, yeah, I mean, 100 percent, because I think I think you're where has breath away from being able to say, hey, give me give me an unreal scene in which this happens and shoot it from this angle, you know, the same way that you do with Midjourney and stable diffusion and get an Unreal project or a Notch project out of it and and and bang that up on the screen behind behind an artist and change it, you know, in in in a matter of minutes to something different. I mean, that was always the the great promise of real time content was that you don't have to re render stuff. Right. But but still, when an artist wants to change, there's quite a lot of work to do to make that change. And so so you can quite happily imagine, you know, a large, a large language model that could basically take over the role of of a notch designer or operator. And you tell it what you want to see. And it makes the necessary changes to the to the unreal project or the notch project. And there you go. And I would expect this to be happening in the next year or so. I mean, I sort of follow vaguely the research stuff that's going on. And there's there's enough there's enough things happening. I would I would imagine that the people that make Unreal and Notch are working very, very hard on making that happen to make that faster and faster. So let me ask this question. We're doing a live concert. We have a big A-list band. And is there a risk component when you go out and ask this generative content to happen live and maybe something comes out the wrong way? Well, I mean, I don't I think that that people asking content we made during a show will probably end up going the way of. I remember there was a big fad for like, hey, let's connect the sound desk to the thing and have all the different sound lines drive different things. And that last approximately two shows of the tour before everyone realizes that, you know, it just doesn't work because somebody forgets to plug something in or the level is wrong or something. And so then they end up just recording it and playing it along. So it's the kind of thing that that's conceptually very attractive. But actually, I don't know how much sense it actually makes for a real show. Usually you kind of want the same thing every time within differences. You want something predictable. So I would I would I mean, again, I'm almost certainly wrong about this. But I would see at least in the near term future, AI being used in, you know, a conventional head. We're going to we're going to build a show. We're going to rehearse it. And it's kind of going to be can stay that way. So so each show would look the same. So if you did it the other way, live, every show would have a certain uniqueness to it to be one of a kind. Yeah. I would I mean, the thing is, you know, I started my career doing shows that changed every night. And, you know, you know, that was sort of the thing that that launched us in the business with the ability to do that. But I think that that stuff is still few and far between. In my experience, like, actually, people just want something predictable. They want they want to know what they're getting. They definitely want to evolve it as the tour continues. And I was going to make that easy to happen. But, you know, I'm you know, I think it probably depends on the artists. Right. There's going to be people who just want to there's going to be people who want to take that risk and do stuff completely live. And I look forward to seeing that for sure. The risk reward. It'll be interesting. This question is to Jonathan. Jonathan, we're doing corporate events and, you know, the the content for corporate events. Do you see AI becoming mainstream for corporate events? You know, maybe budgets aren't what they used to be. And they're trying to figure out a cheaper, more efficient way to create content instead of doing it the old fashioned way. Maybe they will rely on AI for a percentage of content. And I wanted to know, what are your thoughts? Well, in the context we've been speaking, actually, Ash brought up an interesting topic that I would like to touch on briefly is, you know, these AI can answer anything it's trained to answer. It's it is basically Pavlov's dogs to an extreme level where it they're conditioned. So the great thing is, is you can essentially make a conversation or AI that is trained to like a given corporate brand. Like an airline is a perfect example. It's like, OK, tell me what is the cheapest time for for me to fly Delta Airlines, like to go to Barbados, that kind of thing. That's accessing real time data, but it's it's structured and it's contained in the large language model that it is trained to speak from. So there's that. And but but also ultimately, AI in a corporate art setting or corporate experience setting, as it were, it really is more conversational than it is generative art like visual things. Rafiq Anadol is known for being one of the greatest artists who uses real time data and AI to actually generate a beautiful experience. He's at MoMA right now, I believe. But. If we're talking about AI and the corporate sense, it's all about conversation, it's all about using basically the ChatGPT, large language model paradigm. And I do believe that that's that focus. There is more control than people think. You can massage the input where you can tell ChatGPT like whatever that person says, like tell me like like they could say something very naughty, for instance, and ChatGPT won't even know about it because we can actually add. Just respond in one sentence as if you're addressing a child, and that's the input into ChatGPT, and that's what the output will be. So there are certain like safeguards. You know, you go bowling, the bumper, the bumper rails. We have bumper rails for this technology, and a lot of people like to talk about how scary it is, but it's actually not that scary because we actually are able to control it. OK, my next question is to Frederick. Who owns the images that come out of AI? Yeah, good question. And who is the rightful owner? If Larry produces something and he takes an AI image and modifies it and makes it his own, is Larry on it because he's done that? Or is it where does the ownership of AI live? Yeah, that's a big, big question right now that's in the courts, I think. And we're trying to figure that out. Currently, the depending on the the the the AI that you're using, the general consensus is whatever company that you're generating that image from owns that image. Like so for the most part. So you're not going to use any of these images. For example, if I create a great image that rivals the Mona Lisa somehow and I up res it and print it, I'm I'm in trouble, right? If I don't substitute substantively alter that image, I think what does the copyright law say? Something like 30 percent of the original image needs to be altered before you can start calling it your own. And we learned that in that whole Shepard Fairey controversy with the Obama picture back in the day, that that's when that kind of came to the light. And then before that, it was with the in the music industry with sampling. Like how much can you sample of a song and put in your song before you're actually stealing from the original artist? This is all of that mixed together times a million, because we've got a ton of images in this database that that the AI is looking to, like I was saying before, to build whatever image you're asking for. So it's not it's not physically taking pieces from different shots. In most cases, it's taking what it learned and building that new thing based on that. So I think the the controversy around ownership goes back to if you use if you compare AI to humans. Right. So if I'm an apprentice and I'm shadowing one of the old masters and I learn that technique and now I'm painting like that old master, is that work mine or does part of it belong to the old master? No, it's 100 percent mine. So AI is doing this at scale with billions of images and learning how to do that, you know, kind of shadow billions of artists at the same time and recreate their work. But because it's so efficient, does that make it, you know, somehow better than what what humans are doing? So right now, I don't know. No one right now, personally, as a photographer, I wouldn't I wouldn't rely on creating a piece of AI art and passing it off as my own, because the community, for one, would backlash and come after me with with pitchforks and torches and all that. But if I want to use part of that, like the example of the knife photographer, I want to use part of AI to augment another image or be subordinate to the main content of that image. Then I feel better about it. But, you know, the legal stuff is it's in court now. Everything from companies like Mid Journey scraping one of the companies I work for Flickr. So they went in and scraped a bunch of the data from the Flickr database and fed their AI. And now, you know, some of the images come out may have been taught by or some of the images that come out will be have have been taught by Flickr. Same with Getty Images. You know, that's one of the famous examples where people generated images out of Midjourney that still have the Getty watermark on there. Right. So they're they're right. Imagine that. It's hard. It's hard to defend. Right. So I really wanted to add to what you said, Frederick. Yeah. People think, oh, AI art is generating new content. It's not. It really is. If you dial it all down, it's basically a trained model on stuff humans created. Yeah. And it's impossible for us pixel to pixel be like, oh, that came from that photo that was, you know, ingested into this neural network. It's impossible to trace anything. But at the end of the day, this neural network didn't invent this. It's actually interpreting. Yeah. And my my I agree 100 percent with that. But the other side of that coin is what are we doing as humans? Right. We're we're influenced by everyone else's art. And, you know, where musicians are, you can't not be influenced by someone else's art unless you put yourself in a box and create something brand new even then. Right. So the A.I. is kind of doing that same thing. I think that's the argument that we'll hear about in courts. A.I. is kind of doing that same thing where it's inspired by it. Let's call it that, you know, and on a programmatic level, it's inspired by billions of photographs and then creating something that it thinks is original based on that or audio or whatever. So this question goes to Larry. Larry, do you ever feel like at any time as an artist, you would ever be replaced by A.I. or someone who says, oh, I can do what he does, but I'll just use A.I. and I'll give you the same images that you want because you need a picture of the Brooklyn Bridge and I can just put in Brooklyn Bridge and I'll get 100 images of the Brooklyn Bridge and I'll just pick one versus someone coming to you and giving you a perspective and you in mind as to the time of day you want the shot and the coloring, the lighting, maybe I wanted it midnight. If they type in some words and prompts and create an image, then they're an artist. So yeah, if they get results from their actions and it works and it's compelling, fine. And but no, I see nothing but great opportunities and everybody can do great things with it. So A.I. is limitless in your opinion, and especially as it continues to grow and evolve and becomes a tool that everyone relies on, kind of like the iPhone. If you go back to the old Motorola construction phone and then we have an iPhone, well, the iPhone does everything. And so this will evolve as well. Yeah, it's just another great tool. So that leads me to my last question. I'd like everyone to apply. I'm going to start with Ash and then I'll go around the horn here. Ash, it's now 2023. So take me two years from now and say, what will we be talking about about A.I. in 2025 and how will it have evolved? And what will the discussion be about? Well, I mean, I think to me, the big battleground is really going to be between A.I. that you can control yourself and run on your own machines and use in the way that you see fit versus A.I. that is essentially a service provided by big companies that's completely opaque and out of your control. So I am a fan of the idea that people should have their own machines that they can teach to do the things that they want to do and that they can control. I'm not a fan of the idea that all of your data gets sucked up into a massive machine and that whatever it is that makes you unique can be effectively replicated by your boss or the company that employs you. I think that's going to be the battleground really is between the man and the little guy, if you like. The man and the little guy. It's interesting. And so I think both of those things are going to happen simultaneously. It's going to simultaneously be a force for, I don't know, exploitation and pushing down people's wages and making work more scarce. But I think it's also going to be a force for liberation. I think that we are going to see things like the A.I. CEO. It's going to be for anyone to start a business and to have a CEO, CFO, the whole kind of mechanism of a large company in their computer. And there's going to be some extraordinary changes. But at the same time, I think that there is also quite a lot of pretty dark stuff that's going to happen if we let big companies control this stuff, which at the So there's good and evil, and they're both opposing forces depending upon how it is used, right? Oh, I think good and evil is in everything. I mean, there'll be good big companies that do good stuff and there'll be big companies that do bad stuff. And there'll be individuals that do good stuff and individuals that do bad stuff. I think that's just going to continue. I don't see why that should change. Got it. Larry, what are your thoughts? The future? Take me two years from now. A. I., how is that going to impact us? And what do you see in the future? Well, it'll become more commonplace. A. I.'s generative fields are now showing up in Photoshop, so it's much more prevalent. I'm very struck by Ash's comment about, I always assumed that A.I. was something that was a server base that had to be a cloud base and all that data was being processed. But if you can get something A.I. that works on your computer or your phone, I'm not sure what that is, but that sounds really interesting. Jonathan, what is your thought on what will the future lie with A.I. in 2025? What are we saying about it? I've thought about A.I. a lot, so that's for another full-on podcast. But I will say that there's so much alarmism surrounding A.I. right now that it's kind of overblown. Because if you dig down into it, it's just one big function. It doesn't think for itself. So the alarmism should be subdued. However, I would like to say that Kurzweil's singularity concept is at play. Things are accelerating way faster than we are capable of. Our brains, we don't have the bandwidth to keep up. So we're just going to have to accept and embrace what is going on and use it to our advantage until the robots kill us all. The robots kill us all. There's nothing better than a quote from a podcast that we can all kind of embrace. I will tell you, I'm always worried about someone taking my voice and recreating it and calling one of my people and pretending to be me. So yeah, I even came up with some ways around that. And I wonder if people can tell the difference. So Frederick, take us home. Two years from now, can you tell the difference, the A.I. Les Goldberg versus the real one? Yeah, I don't know. Yeah, but if you look at the way things are heading now, I think there will be a challenge there, right? Because I think if you look at A. I., let's look at the large language models like ChatGPT, you can already say things like, write me a short horror story with this premise in the style of Stephen King. It can already do that. And it will generate something that is very similar to what a Stephen King would generate. Same on the artist's side. Create a photo of a beautiful woman standing in a window in the style of Picasso. It will do the generative A. I., generative models like Mid Journey and all the other ones out there will do that today. In the future, there's two things that I see happening. There's a bunch of things. Much like Jonathan, I've been thinking about this a lot, so one of the ways that I could see, and this is positive and negative, is having the ability from an artist's standpoint to feed in a personal, secure, back to the whole blockchain world and all that, but a personal and secure data set of your work. Let's call it your life's work. Everything that you've created since you started using a certain medium to current. You feed it into the A.I. and now, presumably, it can recreate your style. The interesting thing about that is, will there be an industry where you could then license yourself out or license your style of work out to a company that could say, yeah, I need a photo in the style of Larry or an image or whatever that looks like this. Larry has authorized it and boom, it outputs work that's very similar to what Larry would have created. That's one of them. The other one is, we all know what Apple did a couple of weeks ago with the whole Apple Vision Pro virtual reality VR headset thing. I think if they're successful and if Mark Zuckerberg and Meta and their headset evolved to a level where these kinds of experiences become ubiquitous, I think it's the birth of a whole new way of telling stories in that kind of environment that these kinds of language models and AR and VR will just naturally feed into. I can imagine a time when Siri or one of these other kind of now basic artificial intelligence assistants is powered by these new technologies and I'm in a virtual world and I can ask anything I want and manifest anything I want from within that world. Make me a picture of this, put me on a beach here, do all these things. You'll be able to do that, I think in the very near term, maybe in two years, three years time. That sounds amazing. I am going to have dinner with Henry VIII and then I'm going to go around and go with Jacques Cousteau down in one of his capsules. For education, the whole nine yards. It's the holodeck from Star Trek, right? But not quite at that level, but it's close. It sounds like what's coming is going to be more amazing than we can even imagine. I want to thank all of you today for joining. I greatly appreciate Frederick Van, Ash, Jonathan, Larry. You guys are all amazing and I really appreciate your perspective. There's content in everything we do and the source of content is generally content of producers and now we're talking about how it's going to be different in the future and it's going to be interesting, to say the least. Again, thank you all again and this is Les Goldberg and The Road Ahead. Guys, great job and thank you so much.