AI Games Need Soul, Not Slop

Growing up on a farm, you know, knowing all about how to lamb sheep and drive tractors, when I played Undertale at 15, was like, if one person can make this, you know, I can join a team and make something incredible, surely. Welcome. It is the PocketGamer.biz podcast. I'm here, Peggy Anne Salz, your host. But, Craig Chapple, my co-host, just wonderful to have you back. It's been a while since we did this at GDC. I'm doing very well, thank you.
I'm particularly excited about our guest today because well, it's like bigger, better, faster. But, what if the real excitement about AI is not necessarily the biggest model? We have Ambrose Robinson, founder of Parable Studios and the creator of Millennium Whisper. We'll get to that. Great game, first on-device AI game released on Steam, and more to come. Um mission is really simple, right? You don't worry about what you don't own. Everyone

gets what they're owed. He's building on AI tools and games around ethical data capture, actor-led performance, energy efficiency. I'm just so excited to have you, Ambrose, because I don't think I've ever had anyone here who's going to talk about these topics and also ethics. So, that's that's a great mix. Great to have you. Hi. Yeah, yeah, it's great to be on here. I mean, I've already had a really great interview with, you know, in the PocketGamer space. Yes.
Um with Dave Bradley, uh and you know, he's really hot on all this space. So, he was asking all the right questions. Yes.
Um but yeah, it's it's great to be here. Thanks for having
That's true. We proudly pick up where he left off, Craig. This is a different type of interview as well, cuz we're looking into what's going to inspire our audience of Indies, game devs, to go this path. And I think there's so much like potential there. The types of games that there are hundreds of because you really want what will make you stand out, and I think there's a lot of easy

answers to that question with AI, and we've seen them, and we've seen them come, and we've seen them very quickly go, and I think um really the one the conversations that I have are very very creative. That really excites me about um the types of like uh experiences that you see in cinema and TV shows, like core dramas and murder investigations and stuff like this that aren't as often depicted in games, you know, moving away from games about uh you know open world action survival crafting and more towards We started with romance, which I think is a really fun one. Wow. Yes. Do you think that with with some of the AI tech that's being kind of developed now, there's not really any creatives involved, and does that then show through when developers, like yourselves, like people you've been chatting with,

they're trying to engage with these tools? Does that actually make an impact? It's very rudimentary tech, this 3D tech, back in like the late '90s, and they'd done it for kind of like tech demos and things that kind of looked like something that they they really pushed and made something really creative um with Toy Story, and that is now something that even though the tech at the time was so rudimentary, people returned to over and over again over the years, and I think it's because, you know, you can really see the like creativity and the heart of the people behind it. The people returned to games on the NES and the SNES and the original PlayStation because sometimes they've done the best versions, or some of the best versions, of that type of game, that type of expression. And I think going back to the question about like, you know, there's no creatives involved in some of these AI projects. I My

personal kind of devil is like image generation because I think it does so much bad blood for people like me who are trying to do stuff kind of Mhm. It we're trying to make something creative and make and people are involved. There are writers, there are actors. And when people have this thing that's so obviously um got no creatives involved and is just stealing and [clears throat] then repackaging that stealing into something new. Um it makes that what the first thing that people think when they think of AI. And And with um world models with stuff like Genie 3, a bunch of investors pulled out their money because they thought, "Oh, we don't have to make video games anymore. We can just generate them on the fly." Until they realized that actually, you know, it costs about $20 a minute and um it's like completely

surface level. It only lasts for like a minute, 2 minutes at most. Um you know, where you start to see where that sort of perspective has its cutoff point. It has a point where it's just not really feasible. Um Yeah. What the interesting about Genie 3 is apparently at GDC, Google, they had a session there. But they were very open about actually it's nothing to do with games. Um they're building is for robotics. So in the real world, they will be able to navigate Yeah. the real world. Which I thought was interesting. I Like you said, after all these investors were like, "Oh my god. Uh poor Unity, for example." And all these publishers kind of had their share price go down. It's nothing to do with games. Yeah, yeah, yeah. But still, then why were they at GDC with it, I guess? I I guess they want to take those developers. You see a lot of non-gaming companies sort of go there now as well. But

right, you're right though. You have a really interesting background.
[laughter]
Yeah. Before that you got into this. So, it's uh a mix of farm life, theater, AI research, and then game development. Could you explain a bit about I guess that journey and how it's led you here?
Yeah. Growing up on a farm was really very unique experience. I went to like a city school. So, even at school I was the odd one out going, you know, knowing all about how to lamb sheep and drive tractors, and that was how I spent my weekends. I kind of always knew I learned to read on Pokémon Yellow on like the original Gameboy, and I think it there was always some part of me that wanted to to get involved in in video games, but it was always kind of framed to me based off of my background as something completely unobtainable. The thing that actually made me go, "Okay, no, this is what I want to do." is when I played Undertale at 15. Um and

I was like, "If one person can make this, you know, I can join a team and make something incredible, surely." I ended up building a team. I I essentially tried for a couple years to get into games and um found it very difficult. And when I eventually got offered a PhD uh in AI research, um which, you know, I was very good at AI like stuff at at university, but it wasn't really I wanted to make games. I played the new Animal Crossing during COVID. I remember thinking, "Why do they all say the same things? Why do they not have changing perspectives of each other and emotions that kind of change their relationship. So, that's what I did for my kind of like undergrad thesis as I made like a a fully automated like emotion system. Wow.
And that kind of really framed the direction that I would end up going. It's kind of like this weird mix of psychology

and AI um to make something unique and I integrated a lot of that into Millennium Whisper um where all of the characters generate emotion tokens that we then use to kind of figure out what's happening in the conversation at any given moment. I was doing that PhD. It all went kaputs. Not all AI research is particularly good for humanity. Having gone through that uh when I came out the other side and you know, I was working for like a an environmental carbon accounting firm. So, which are brilliant Mycelium. Go look them up. They kind of inspired me to make my own thing because I was surrounded by uh theater kids writing and directing plays at the time. And I thought, well, you know, it made perfect sense to me. There's this idea of where do you get your data from to train the models to do what you want them to do. Right.

And a lot of people go, okay, let's there's this website online and if we scrape it in this one specific way, we're going to have it in this structure and we can restructure it to train an AI model to output this type of thing. And I did a lot of that um in my undergrad and part of me was just dying to make my own data and direct people within a system that I'd created so that it was perfect, so that it did exactly what it needed to do. um And because I had actors who could who were willing to come in and give it a go. Um That's That's what I did. I rented out a library um like a local library, and we all sat in the library, all of the actors, and um we had people kind of calling in to this system, this like version of Man in the Middle Whisper, and um it recorded everything, and that's how the data set was created, and it was

kind of from that that I ended up um where I am today, really. That is fascinating. I'm just blown away because we are constantly living and scraping and learning from the data we have, and so the more we generate with AI, people are saying we're going to go full circle because we have AI-generated content, data, you know, theater, whatever you want to call it, and it's going to learn from itself, and then we're going to get to a point that it's not learning from humans anymore. It's just learning from AI-generated stuff made by humans. Mhm. Yeah.
So, you get to a point where it's going to go flat, right?
The Ouroboros. Yeah. Yeah. And you We're kind of already at that point, if I'm honest. We're on our way there. We're on our way there.
Yeah. Maybe says something here, Craig, about what it is that also gaming studios are going to need to be thinking about because we're all living when we're all sort of building based on the same base, and that's not what Parable

Studios is is about. Mhm. Yeah, I guess uh the interesting thing at the moment is developers trying to grapple with how and if they should be using certain I mean, it always used AI technologies. I guess in this case we're talking about GenAI. It's always kind of uh easy to get confused with the two, I think, these days. You've done a good job leading up to kind of where you are now, and the thought process behind what you're doing. Could you explain, I guess, a bit more about what it is you offer now with with Parable Studios? So, yeah, we've kind of iterated on the Parable Engine quite a bit and it's essentially your one-stop shop for the whole AI development pipeline. You set it up, you create what is essentially a multiplayer game, like a single interface for your characters that is both for actors and for AI, so that when you kind of use actors to collect all of the data, it's directly

it can be directly used for the AI training without you having to really like, you know, data engineer anything. And then once it's done that, it kind of runs all of our optimizations on it and will give you a little model that can be ran on device. It's really up to the developer what model they want to use. If they want to use a super big one, well, A, you're going to need more data. And you might not really see that much of a benefit, but, you know, if you wanted to self-deploy it on some massive server, that's fine. You're kind of not really getting the benefit of it though, because, you know, like Peggy said, a lot of AI conversations are about how do you scale this thing? And really, the only way that you can scale it is by making it, you know, what's the best answer to how much does it cost per user other than

zero? Other than what Steam takes, right? Steam is already taking 30%. You want to take another 30% to OpenAI or, you know, for Azure or something? No, you want to be able to ship it directly to them and I think it it's something that is really standing in the way of developers making crazy cool indies with this tech, because an indie, if they're being creative and then you know messing around and taking a real risk because let's face it like every amazing indie title that you've ever heard of is ultimately been a crazy experiment and risk to these people and they don't want to think okay well you know that's going to cost me this many per user and therefore if I have this many servers and like I'll have this sort of margin that's not how game like indie development really works it works off of

like passion and people caring and like flash in the pan stuff you know and I think that's kind of what we're offering where sure you can be a multi-million dollar company and go to Nvidia and try and get them to do something with you or you could like come to us and we'll give you the tools to be able to do whatever you want um and I think the actor thing does put off some indies because like indies can't get access to actors always but I think the the nice thing about the Parable engine is it's quite easy to get MVPs that can get more actors involved very quickly without you know having to fully invest in in that side of things and and with Millennium Whisper they were happy to not get paid initially and go off of royalties right Wow

building a team indeed just tell someone to spell this out very clearly for our audience because some studios don't even understand fully or have a name for the problem that they're running into what what is the problem you're solving for the one that they're going to be seeing more and more of and seeing and experiencing more as well I mean, it's tricky because there's there's obviously the scalability of AI thing, right? But often times when I say that, they go, "But why even have AI in the first place?" And that's really the crux of the problem that's being solved. It's like, how do we have um NPCs or, you know, gameplay mechanics that work off of uh kind of continuous linguistic or kind of like very large state space inputs, you know,

um we have like these really excellent video games like Fallout: New Vegas that go, "Okay, let's approach it from a writing perspective." And Baldur's Gate as well. Let's write as Let's try and figure out as many of the scenarios as possible and try and fill in the gaps for each of them. first released Millennium Whisper, uh somebody We sent it to a streamer and they emailed about how "Oh, you know, it would've been so much better if we'd had if you'd had like a really excited team of writers writing every situation instead of um you having AI for it." But Millennium Whisper is based off of um like heavily inspired by Tokimeki Memorial, which is a Japanese game that never came out in English. And And I wanted to play it, but I don't know Japanese. So, I wanted to make something like that. And obviously, it was in 1994, so there's no AI in it. And they did that. They wrote for each and every one. Um

and I still would love to play something like that. It Millennium Whisper isn't going to replace uh a product like that, but I think um it opens up kind of an alternative way of doing it, potentially one that um scales slight like in a way that kind of matches the way that the industry expects you to scale now because Mining Whisper had a well Tokyo making Memorial had a crazy amount of money back then and a lot of kind of some of the best most pioneering game developers in the world working on on that game for a very long time Indies don't have that sort of luxury and I think you can get something here that unlocks similar sort of experiences because you still need to think like extract the information from these AI systems and develop something I guess is people being replaced by AI actual creatives so

Mhm. you're [clears throat] trying to empower creatives but also by as as Apple company to put it by creating efficiencies etc. You that basically means don't need certain stuff so is your engine sorry are your tools ultimately going to be a way of that indie developers don't have to hire certain person or do you see it as something else? I will like we still have a writer who does an awful lot of writing and we're kind of on the way to hiring another one and it because of we basically have like an internal scripting language which means that uh you kind of we have creative writers and not programmers writing this stuff um and that side of it is still very necessary. I think in the same way that like you know Balatro didn't need a writer there are going to be some AI games that don't

need writers I think if you're doing something narrative based you really do but that sort of opportunity has now been opened up to people who um aren't kind of won't necessarily will potentially really want to work in games, but aren't don't necessarily have the kind of incredible programming tool set to know how to program an LLM, right? We have um a couple different methods of doing that. And for instance, the actors, um they're kind of like new people getting into the industry in this way, but they're not quite like actors that we currently have who are given a script and have to deliver stuff. It's more like a dungeon master or like uh like Critical Role, if you've ever seen that, like those D&D podcasts. The same way that they're writers in that they're given a scenario, um they're given a kind of environment,

and they have to improv write their way through it. Um and that's like incredibly valuable. And I think there's a lot of people in the world who I can imagine being like, "Your job is to play D&D uh and and get it recorded uh and sa- go, that sounds incredible." Um which is exactly what I think. Like that's the end goal to me is we have an industry, kind of a a a proportion of the industry that are basically um big D&D games that have been kind of uh built on top of look the creative the the real-time creative input of a bunch of uh really excited people, you know, who love what they're doing. And and I guess what you're doing is you're focusing on developers have to have to use their own inputs, create their own assets, and then yeah, your engine kind of built upon

that, rather than drawing from anywhere else. Like for example, if one developer uses the engine, does that feed into some kind of algorithm for someone else to use? So, when they're building their data set, so for example, with our data set for Millennium Whisper, the actors signed it so that the data could be used on Millennium Whisper. Like, we can't use it on any other game. We can't use If we wanted to do, you know, our next game, we need to get new actors in, and we need to do it again. Um or we could, you know, technically get them sign another contract for the use of that. But that's how I think that it should be done. Right. Um and how we're going how we're approaching kind of like our pilot schemes with other companies. This is how we think that it should go. Um the it doesn't really make sense for it to kind of cross-pollinate because the idea and the reason that the models get to be so small is because they're not you have like Claude or

ChatGPT or Gemini, and they're trained to do everything. So, they're like trillions of parameters and need massive amounts of server input and whatever. If you get them to do one specific thing, um cuz realistically, what's being used in that trillion parameter is probably only a very small proportion of it for the specific task that you're doing. So, you know, there's a there's a cost thing there, but there's also an environmental and just like waste-based thing. And I think the game industry's always been very good at using what's right, the just the right amount of like innovation and grit and determination to make stuff work that wouldn't otherwise work because we we're good at making it small and making it optimized. Um and that's why it doesn't make sense for other people's data to intersect with yours because you're getting it to do the specific things that you want it to

do for your game, right? Okay. One of the concerns that people have is is AI slop. And so you're also saying that program games and this kind of opens up to them, which on on the one hand is a good thing, but the other hand people are worried about AI slop too much stuff coming out. Is that also something you're trying you're thinking about? And that's the way you're approaching it.
To sort of defeat Well, I think it's like we don't have AI-generated [snorts] art, obviously. We don't have AI-generated music. We don't have AI-generated code. Like and I think the real like AI slop best thing is, you know, vibe coding and trying to put together I think that there's a lot of um benefit to those tools in certain ways um in terms of like cutting out uh really mundane work, but I think the

idea is if there is a creative input, if that if you're expecting an an AI model to come up with an idea or develop a creative idea from you and not in terms of purely functional yes Mhm. this works or no, this doesn't work. Millennium Whisper, like the models aren't perfect. We're still iterating on them all the time. You can always tell when you talk to them that this isn't ChatGPT or this isn't Whereas when you play these other ones, you kind of can always feel it. You can always feel it in there. When I read an article by an a news article that's AI-generated, you can feel it. There's like no one on the other end. And I think you know, that's really deep down, once you've uncovered all of the other kind of things that we've had to do to facilitate this, that's what we're solving. That's the problem. Um

So, to step back to that problem, you know, it lets studios train, ship their own AI. They're not depending on a large external API. They're not depending on some sort of cross-pollination that's going to sort of water down everything. I get the value proposition. Who's it relevant for? Is it for the Indies or even potentially mid and large studios because, you know, they also need to be thinking about doing this differently. Who are you focused on and who would it be relevant to right now? I think that it has got to be Indies to start with because Indies are always the ones experimenting and and pushing the boundaries. And when I speak to like a huge uh company, it's always about they have this game, you know, they have their pipelines for how they get their specific thing together. And they want

to be able to kind of stick it in. They want to kind of get the AI Like like jerry-rig it into it. Yeah.
Which isn't the point of this. I think if you that's how you get the kind of like tacked on Mhm. kind of what what was the point in even doing this sort of AI integration. Really, you have to start from the ideas phase, uh from the very kind of core of gameplay loop. Why does this have to have an AI in it? Mhm. And the Indies, they're ones who are trying to um make a name for themselves, the ones who are have room to be really experimental and creative, um and pioneer their own method because, you know, we've done a dating game. And there's a billion different ways that you could have done a dating game with this sort of technology. There's also a billion other possible things that we've been talking about like court dramas and

like interrogation systems and all these different types of games, but that doesn't uh neces- like even that's not even really scratching the surface and I mean we kind of go, "Oh, wouldn't this be like a a super cool way of utilizing this tech?" Um and yeah, so that's that's who we're targeting. I think eventually triple A's will it will leak into like bigger titles, but it's got to it'll only be once like an indie's really proven that that sort of gameplay loop or that sort of development pathway really, really works. Well, that's the thing, isn't it? I mean, the more we have AI, the more the creativity stands out or something different and indies are great for that. They're going to experiment, they're going to push the boundaries. Let's get to how you're doing this, Ambrose. I mean, we have the proof point in Millennium Whisper. We didn't talk about it yet. So yeah, let's do that. You didn't just build tools and stop

there. You built a game. Tell us about it.
Yeah. So so it's in early access, so we're still kind of adding to it. Um and I think essentially, you know, the the pitch is it's a dating sim set in 1999. Um you redo the year of 1999. Um like however many times you do do a run of it and the there's a rumor going around that if you kind of confess to somebody on New Year's Eve 1999, you'll have a a thousand years of happiness together. It's all based off of that framing and you're trying to convince one of these characters to confess to you at the end of the of the game. But it kind of devolves a lot from there in terms of the things that you can get up to. It's uh kind of like a stat um balancing game um and a and a role-playing game in a in a lot of ways. There's a lot of branching narrative

that the AI then gets kind of slotted into. So, you know, you can be trying to grind up your fitness stat by going swimming. And uh I think this is like cur- maybe this isn't the best example, but there is one where you can uh you'll be swimming and one of your uh like favorite characters. Like we we have these metrics that track which characters the player likes the most. So, we'll pull in one of the characters' favorite um ones to watch them and then they'll kind of uh a kid in the pool will have weed and everybody will be rushing out and it'll be really embarrassing. And then that character can go and tell the other characters what they saw and it'll be kind of like a rushed uh uh cleanup duty of of trying to damage control any of the embarrassing or like mistakes that the player might have made whilst they're in this kind of simulated world. Um

and we're we're adding in like inventory, so that means that being able to dress yourself and having the characters dynamically react to the things that you're that you're wearing and doing and giving them. Mhm. Um cuz they already have that in terms of the places you can take them and the um decisions that you make between the characters. And I think I haven't seen anything where it's had multiple AI within a world communicating and and transferring information amongst themselves that can impact the player. And that's what I really really want to double down and experiment with um within Learning to Whisper cuz we've we've got like a really strong base there. Um Very cool.
Obviously, there's the the hardware limitation of that, where you can't have loads and loads of AI running at the same time. Um so we're trying to balance that and still target the kind of mobile eventually market, because it right now it's running quite well on um like

laptops and Steam Deck and stuff. So, Mhm. uh building upon that. You're saying that tools like this it's not just good for developers, where you can kind of I guess potentially create more out of out of the assets that you've already made or out of voice acting that you've already done, etc. It's you can actually create new types of experiences that users didn't have before. Mhm. Yeah. And um there is like a really fun iteration approach here, where um you know, you have the initial actors for the initial game play loop, and then you [snorts] can flesh out the world, build more in, and then you can collect more data and do do more training to have them better adapt to the situations that you've added. I think it works well for early access, especially because then you can kind of see the community that grows around the

different characters and what people like and and what people don't like. Yeah. A lot of layering is possible, I'm thinking the whole time. It's like it's self-generating its layers, as well as just the AI, you know, because of the possibilities.
I think the thing that's going to be the really fun stuff to start with is let's take preexisting systems that gamers have built a million times, like an inventory system, you know, super simple, but built into an AI game, where, you know, uh characters can have items and they can gift it to you, they can gift it to each other, they can have them, they can use them, they can sell them. You know, what does that what kind of gameplay, you know, is there like a stockbroker game on the horizon with this sort of stuff? I think there's a couple games that have kind of tried to do shopkeepery stuff with this. Um and I think

uh like combat and like adventuring and a lot of these have some quite difficult design decisions in terms of I okay, so how what are the decisions that the models can make and how much can you rely on them being real and how much does a mistake from the AI just completely obliterate the experience for the user? That's something that you really have to balance and and how much are we going to try and like calibrate this to make them dependable? Um but the there's a lot of like low-hanging fruit, I think, at the moment. Mhm. Um And what did so build what did building the game kind of teach you about your own technology and how it actually works in the hands of creatives? It was approachability first and foremost and having like layering of the the actual implementation. So, we had originally just like a bunch of code and

you kind of had to code everything in and it just became a nightmare and so we have this the scripting language that I said before, which makes it so much more um kind of like the people who are writing the the scripts and the AI scripts and the games flow and the agents flows, they're very much can self-contained within just a couple functions that can be altered on the back end by a bunch of very very smart complicated coding people. Um and having that distinction is absolutely paramount because if you don't have that, um I think either you're going to have uh very smart programmers um like creative programmers wasting a lot of their time or you're going to have um a lot of like creative designer people getting very confused. Um The other thing is speed.

Like latency is so important here. And that's why I think that these big models and this kind of direction towards reasoning uh as the way that models um kind of generalize to multiple tasks, why I don't think that that is really applicable here and we're kind of working on an alternate solution to that. Sort of narrowing the focus of AI when actually it was built to be exactly the opposite. So, that's going to be interesting.
Yeah, meant to be general, yeah. That's an interesting challenge, but one that's going to be important particularly as you look at more ways to bring this game to players as well. Thinking about latency just there. I always think about how when we were making like physics engines and we were making um graphics engines for video games, a lot of that work ended up being the pioneering work that spilled over into almost every other industry. Um And now

people use on Unity and on Unreal for all sorts of stuff, you know, not just game development. And I think that the work that game developers do because you know, often times they are the most creative and especially more creative than this rigid view of AI that these huge companies have, um I think we might have an indie dev one day come up long and design some sort of architecture that absolutely blows all of the pre-existing stuff out the water. If you look at the triple A ends there's some cases where you know, budgets are kind of expanding far too much. I think it's uh is it Ralph Koster? He has uh some sort of chart that says over time how dev costs are rising. In the end, who do you think's actually going to benefit the most from these kind of AI tools? Do you think it's these large studios? Or do you think it's small teams that can actually

probably things I don't understand, but adopt these new tools into their kind of production pipeline a lot easier than maybe a massive company that's kind of set in its ways. And games take so long to make behind be a long time to actually see that integrated. Or do you think like it can actually these AI tools they're just kind of for everybody? Yeah. I think ideally they're for everybody, but I think indies are going to benefit the most the earliest. I think it's really valuable at the moment when it's really hard to get your foot in the door to be able to say, you know, we have an understanding of these tools and it's not just everybody can write a chat GPT wrapper, you know? But if you can come along and say we trained our own models and they have this specific thing that no other AI model can do right now. Uh and maybe it's just like they uh are like a an AI jury, you know? And they can decide whether somebody's

guilty or not off of this thing. That's still like a super unique and innovative thing that's like never existed before. And if you can go to like, you know, a grant funder or an investor or something like that um to in order to really try and get something off the ground, I think that's so valuable because it's it's so tough. You go to GDC and as an indie and it's so uh forboding. It's so unfriendly. Good luck, kid. I don't think that's really in the spirit of games. This is supposed to be a fun industry. Um And if I can facilitate more of that for more people, I think that'd be brilliant. I I like that you said that cuz I had a conversation with someone recently just about writing and it's like for creative is like that's the fun part. I don't want to take away certain some of these parts of that process. I I get business reasons and that. But bringing it back to actually

why we do this. I think there's a really interesting message here, an important message for our audience is that it's going to be about the focus, about the specifics. As you said yourself, you know, if you have trained AI to do something very different, like the first discussion we had in AI when people were saying, "No, I'll get a patent on the prompt, right?" Um that's where the differentiation is going to be. That's where the business is going to be. Mhm. I think there's real value in having something that is yours. And if you make an AI game that is tied to an open AI wrapper, it's it's this ephemeral thing that can just disappear overnight. And I said this to Dave Bradley and I said it at Pocket Gamer London and I keep saying it, but it's like Millennium Whisper, even if I like disappear off the face of the earth, there's always going to be this like

memory of this thing that existed and it was the first thing that existed, which is impressive, but it's also just like in and of itself something unique and and persistent that can if somebody downloads it onto their PC now, they can play it 5 10 50 years in the future where their computer can where they might when their microwave can play it, you know?
[laughter]
So so where are you actually with with the game now? You know, you're on Steam. Um Steam Deck uh mobile? Move to mobile? We're kind of in a bit of a transitional space because the non-AI game development kind of team have a lot of ambitions for just mechanics and stories that they want to tell within this space um before it's wrapped up. Um we're hoping to get it out before the end of the year. Um but obviously having it on mobile is going to make that a lot easier cuz then

it's going to be a lot more approachable for people. Um and then we have we do know what game we want to make next and we know the innovations that we want to put in that. Um and we want to kind of expand our actor network and and start doing that a lot more because um one thing that I must say about the actor thing cuz a lot of people hear that little game devs hear that and go what a pain in the ass that sounds. It's so fun. It's unbelievably fun. Partly some of the best fun I've had doing this kind of work. Um it cuz it's just a multi a big very high quality multiplayer version of of the game that you're making. Um you know, and I think there's going to be so much value in recording those sessions. I wish that I'd recorded them like you know, I've got the data, but I don't have record recordings of everybody doing it and the the storylines that happened. Um because that could be really great, you know, marketing material, streamer

fodder, you know, stuff that people can get involved in. There are a lot of companies out there that you know, it's almost the making of that's more exciting than what happens in the game. Absolutely. Yeah, I hope that's not the case with Millennial but we'll see.
just just a side gig, just a side gig.
Yeah. Um but we do need to be wrapping it up. So, I would like to suggest that we've we've we've done a lot of the questions already. We just went ahead of ourselves cuz we've talked about what does AI unlock for developers and what does it look like against larger studios. We've even talked about Toy Story at the very very start but really we're really left with the bigger picture and what comes next. You already told us you know, what's next for you and Parable and you hinted at the kind of games that are coming. For developers, indie developers looking at these tools, maybe looking at Parable AI tools, how should they how should they approach

taking advantage of these? You've got to go in with just like a complete open mind. Be willing to sacrifice some of the ways that you've kind of done things before and really take the leap on a on on the most kind of out-there idea that you that you can get. I think a lot of indie developers have kind of heightened expectations of how much things are going to do. And this doesn't make things infinitely easier. It just it facilitates new things. So, other times people come to me and they say, "What about this idea?" And I'm like, "That's actually excellent and you should do that." You know, it there's kind of like a weird middle ground of there's a lot possible but there's some odd things that are actually really hard just because of the restrictions on the tech right now. If it's been done in visual novels, it's

def- definitely fine. Two questions that we ask everybody. Um what is your favorite game that you're currently playing? Um I'm currently playing Baldur's Gate 3 because I still haven't finished it. I'm also playing Metaphor: ReFantazio, which is by the Persona team. Oh, yes. Uh Katsura Hashino directed it and I love I love Persona so much. Um and I I like that they really took a risk on trying something new and weird. I've been highly recommended that game, so uh I might might check it out.
Transistor on the Switch. A lot of good monk chanting.
[laughter]
And then and then uh what is your favorite game of all time? It's Undertale. I always say it's Undertale. Yeah. Well, it started it all.
very very inspiring. Yeah, it started it. Yeah, and he he's he's such a funny funny person, Toby Fox. I strive for that level of

of um of myself in the end product, you know? Um or at least like the people around me in it. It It's incredibly emotional and inspiring to just have that sort of thing from from somebody who who was just inspired to do it all themselves. Um and I hope yeah. I hope to be a part of like an entire generation of people who were inspired to join the industry from that game. Wow. Um The perfect way to end uh this podcast. Something quite inspirational. And it goes back to where we started, so
[laughter]
Yeah, yeah. Absolutely. Goes full circle and inspiring. I think I think that's the point, Craig, you know? We've shown literally what is possible and we've shown how indies can grasp those opportunities because they can talk to Ambros. I mean, Ambros, what would they do? How would they engage with you? How would anyone engage with you? We're We're We're wrapping up. They want to

continue the conversation I am all for There's a website uh that's got a bunch more information about how a lot of this works and and what we're about. And there's kind of like a sign up thing for a waiting list for the engine itself. Okay.
You know, stick a little couple of your ideas in there if if you've got some ideas for what you want to do and we'll be in touch. Like and the The other way is to just like you know, reach out on LinkedIn if you've got that. A lot of indies hate LinkedIn. I hate LinkedIn. But it is it is there.
[laughter]
If if it if needs must. Love it. It's true. It is true. Well, as as as Craig said, inspirational, fantastic, and Craig, always a pleasure I would say. Until next time.