From Prediction to Reality: Navigating the Next Era of Screen (live recording)

Frances Valintine (CNZM, academyEX) last set the tone for Spada in 2016, exploring disruption, AI, and the future of screen. Nearly a decade on, the world she forecast has arrived at speed. In this keynote Frances revisits her predictions and charts the next frontier for a sector in rapid transition. 

Recorded live at the 2025 Spada Conference, held 20 & 21 November 2025 in Auckland, New Zealand.

www.spada.co.nz

Subscribe for more screen industry conversations from Spada (Screen Producers NZ).

Transcript

0:01

Frances Valintine's presentation ‘From Prediction to Reality: Navigating the Next Era of Screen’ was recorded So when Sandy and I were talking about this keynote, I thought, why don't we just go back and take a look at some of the predictions? Because it was the first time we started talking about artificial intelligence and film through a very different lens. And I have to say, we've overachieved over that time, depending on how you look at it and which side of the you use it. But certainly we're a lot further ahead than we ever imagined. Back in 2016. Some other things that have changed in that time is when I first came and spoke, I was wearing the head of Tech Futures Lab, which I had founded back in 2016, which is now part of my larger organization, Academy X. So we have two logos on screen, and this is just my way to kind of remind myself, between which era am I in? And the other big difference today is, you know, ten years on of knowledge and ten years more of speaking with professional students who many of you in this room. I was thrilled to know that year Te Aroha Grace, one of our current master's students, opened this morning and just, you know, really make me feel awkward. My feet, my fabulous life partner, William McKegg is in the room as well, which we don't often be. We're not often in the room together. So this is like coming home in a little way. I've been involved with the film industry for my entire life, but I've done that through the lens of a technologist. So to get going, this is literally the screen that you saw if you were here all those years ago. And if you weren't, this is how it looked. It said “the more overdue a disruption and the more sudden it is when it finally occurs, more the more off guard the incumbents are caught”. Now, that was then. And so I thought we just do an intermediary kind of moment in time between 2016 and today. And so I thought I would go back to when AI first made me think, well, we're entering the era that I imagined for many years, and that was three years ago. Now, three years ago, something happened. And many of you will know because it was when ChatGPT was launched through open AI. But just before that, in November, all those years ago, in September, we started to see the first platforms of creative options being formed. And the first for me was with Midjourney. So I thought I'd go back. And so what was my first prompt? What did it look like and what was the result back then? So to me, to be fair, it was really terrible. My prompt was the 2050s diner with a dinosaur waitress and robot customers. Anyone who knows me well, I have two complete fascinations. One is robots and the other is dinosaurs. So I thought, why not? Let's do that. So three years ago, I went back to my history, and this is what I had. And just because. You will want to see this. So I persevered from this moment of time. In fact, I actually got this printed by for a number of people and got posters made and sent it to them saying “One day you’ll look back at this poster and you’ll laugh” we all do look back at these posters and laugh. But I thought it did evolve. And not long after this I keep prompting. And it came more questionable, and I'm not even sure I could explain. The two naked men on the beach. I trust you can. You must trust me that there was no nakedness requested in this prompt. It was, also the Mr Potato Head child. I don't know how that happened. And then you can count if you saw close enough the number of extra limbs and people with tails and strange body parts in the hardware store. So we have gone in just three years. An awfully long way. And so I thought, this is just a kind of a point in time, because when we talk about exponential technologies, it's amazing or easy to imagine that we actually have a fairly sort of a smooth approach to innovation. But actually this is not, you know, the whole hockey stick It's coming at us fast. So I thought, if you're not a tinkerer and you're not playing along with a lot of these tools, I thought, let's just show some of these things. So going back from those early prompts to now, a prompt which I'll just play you, I've got an audio prompter. So when I do prompting I use voice only because I'm lazy and a terrible typist. So it looks a bit like this if I'm using, the Google AI Studio. Okay, so here we’re in Google AI Studio a new tool which is called Nano Banana, under their demo so with this new tool and let’s control the image And now I change prompts. So. So this image replace the background. So that we are in a beach park. And so there's a beach party. there are ferris wheels and change the small boy that we carrying the boy in the middle by a big soft toy teddy bear that she has just won in this maybe this theme park. Swap that out. So I'm leaving this to run in real time, just to show, in case you haven't been using these tools. And so here we had this image where I'm standing in Queenstown with some friends. This is the most advanced in terms of of image replacement. And I use voice you text it in or message it in or type it in.(poor audio) So here it is. Not only do you see the full image change, but you see my tilt of my head changes the tilt of the person in the middle. It's actually realigning the body position around this new object of a bigger teddybear. And so you can sit there and go as a creator. This is quite horrifying, this idea that years of Photoshop skills and actually having, you know, to really navigate these technical platforms. But actually we can also see it another way that actually now we can bring these two worlds together. You can bring the skills of a great photographer together with the ability to be able to utilize them in very different ways. And so we're augmenting our abilities constantly with these new tools. So with the augmentation with AI isn't just a creative endeavor. It's not just that we're going to disrupt the professional services or sectors that are already playing in the data space. But in 2016, I played this video. I’m Bryce and I want to talk about full driverless cars. will be made available to some lucky Californians to ride in on roads all over the state in just one year from now. If California is able to finalize the regulations by the end of the year, the car should go from testing to use by members of the public next summer. So by members of the public By 2017, we'll be able to go into driverless cars. Now we have entire car parks dedicated to driverless vehicles. And this is a Waymo One in San Francisco and 7000 autonomous cars on the road, and half a million rides every single week. Now we have to start understanding again this very steep exponential curve that happens. So we imagine it's not going to happen, It feels like 2017 is a few little cars driving around the city. Then you get to ten and 20 and people start talking about the novelty of going into a driverless car and then suddenly walk down the road in San Francisco and a Waymo is your first and second and third probably choice. One because of the number of cars, but secondly, because people prefer them. And then other countries pick it up and other cities grab, and then things like Robotaxi with Elon Musk and, you know, other programs and cars and countries. And then you once you get to a half a million of anything of subscriptions, a half a million, that means the number of subscribers to these is exponential as well. We can start to imagine how easily this will scale into other markets. So once you have AI enabled, you start to see the lens through a very different way. So of course we need to go back to film. So again, original screen from 2016 when I talked about machine learning, I wasn't talking about generative AI because generative AI was really not a sort of a subfield that anyone really had backed their hopes on. And actually, if we talked about creativity back then, we were still imagining that most creative outputs were still in that very precious little snowflake area, which we were all going to hang on to, that they would not be disrupted because the creative sector wasn't about zeros and ones, it wasn't about data, but how wrong we were. So I talked about it initially about how fast exponential growth of communities were happening. And this is to the graphic I used to showcase how fast it took to get to 100 million users in it, 75 years to the telephone, because you had to get lines in and you had to have a whole bunch things, like a telephone itself to sit on the desk and, and then getting down to this period where in two years you could get to 100 million people with Instagram. Now that was a record breaking number at the time. And communities over 100 million were pretty small, even just that small period of time ago. Now we look at ChatGPT and on the right hand side, all these ones are over a billion in their communities. You know, you think about the world being 8 billion and just a little bit. Then you have all these platforms now have grown to be over a billion. So of the 8 billion people on the planet, and ChatGPT comes in and actually grows to the point where it's getting there - It talks about a million subscribers in an hour. So we have grown these connections online, and eyeballs are increasingly targeting people's time through online streaming media, including social. So going back to where film was and this was in its time in 2016, the absolute pinnacle of AI. And so this film was talked about widely and talked about in festivals, but it was actually just written by an AI. It was no aspect of the creative output. There wasn't the filmmaking, the editing, the soundscape. The actors acted out verbatim the script. Now we'll play a little bit because it will show again how clumsy at the time AI still was. In a future with mass unemployment, young people are forced to sell blood. It’s the first thing I can do. You should see the boy and shut up. I was the one who was gonna be a hundred years old. I saw him again. The way you were sent to me That was a big, honest, idea. I am not a bright light. Well, I have to, ahh go to the skull. Right. Well, I don't know how long we'd watch that one for. So now we jump forward into this new space where creators, bringing together real footage with AI footage and real soundscapes with artificially generated soundscapes. And the internet is full of them. But here is one particular creative which I follow, and I'm really interested the way he brings these two worlds together. I thought you might also appreciate it. You’re gonna create the storyboard for the trailer. Here you can input your full script or general overview prompt for your story. this is the prompt I put in Then I’ll let this AI introduce a film trailer add in some strong keywords(poor audio) Once you’re happy click next This takes you to the storyline part which you can edit and give the overall description of your story And once that’s done it becomes this really fun part where you actually create all the characters for your story. So click on one of the characters to choose edit, I can then make any amends to refine how I see the character changing their age, appearance, clothing, etc. Once we’re happy, click apply. You can always get back to this part, add more characters if necessary. And the other part is the breakdown section. We can describe the location, what actually takes place in the scene, the characters that are involved. Once we’ve put this To do that we just print start button. Let’s try and get into the storyboard session. these are our past storyboards started to get a film trailer. If I zoom out you can see that each (inaudible) represents each scene And you can add as many shots as you want simply by clicking (inaudible). So one of the things that I’m really a big fan of when it comes to LTX studio trying to give you a lot of control over what you're trying to create. So we take a look at one of these shots you can see (inaudible) This here means (inaudible) Again, we can focus on characters here by typing the ‘@’ and then their name and then describing the scene and shots we wanted. We then have (inaudible) where we’re able to choose things like close up medium shot, wide shot, etc. there's also this create composition button, this will open up new page where you can actually draw a shot you’re looking for, to give the AI a better idea of the shot you’re looking to create. You can also offer a reference image as well. So once you’ve got all your prompts sorted you can actually regenerate the image if you’re not totally happy with it just by hitting the refresh button and add auto (inaudible) Also the (inaduible) tab if you click here it allows you to upscale the image generate an image, basically just like photoshop where you can paint in and afterward generate something in that place that you selected this. Also you want an object removed at all and do the opposite of that so that’s I think in ‘remove from image’ And if you also change the facial expressions, and camera angles as well. And a quick final thing to show you just before on the emotions then is a sign that here you can actually edit location of the scene adding things like time of day adding description, it’s also a style, where you can change the aesthetic of the actual storyboard, so changing it from a cinematic shot like this to more like drawing or a cartoon, clothing, the characters clothing, and sound where you can add voiceover and scene sounds. Now a finished storyboard step you can also add motions of all these images simply go into generate motion picture ask the AI to autogenerate motion for the scene alternatively, you can click on motion editor which will take you to a new page where you have a lot more control of the motion to try to get. All of that there is the storyboard which I created and on the right Are the shots from the trailer. So now we're gonna walk through the trailer And what shots were AI generated. And, I’m gonna show you some techniques that you can apply yourself. So the first three shots are all real footage Myself and Jenny went out to shoot at this location at 5AM. So that is actually a reason I started this trailer with real footage. It’s a technique called perceptual fluency. In simple terms, if I take what’s familiar to us our brains are much more likely to accept it as being real. If I start it with real footage where the lighting movement and texture are grounded We set the expectation, then when the AI shot appears your brain doesn’t question it because it expects the same consistency. Shot four, this is fullly AI generated, that’s me describing (inaudible) and especially using AI to generate this wide establishing shot. This is fully generated - Now I'll leave it there because there's so much more to it, but I just think it's worth going through and understanding this, this process of augmentation that's going on across all sectors. So just delving a little deeper into that one just shows this, this new work, their process and systems that people are building around it. And so we see it there and film and I'm going to jump back to there and again. But I'm just now popping across into the writing side. And the first time I ever heard about an AI writing was actually back in 2016 at the Rio Olympics, and they were talking about that. The China’s AI reporter published 450 articles that were written by an AI at the Olympic Games, and at that time it was actually literally Olympics were playing while I was standing on stage. And I was a little bit horrified. This idea that an AI was trusted to write, given that we had no history, there was no precedent at that point in time, apart from the likes of the movie we'd seen. So you can imagine my concern about, you know, Sun Spring and then suddenly talking about writing articles. Jump forward to the more recent Paris Olympics and just Intel alone created 100,000 AI videos, from, you know, from the actual performers, the players, the entrance, whatever you want to call them. And so we've we've changed in that period of time between two Olympics is exponential lift of what you could do with AI even within the sporting arena, within the writing journalism type of environment. We also jumped from sound and music in a very fast way. And this video that I played again back in 2016 was the first time that I'd ever heard of an AI creating music, and this one here was designed on the sounds. That was designed in the style of the Beatles. And at the time, it was almost possible. And I was thinking, yeah, this is me and this is okay. This is I can see where it's going. It's a little play. And then it gets a little bit off. Jump ahead. Now with multiple different music creation platforms. Entire channels within Spotify. There is now the number one artist on the country hits in the US right now is an AI band. We have definitely moved a long way ahead. And in fact, I don't even think we'd ever imagined those words coming together. Country and AI. I'm not sure that was even designed that way, but here's just an example in Eleven music with me putting a very simple prompt in and asking for an ironic song in a smoky jazz with a saxophone and a few things around the technology. So. So this is what you can now create in this, you know, period of 15 seconds.(Music plays) so you can see a pretty big advance in just those few years. And what you can do now in any genre, in any language. And you can add or whatever you want to instrumentals, you can move in shape and edit as you wish. We're now also seeing a whole, industry booming out of people creating for festivals, and this one here is for a music festival where you have to create the song in AI and the video to, to attach to it. So here is a submission done through a platform called Higgs Field of an AI generated video with an AI song. And you start to see again how these genres are starting to appear in mainstream platforms such as Spotify. Can we just turn it up a bit? Right. Better luck next time. Okay. So, so then we start thinking about where people see this opportunity. They start building their own platforms. And there are many platforms where if you're looking for music and sound effects and voices and narration, that they now enable you to do exactly that. You can drop in what you need, you can request it. And here is just an example of just one of those, where they have a huge range of music options that you can just kind of bring in all created again by AI. And I'm just going to play a few examples of this. As I was playing around with it for the first time. Just a couple of ranges of types of music here. Now, the reason I want to show you all these different styles and how people are using is because right now in New Zealand now AI adoption is extremely low. So we are at the moment there's a number of different research studies have been out, the most recent with KPMG and the University of Melbourne. We sit on the bottoms of 10 to 20% on adoption, AI learning and training, understanding and trust. And so where you start to see really massive adoption is in developing nations with younger populations, where you've got a vast majority of the population are under the age of 35, we're certainly not that. And so we starting to have this potential us and then an inequality of what people are using to be able to create it faster, cheaper, not necessarily better, but a new type of content that is really starting to be built out by a generation who are really happy to adopt in ways that perhaps as craftspeople, as people who’ve been around the block a little bit longer, we we so stand back and saying, it's not for me. And so we do need to start thinking about our own, you know, our own biases around what this is and what it could be for us and how we could actually amplify content from Aotearoa to the world faster. If we start thinking about some of those tools that maybe will help us enhance and do things quicker. So jumping back into where things were back then, you may remember this because this was around a lot the Soul Machines New Zealand company headed by Mark Sagar, an absolute legendary guy in his own right who really spearheaded a lot of the advances in this idea that humans could become digital. And so back then, he was starting to think about how you could have speaking AI avatars... can even express itself to you in a human-like way which is emotional it’s cognitive No. That’s where you’re saying - It’s a great Kiwi accent. So that was then. And so that took, you know, years and years of work to simulate, to build the data processes, the cognitive aspect of it, so that the baby or the person had the brain and the knowledge. And now you can do that yourself, something you can do in a very short period of time with a large language model, you build out with your content. So if you've got academic papers or YouTube videos or things that you've been, you know, maybe on screen for to capture your voice, then all you need is an image. That's it. And so here is for my digital avatar, Chances are, you've heard the word blockchain and instantly thought of Bitcoin bros sketchy crypto schemes. Or some startup trying to reinvent the wheel with web three. So you can see again how far that advanced - we were literally at the frontier of advancement nine years ago with Soul Machines and where we were. And now we've brought it down. democratized access to those sort of data. We're able to upload our own content, our own data pools, to large language models, and literally not even a video, just an image. And there you go. And it will sync the two. And you can question, you can interact with your own avatar. And of course you can do that with other people's avatars as well. Depending on which platform you use, I want to just show when you take away humans and put them out of the loop. What if AI just creates content? And we're seeing that particularly on social media, and I'm sure most of you heard the term ‘AI slop’, and AI slop is definitely a domineering The internet right now in social media, it's about 82% of all content on social media now is AI generated. So when we're thinking we're watching these influencers and people who are doing great things, none of this is real anymore. And what we need to understand is the good thing is, for the first time, social media is on a decline. So the very nature of social media in its infancy was you could invite your friends and you could find out what they were doing. And when you went to your your page and early Facebook, you knew what your friends feeds were doing because you could see them. What was there? They took away that over years, TikTok came in and completely obliterated it. And now we have infinite scroll across all of these platforms, which means you never get to see your friends. You get to see what the algorithm serves up, and it serves up increasingly content that's got no no particular interest you. But it's AI generated. And so the savvy young generation has started to go, this isn't social and this is not what I'm here for. So you start to see the diversifying across to more of the fringe platforms, and you start to see them in Discord and Substack and other areas away from these mainstream platforms. But if you look at the type of content that you now can find online, and you can start to see any day of the week, by generated entirely by AI, it's worth knowing what is possible.(inaudible) two years ago now this video came out of Will Smith eating spaghetti And like we all thought it was crazy look at this it’s still kinda mindblowing if you really think about it. But now something insane just came out look at, look at like watch this and then look at look at what came out. Is it actually? Look at this one these are fully, like you can turn this into a prompt And it makes this, it does all the sounds. It does all the visuals it does everything. Watch this. This is fully AI What’s one thing with AI Oh ya’ll gotta give them that This is wild! it's over. We are cooked on that thread. You get me? There's still some. You know, obviously some tonal things that but look at the jump that we made from 2023 to now. Just think about what 2026 will be like. I’ve saved a couple of these. There’s no way - is this... This is AI? Good. Yeah. Hi and welcome to the channel. Look at this. I’m AI trying to get right in your brain. Nothing is real anymore. Haunting, haunting sentence to hear(inaudible) Prompts. Like seriously, dude, you're saying the only thing standing between me and $1 billion is some random text? Honestly, the biggest red flag - That's why it's so true. They still can't get text. Unbelievable. I feel like that's the easiest thing. Like CapCut (inaudible) Why can’t AI do that? why is it that the thing that it struggles with? Oh my God when you use the Prompt theory. like really we came home from prompts? wake up man! You want to convince me Nah this is - this is freaking this I don’t like this! I don’t like this! It’s like very simulations (inaudible) And someone is this is our whole world is going on the result of ones and zeros of binary code and nothing more. It makes no sense. So there you go. That's where we're at in this new kind of world. So we also saw things, and again I touched on the idea of AI and law. They got lots of people's feathers ruffled back then. And of course that is very much the new element of the law. Scarcity issue has been completely obviated. And now we have lawyers that can be completely AI for a lot big chunks of the profession. And so here is an example of the types of products that are coming out in the law sector. We are using a contract matrix internally across loads of different contract types in lots of different (inaudible) It has been developed in collaboration with Harvey and Microsoft. There are three main functions of the contract matrix. The first is open ended question. You ask the question of the provisions you're drafting or reviewing So it’s a ribbon at the top of the page and you click on it to open up(inaudible) directly in Word you choose the setup provision and gonna apply AI to it Harvey is being channeled over that particular loan setup provision and is responding against it. So we won't go through all three provisions. But you know it's a big chunk of that contract drafting. It's really taken over. And here they were talking about Harvey. And here's just an example of Harvey. Now Harvey is in 34 of the top 100 law firms in the world right now. And the adoption again seeing the exponential curve up. So when these law firms actually understand that this is an advantage for them, they're, you know, embracing and bring it in the right place. They able to tell their customers and clients, I'm using AI to keep the cost down for discovery and for some of the easy drafting. And you're going to pay me my $800 an hour for the things that really matter. And I think this is the world that we started to step into, where we have to be upfront about what we're using it for and when, and then also talking about where we bring in humans and to make sure it's correct. There's no misinformation that all the, the, the data points are all correct. We also talked about Skype now, not just any old Skype, but Skype with a written translation. So you could speak back then and give it a few moments. It would have a little captions coming up in another language. And this was blowing everyone's mind at the time, this idea. So this is what it was. Imagine being able to speak in German and have your message conveyed grammatically and somatically correct in English that future is here. Doesn't this look so old fashioned? It all starts just as with any other Skype call. You just call someone. But now the difference is the person you're calling doesn't have to speak your language.(inaudible) So again, jump forward. Now the AI can translate and it can do that. And now it's actually almost every language in the world. The latest count is 3000 languages. It was about 180 about two years ago. So this is going down now to these very small, almost dying languages and some of these very small pockets of the world. So now, just a new if you, if you have an apple and you've got your little pods in your ears, if you've both got them on the new version, you can do this, I agree. Yeah. Let's look at - Spanish translation(different languages spoken) Definitely the clients will love that. So if you don't have both of you receiving, you can obviously have it. It will go to your phone and you can read it. So someone saying it. So so if you're if you're speaking they'll hear and then they can type back to you. But this you know, this advancement again as you start to see these extreme jumps ahead the same way that you can do translations live in any language. I wish as much as I could to speak Hindi, I don't, but you'd never know.(Frances’ avatar speaks Hindi) There you go. We also talked about in 2025, this was a screen I heard back in 16 was this year, our relationship with search engines will be like that of a human assistant who can understand complicated ideas, but has read billions of pages. Good to hear from you Charlie How are you going? I’m doing great thanks. It’s always awesome to hear from you. How's everything going on your end? Pretty good. Just wondering how the kids are, is it school holidays yet? Yeah, they're about to start their school holidays next week, so ah, they’re really excited about that. Um, how about over there, are the kids on break yet? No, a few weeks away yet. so. I don't have a brother, Charlie. I don't have little kids. And, Yeah. So this is my AI, and so when we talked about conversational AI back then, this is really what we were imagining, but not quite this far advanced. So what's amazing about that? When I'm talking to Charlie, he starts talking about tents and summer time or what the kids want for Christmas presents is that's all improv. I'm not asking. I didn't set it up. That is literally just freeling, free Kind of styling a conversation with an AI companion bot. So the companion bots now come in many different forms, but the latest estimate as of last month is somewhere between 1.6 and 1.8 billion people using them on a regular basis. Now, some of those are just friendship, some are advice, some are a guidance, some a coaching, some are sexual. There's all sorts, every flavor. But actually and some of them are visual. So there are ones like if you go in to Grok and you start to see the kind of content and Annie which is a highly sexualized visual anime styled character, but most of it is just this conversation. So you might be sitting in your car having a long drive, and your conversation might be about things you need to do. It might be just general knowledge quizzes to keep your mind busy while you're driving. It might be asking it to freestyle a story about your family, whatever it might be. But this interaction now is becoming very mainstream very quickly. The first of the chat bots. We've only been out for a year, and so now almost every platform has one. And so ChatGPT is certainly the one that most people are using, which is you just see there that there is 800 million people using it weekly, which is a pretty big chunk of people. So the new players in the block I want to talk about, because back in the day, it was all about the tech companies starting to take the place of all the oil giants that now is shifted. And what we're seeing here is how long did it take these new emerging companies to become unicorns? So a unicorn is $1 billion US company with a valuation of $1 billion. And you'll see here at the top Cognition AI, which I'm going to assume none of you have heard of because I hadn't, it took them just five months so a company could go to five, five months later this year. By the way, we've had two billion-companies form with less than 12 people. So you’re starting to see where they're fully AI. But then you start looking across all of these top ten, most of which you won't know. You might know Anthropic, for example, the Stability AI XAI of course you all know, but some of these are becoming billion dollar companies within months of launching. Now this is so unprecedented. You know, if we had said that these is to get you$1 million company, you'd be impressed ten years ago if I said, you know, five months after I opened the doors, you became $1 million company. We'd all be going through this pretty good. But now this whole parameter has changed again. And so this is this growth and the investment that’s flowing. 92% of all investment in the US this year has gone into AI enabled businesses. So if you are not in the AI enabled business, there isn't funding to be found. I'm just going to use the example of Anthropic. So Anthropic here, what it shows is when they went to the first billion dollars, it took 21 months. And then can't see with my glasses. It took another three months to become a $2 billion company and another two months to become a $3 billion company. There we go. You can see it now. So this so once they get going, the ability to scale is really tremendous. But just thinking about this. So this is now a $3 billion company. But actually not long ago it got a Series F funding on a valuation of 183 billion. So the period between a $3 billion company and $183, a $183 billion company, is just a couple of years. And so when we thinking about what does the economy look like, what does financing look like? Where is money flowing? These are just an era we've never seen before in the big tech giants in the US, particularly right now. Salaries for AI researchers A year ago were 1 million USD. Then they went to 10 million USD and the most recent ones the most controversial hires. Instead of buying companies, they're buying the best talent from companies and they're offering them $100 million salaries. So again, back in 2016, we were talking about Amazon coming in and eating Netflix lunch and talking about the amount of money that they were putting into this new era of streaming. Now, what is really interesting is that did continue. Amazon did continue putting money in and others as well. But this graph really shows it. So you've got the date back at the orange back in 2016, and then you've got the blue date of actually how many subscribers they have. So you'll see that Netflix continue to Amazon is also growing along with Disney and a whole bunch of others. And there is you know, some of them are actually new on the block. So Disney didn't exist back then in a streaming format. And you'll see how many new entrants there have been. So we are again just pulling eyeballs away across multiple different streaming platforms. We also talked about Facebook coming in, and actually it was the first year of people being able to create video on Facebook, and they were saying that it would be really interesting. Maybe all video in five years might be kind of the focus of Facebook, which of course now is across multiple different channels. And so they were talking about this is going to be bigger than anything else. Before back in 2016. By 2021, they stopped reporting on their growth because actually they weren't growing. So the slow demise of Facebook and the way that they started to disappear as other platforms Snapchat, TikTok, Instagram came in. And so for a moment in time, they set the new standard and then they disappeared. So it's good to understand then, how many videos are actually now uploaded on YouTube a day, because that's certainly the place where most eyeballs are going over all of them. And so they have 20 million video out, uploads a day just to get an idea. So a 38% increase from 2024. So you've got the 5.1 billion videos serving 2.5 billion videos monthly active users. And so if you look at moment of creators uploading 500 hours of new content every minute, every minute, that's equivalent to 82 years worth of content. So this is really the big shift. So we've gone from what we thought was Facebook was going to own this domain. It's now YouTube really owning it. And you can see the India leads the global uploads with, with 21% in the US following 18. It's a very much a diversified and distributed uploading across the world. And then TikTok came in with 34 million video pauses posted every day on TikTok. So this was the new biggest new entrant since I last spoke to you. And now is the fifth most popular social media platform in the world. And of course, it takes a global look across other markets where they had their own different platforms. So all of these are starting to fragment again across into this content. But it does mean that more people can create content because of AI, and of course, more people are creators that don't have the significant outlay of hardware and software that they would have had to have before. And so if you're going back to where these markets where the most content is coming from, you're starting to see it again. As I mentioned, that top of the talk was these are emerging economies. So the biggest, users of AI in the world is Nigeria, also the youngest population in the world and also one of the most skilled. So that they have one of the highest participations in higher education in the world. So you have these pockets like, India and China and Turkey and UAE where you have younger populations that are outperforming the developed world aging populations. And of course, you start to see that across the way that they are populating these platforms. But they are moving into their own platforms and increasingly becoming creators in their own right. So just finally, before I wrap up the questions is to some of these just kind of general stats as of last month. So AI companies worldwide nearly doubled since 2017. 25% only of all these AI companies are in the US. So we do take this very US centric view of the world and actually right across the world. If you look across the top 20 or 30 AI companies, very few in the US, it's just that OpenAI has a very big dominant footprint. And so it is good to understand that wherever you go in the world, there's going to be other, other organizations and companies doing great things. We talked about the AI funding. It's reached $103 billion last year. It's increased significantly on that this year. Google search is really changing. And in fact, most of you will be aware that actually, SEO is diminishing all the time. It's very, very hard to get Adwords and SEO on your on your websites and to get your eyeballs onto anything because they introduced their AI option at the top of a search page on on Google. And I'm sure you'll be aware that most of you use it now over going scrolling through what is listed below. And that's really changed the model of Google. And so about 65% of all searches stop at the AI piece at the top now, and people don't go any further. So they sort of ate their own lunch. But I would say if I was a betting person that Google will become one of the dominant players in the next ten years. So you have to ask me back and see if that came true. And then just the number of people using and looking for, you know, Gen AI tool videos got a 1.7 billion views just last year. So people are looking and learning, and this has become almost the default place of how to understand AI for a whole generation who are YouTube natives. And so that learning on these platforms, they are becoming teachers. Many of them had their own AI avatars. They're not even the real person. They've literally they start their videos. I would say probably five and ten videos I watch from these producers saying, hi, I'm Dave's AI avatar, this is my weekly update. This is what I'm doing. I'm going to show you through a process. And so this the tile content has been sort of self-teaching a generation who are coming through with completely different way of imagining the future of film. So I will end there, but happy. If there's time, I'll get some indication if I've got time for maybe a question or two. So I do appreciate it's a lot to take in. And a lot of information and a lot of change over certain small period of time. I guess the opportunity for us here in New Zealand is we don't have the data that some of these big companies and countries have. In fact, if we took all of our data in this country, including all government data, all private company data, we put all of it into one large language model. It wouldn't even be 10% of the some of these models that we're using on a daily basis, like ChatGPT or Copilot. So we're not going to play there. We could play on trust. You know, we naturally lean towards a country where we could look at how we use AI safely. When you think about data sovereignty, you think about the wise use that we could have, but also how could we use it creatively to tell our stories better and actually have amplify, use the channels it has to send more content into the world and making sure that we have people within all that organizations can help steer us in the right way as we think about this new future. Because there's the flip side of that. As we continue to do more of the same and our stories get drowned out and we see less of our content, and actually we become invisible to the world and the next era. People will have no connection between film and content in Aotearoa New Zealand. So(applause) Now I’m sure - is there time for I can't see anyone, but is there time for a quick question if possible? over here you might to yell very loud. Oh, I was wondering, you know, just in terms of, you know, 25 years ago we had the dotcom bubble that burst. All these companies that were had these huge valuations but couldn't really figure out how to generate revenue, or profit. And we have the same thing with OpenAI now, you know, these companies that have these huge valuations but can't really figure out how to actually turn that into a consistent profit, how to actually generate revenue from their customers. So what are the potential, maybe risks of investing in AI if this bubble were to burst? Lots of good layers of question there. So the first one is in the dotcom bubble They were promising what the internet would be and not what it could do. So it was all about the promise. Now, it did take a while and a lot of people lost money, but now we have the internet, so it actually delivered on what it said. It just took a while to get there. And of course, if it didn't have that and people didn't put that vast amount of capital in at that point, the internet as it is today would probably be quite different. So there were winners that came After the after the big the big bubble burst? I think the difference here is AI is over delivering what it said it can do. It's not saying the future will do this. It's literally coming out and companies saying, look what we can do and you should give us some money. Yes. They're everything is overvalued at the moment. Absolutely. There are going to be massive losers. And if you if you are dabbling at all in the sharemarket, you'd want to be picking your companies very, very carefully. And certainly this is not financial advice, by the way. But this is the time when these valuations will consolidate those who are truly going to play in the long haul, and they will work out how to make money. And I think it's a really interesting example I just use with Google. Google has made so much money over the years through ‘Search’ that they literally stood up one day and said, you know what? It's a sacred cow, but we need to kill it off and we need to do something in AI and we need to replace it. So they've done that to the detriment of everybody who has SEO, and people are starting to move away from it because they can't get the rankings anymore. But clearly they have a model going forward. It's a it's a domination model. And I think a lot of this is domination. I think some of the big ones will get bigger. The likes of Nvidia, clearly when they've got the GPUs, until such time, there's huge competition. They'll continue to value. They will start to devalue once they're, you know, lots of other competitors in the market. But I don't think we should sort of do an overlay between these two scenarios. And the investment side right now is, you know, I have so many AI subscriptions every month, but I only ever sign up for a month at a time. So even if they do great deals for a year, I do one month. And then at the end of the month I go. So I sign up and then literally immediately unsubscribe. So don't forget to. And so then I have a month's worth before I forget to unsubscribe. And then I go, how many times in that month did I use that tool? Now most of them have free versions to a certain point, and even when you are paying for a month, the most I've ever paid for A month for The very best tools is $80 USD, so it's not like it's investment on the days of, you know, having to bring in some really serious processes and speed that you needed on computers and hardware and software. This is quite a different era. So at the moment humans are winning because we are getting a lot of these tools which companies are losing money on and we're getting to dabble at their cost. Eventually Those who are the tools that are so good, we will start paying much heavier, for there's no way that they're going to continue to be this cheap. So I think we're in a sort of era right now where we are the testers for many of these companies, and they're all trying to work out how humans are going to use these tools. And then the other side of that, we'll start to settle down. Which companies are the winners, and then how are people using it, and how are they making money out of it and justifying it to their board and their investors? So it was sort of a weird period. I think the next ten years, if I was to come back in ten years time, I think they are going to be the probably the messiest ten years that humans will have in the last 100. And because we are literally running in sort of two different speeds, you know, we've got those who are kind of rapidly adopting and adapting and those who are sitting back and saying, I'm just going to wait and see. And I think the inequality over this period of time could be quite profound. Particularly as we start thinking about what this nation looks like in 2050. And just on that basis, there is about four current New Zealand reports out with a name in the title 2050. So New Zealand 2050 is in the title of four reports, and they all paint a pretty grim picture of this country. If we don't do some pretty significant digitization for productivity gains. And of course, a lot of it is the technologies we need to augment what we have here. If we're going to compete in the markets that we already play. So long answer. But hopefully that answer answers. That okay. Thank you everyone. Brilliant.