Everyone’s Using AI — So Why Is the Content Still Mid?

[music] All right, Laura, take us through your uh AI is eating the world.
Actually, my title was presentation.
I summarized it. So, we're not going to go through every slide even though I did think about that. I'm just kidding. So, it's vibe coding and AI generated assets are the next platform. So RIP ARVR what's growing from a dashes is generative AI. This is Benedict Evans. So he does I think a by-early presentation of one of his predictions of and the assessments of the market and he did one on generative AI and I was like oh okay this is long. So I started going through and I was like actually I think he has there's some takes in this and one bullet point I pulled was ChachiPT which I learned is at 800 million weekly active users. So what does this mean? So one, everyone's using it. So where are these eyes going? Well, if they're not going to mo well, so they're still on mobile, but then if

they're not going to AR, VR, what are people doing? Well, a lot of them are using AI tools. And then how does that apply to gaming? So one of the one of the perspectives he has is gaming and gaming content will have a new look as vibe coding grows and AI generated assets improve. We know I mean we know just being in the industry, they're not perfect right now. you want to make assets for your game. You can use AI to help you, but it I think right now is a struggle, at least from my perspective and what I've used in the past. It can't go from start point to end point in something that you can necessarily ship at the at a very high polished quality doing exactly what you want it to do. But this also means a couple things. One, we look at the new games coming out. They don't need that necessarily that high quality the quality of polish. And something really stuck with me, which is this is going to open up a new wave of hobbyists to be able to find kind of this path towards professional or path towards releasing content for us without needing the same skills that we

needed growing up to start. Like I remember this stuff doesn't didn't exist. If I wanted to make something, I'd have to draw it myself. And let me tell you, I'm a pretty shitty artist. And then looking at kind of the overall like we look at the younger generations, they're the ones that are going to be playing games. There's been a lot of studies on Gen Z and alpha shortened attention spans. I mean, if we look at Australia's ban on social media for youngans, which probably came a little too late, it's a growing concern. And when you're a part of that same gener, you're making content for your generation. I believe these users will make better content suited for their peers, potentially better than the current incumbents that are in place. There's two sides that I wanted to call it on this and the biggest issu is issue I say this slightly tongue and cheek is that I use chat GBT so in this example I will use that when I've interacted with it and I got to say I love copilot too co-pilot's a little bit different but chat GBT kind of is a very big one it is not as mean or as honest enough to give feedback that will help mold a good game

so pitfall one and assuming we fix that problem so let's say chatbt grows a backbone is actually giving honest feedback. The issue is will there be an acceptance of critical feedback by this tool? And so for anyone using this and this will shape how this platform kind of grows, is AI viewed as a low-level collaborator or an expert? So if AI and we when these you know for gaming tools is viewed as this low-level collaborator, they're viewed more as like a peeon in the creative process, then the more positive suggested feedback might be preferred by the creator. So, if you think that this is like some low-level person giving you feedback, a lot of time you want praise. And I see this in the hierarchy of most companies anyway, you're going to be getting, oh, this is so great. Of course, boss, you're doing a good job. And then you're viewed as the subject matter expert regardless of whether you should be evaluated as such. Truly, if the AI on the flip side is viewed as the creator or the expert, then that person

that's using AI is going to be heavily relying on the AI to be the expert and be creating things for players. Whoever is doing the creation bit will be taking this as gospel. So, normally would this be bad? I'm not sure. But the only place that I think draws a question is in every little like subtext when you use any sort of tool, it's like AI could be inaccurate. they might generate accurate results. So we're opening up a little bit to if the person that's using this tool has decided, ah AI, you're the person that knows what to do. It's going to be heavily dependent on whatever inputs that it's using to create the suggestions as this gospel of truth. And in many cases it's not. Maybe it will be, but right now it's not. I mean I'm a big Tik Tok lover as you know. half the content, that's probably an exaggeration, but much of the content is AI generated and we don't know what side I think whenever someone's making content they're sitting on. Are they using this as this is what it should be or is this something that's just kind of

boosting my ego for whatever I'm putting out. So effectively I put came to the conclusion we need different creator modes in AI or AI pre-prompting that sets the style and the type of collaboration environment to make the right type of content that we that audiences will find engaging and good. And we're not quite there yet but I do think this is where I'm bullish. I do think we will get there because of this high number of usage, the amount of inputs and the amount of time this is going to be able to iterate and get better. I think it's going to be moving much faster than VR ever was able to. Any other thoughts on my little hype train for AI?
Can I just say one thing? First of all,
Benedict Dick Evans doesn't know [ __ ] all about interactive, right, or video games. Like this guy is a professional level.
He is B he is Matthew Ball of AI basically. He's aing [laughter] grifter
and he worked at A16Z. So that all that's that's all you need to know. And so like [laughter] his whole analysis is [ __ ] [ __ ] right? He doesn't know what the [ __ ] he's talking about. And

again, I'm gonna say this a million times. I believe in AI. I think it is the next tool that's going to be revolutionizing our economy. That there's all kinds of benefits AI. But creating content, like compelling content for the consumer, again, this ignores the entire whole point of what makes content great is the consumer appetite for it, right? And so, like, yes, it could be good. Even the Tik Tok [ __ ] is crap. But building like what consumers want from a AAA production value, movie, television, and video games, it's gonna be really almost impossible. So
slow.
I don't think we should be listening to Benedict Evans as we should not have been listening to Matthew Ball [laughter] about what's the future of video games
period. Let me come in somewhere in between [ __ ] [ __ ] [laughter] and and and the beautiful words that Lara said. So it's uh almost 100 pages 100 graphs in this presentation and I assume it would have been better as a presentation because when I was going through it I felt that I'm missing the

context. It was just a graph after graph with like oneliner that didn't really
Yeah, I think he presented it slush because I'm like AI is great but like do I need 100 slides telling me that AI is a great tool?
You know that I already know it's a great tool for anyway. Go ahead.
Yeah. So, so TLDDR capex investments are getting out of control. According to this presentation, what I took away there's limitations to AI growth in form of data centers power just to power those data centers chips clone of LLMs and just raw funds money to make this happen. And he talked about LMS having a risk of becoming commodities and also pointing out that the penetration is very high. Then the big thesis was big corpse bundle while startups unbundle. So basically whenever you open up like claude or athropic has launched a new tool that is part of this big chatbt health and like there's just like endless amount of products Google the same thing and then there's startups that are focusing on niche and either

getting evaporated by these big LMS pushing in a product that essentially makes them obsolete or finding a profitable niche that they can go through grow their ARRS as they like to put it while being you know live for 8 months or 3 months and get acquired by one of these big companies and The final slide of this presentation like spoiler alert it ends with when automation works it disappears and it shows a graph of elevator operators kind of like how many there were and at one point there were many of them and then there were very few. So I'm clearly missing the context that might have been a banger ending but for going through all these slides I was like this is it anyways. So in short, the way I've been kind of like if I would have to be very hyperbolic, there's like a two short-term futures that happen from AI. And one is everybody who is super bullish on AI is correct. It's going to revolutionize everything in an incredibly fast speed.

In one year, humans don't need to go to work. There's no more tasks. yada yada yada. AI overtakes everything. Rapid pace causes mass unemployment and a lot of trouble in the economies. Then there's the other one which is the opposite case is that these promises are inflated and the current market as we know is prompted by a handful of tech companies which are relying heavily on AI for their valuation. The valuations get deflated and then that causes mass unemployment. So you know those are two views of what's happening next. But I wanted to educate myself more because I saw this in the documents and I was like, "Shit, I need to say something smart." So I watched two interviews. One was with Jensen Hang, the Nvidia CEO, and I took a couple of quotes from him. He said, "What AI will do is make our tasks that we do in our job more efficient." He sees that our job is more meaningful than typing on a keyboard or wrangling spreadsheets. And the point

that Jensen was making in his video interview was that AI will drive productivity and productivity leads to more revenue and more revenue will lead to more hiring at the companies which in turn ultimately results in the growth of the overall world GDP because it allows a larger portion of society of people to contribute to GDP growth by powering these people with AI. And I was like, "Oh, that's a very positive view of the future and the impact of AI." And even Trump tweeted that there will be more jobs because of AI. So I'm like, "Okay, Trump and Jensen are on the same page. Let me find another page that has a constructive criticism towards this." And I found Greek economist Janis Varufakis who is the former minister of economy for Greece, which says a lot and is actually one of most deep faked person on the internet. And so in his interview, he's an economist. It was

very interesting because he pointed out right away it's not the question of is AI useful or not. Of course it is and technology is amazing. But his point of view and his perspective was more about who owns AI because when people engage with AI, the AI or the machine pretends to be our servant. whether it's Siri, whether it's Alexa, whether it's the Instacard. And what's really happening according to Yiannis is that the machine trains you to train it even better to know you so well that it wins your trust. And once the machine has your trust, it can direct your choices. So the next time you're making a purchase decision, you will trust what the machine recommends you without comparing it on your own to what exists in the market. And that means that the owner of the machine or the AI will make a profit on having you to buy what the machine has recommended to you. Sounds like social media. And that profit is essentially what he sees as an

economist, a rent on the economy in terms of price inflation. Because that price increase by the machine or you know the Alexa selling you whatever Amazon wants you to sell. It's not the money that goes to companies that deliver to you or the companies that design the product or the companies that manufacture the product but the money goes to a technology company that is recommending you the product and there are examples of this happening. So according to Consumer Reports and Ground Work Collaborative, they revealed that Instacart, which is probably the biggest, it's owned by Amazon, but it's the biggest grocery delivery company in the world. When they used AI technology, and they did price testing, they saw that customers paid significantly higher prices, up to 23% more for the exact same items at the same store, leading to increased cost for some shoppers. So they trusted this platform, and the platform increased their costs. And in the end kind of like AI is not machines that produce or increase productivity,

but they are machines that determine how we think with the goal of according to Yiannis extracting rent on the economy and in such actually extracting from GDP and not growing it. So that was uh what the [ __ ] are you talking about? Dude, do you realize that Greece Greece's entire economic output is less than what Google makes in a year? Like why are we listening to this guy?
Because he was also an economist on CCP. So that's back to games CCP. He was in the Russia.
No, no, no. The uh the gaming company example is
online
that's not even having to do with AI. That's what airlines anyway. Anyway, Chris, let's why don't we wrap up on AI and gaming because I do think this is interesting to the audience.
Talk to us about the Larian topic. All right, let me let me let me just let me AI is a tool like it makes part of the process more efficient and may ultimately reduce costs in game development like in theory, right? But it still needs design. There's many many

other parts to this process and I think like all these takes are just very simplistic in their execution, right? And I'm not saying that AI is not revolutionary in its own right, but for game development, it is just part of the process, right? And so there was an article out about the darlings of the industry, Larian, who made Balders's Gate 3. And the original controversy came from when Schwven Vinky, who I think he's the CEO of Larian, he was in an interview with our friend Shrier at Bloomberg talking about AI tools. And they were talking about using AI tools to produce their next game called Divinity. and they were using generating like concept art, producing powerpoints and other early production work of the goddamn game to kind of speed up the prototyping phase and to get to some agreement on the visual direction of the game, right? And so he talked about this interview and it was a very small part of this overall interview evidently and the backlash wasing immediate and it was

primarily Reddit, social media, some former lawyer employees like you know bombed them a little bit and some fans and the but the honest truth was that the coverage out there was actually relatively balanced, right? I mean, in the interview anyway, Shrier did show his clear bias in the interview against using AI tools, but nonetheless, I mean, it was relatively fair in terms of the coverage, but what happened was immediately he rescended that and said that he they're not going to be using tools for this purpose anymore. And it's like,
dude,
just stop cowtowing to the mob, right? It's your job to experiment with new tech and use it to the benefit of process of making [ __ ] games, right? And great content. AI tools are out there. Like you can't like put that [ __ ] back in a box, you know? Dev, should you be using this to help drive efficiencies? That's what makes and it will make games better. And this could be at the expense of some employees. Like I'm sorry, but that's the way it works. Like this is the nature of

enabling technologies, whether it's AI or other things, you know. And so what you do is you displace some people and reorg around other skills that ma make the process more efficient. I do think AI will be a boon for game development and you cannot limit your usage because some idiot on Reddit and your fans who are 100% going to buy your next game anyway, right? Are bitching and moaning about the use of technology of these type of technologies. So I think this quote was like really good and it was one of the Reddit type comments was developers should use whatever tools are available to make the best game possible. There's no reason to cater to artists as though their work is uniquely sacred simply because it produces something visually appealing by that logic. Should studios also abandon software that makes programmers more efficient and reduces the number of coders needed? Of course not. Embrace the technology and let people adapt. Period. End of sentence. Stop apologizing for running your studio and do what makes sense regardless of what these idiots are talking about on the internet. Right. That's my kind of
or just don't talk about that and you

no talk about it. You should you have the freedom to talk about whatever you want.
Well, sure. I think the challenge is in, you know, these weird rules that are getting set up on Steam about AI usage or in the award show. Remember, I think Clare Obscure had to have an award taken away from them
because they had used like AI for like two seconds on something. So part of this is we need to make sure that we don't have rules or guidelines set up that prohibits Chris. Exactly. I agree a thousand%. I use AI a thousand times every day and it has made me more efficient. It has made me smarter. It's made me better. I just can't imagine in the old days you'd get a bunch of people sitting around the table and brainstorming, you know, to come up with an idea. And I feel like that's what AI allows me to do. It's it replicates the idea of a brainstorm partner and a thought partner. Anyway, I think everyone is being too black and white on this topic just like on all these like we just need to chill out and have

better guidelines is my humble opinion.
The comments from the Larry and co-founder were weird because like two sentences stood out when I read this article. The one was quoting, "We decided to refrain from using Gen AI tools during concept art development." And most of the listeners know what concept art is, but it's in gaming, it's a visual research and development phase of games look and feel. It's the process to explore, define, communicate what the game should look like before the team actually commits to expensive production such as 3D modeling, animation, VFX, UA, marketing renders, so forth. In other words, this is throwaway. Larry is not willing to use AI because of Reddit comments in a throwaway asset creation that helps them to ideulate faster. Like what is it? It's like, you know, like saving the pens from your meeting room. It's like we can't throw them away because that's that's against the nature. But anyways, he kind of also goes against what he said because the

next quote is like the more iterations we can do, the better in general the gameplay is. Yeah, we think that Gemini can help with this and we're trying things out across departments. I was like, this is in the same piece. So, so it doesn't make any sense what he said that they won't be using. And then he's saying it really helps us to move faster to come up with ideas and test them so that we can make better games. Literally, what he should say is like, [ __ ] off. We're going to use whatever [laughter] tools are at our disposal to make the best game possible based on our design. Period. End of sentence. So Chris probably not the best PR consultant you can find but if you mince your word and have that point there that would be good but I just wanted to describe like how most of companies like the beloved developers are using AI in asset creation in graphics for example so the first step is you train your custom models by uploading your database into a selected tool whatever it is let's say uh scenario and once the

custom models are ready so from your previous library you generate across different multiple media types. So you can do images, you can do video, you can do 3D, you can do skybox, you can do textures, but it already has your library. So this is of course for companies that have that one. Then once you have uploaded and once you're able to do these custom versions, you generate output that can be edited inside that chosen tool with either prompts. So you can say change this, change that. You can do retouches for like fixing specific regions in the images or the videos. You could do outpainting where you just expand beyond the original bounds. And you can do enhancing where you improve resolution, detail and so forth. And everything is then stored indexed and searchable inside this tool and accessible only by your organization. It makes sense because your data remains private by default with these tools. The model is customtrained. It's generated content securely contained in the workspace that you have and isolated by a project. So these tools make ton of sense and I

don't understand kind of going through like what these tools are. It's like why can't you use it? Dude, you can't assume everyone's using it in this way. I mean this is just one execution of how you use AI tools. So like they could be creating it from the ether and you know and models that are out there like
we're talking about for companies this size that have their own IP that have their own style. This is how
one way though. Yeah. Chris, I mean, there's so many different This is even just creative. Like, what about, you know, co-pilot on coding or pipeline? There's so many other ways.
The fact is that you should be able to do whatever the [ __ ] you want to do, right? Unless it breaks the rules and breaks laws, right? And the only point that they make that actually is somewhat valid is that if you're using art that is out there that the models are trained on to build your own visual style, like there's some issue there potentially, but that needs to be resolved by the courts, right? Not by like some guy on Reddit and bitching to Larry CEO, you know?
Anything else to wrap up our massive

amount of AI talk today? [laughter]
I think I think we're good. I think we're we're going to cut some portion of it to make it lighter for for for [laughter]
for the listeners to go through.
All right, Laura, lovely to see you. See you in a month. Uh see the rest of you next week. Thanks everyone.
See you. See you in Cabo. [laughter] Not not in real life.
I'll be there.
Not in We'll be with you on your Instagram. [laughter]
Call me, man. Bye. [music]
[music]