We're Making... A New Game Framework!? Survey at the End! Your input matters! (Patreon)

🔔So before you start reading, this post is very long, likely the longest one i've ever written. So if you don't feel like reading it, I'll put a TLDR and survey at the very end that wraps things up nicely.

Also I figured I should post this now, while our animator is recovering from a fever, and so that people can start to voice their opinions and shape the engine sooner rather than later.

I included lots of images so it's easier to pay attention... Or does that make it harder? Ah, oh well.

📝NOTE: All Imagery shown is AI made (not the charts), its just to showcase what's possible, its not apart of any game or anything of that nature...yet :P

🤔You might be wondering, why a new engine exactly? Well it started last year when dalle2 imagery was just starting to bud it's head and AI writing was just getting traction. I sort of realized Generative AI would only get better from here and in theory, you could develop not just a game... but entire games if you constructed scripts, APIs, Macros, etc. properly.

I think of Life in Woodchester as a fine homegrown wine. Its a long process but the result is a very high quality visual novel that I'm very proud of, and will continue to work on with the goal of being one of the best adult western visual novels. That hasn't changed, and won't change!

This game framework, however "Linkia", isn't really a competitor to LIW, but instead a new way of producing games. Instead of a fine wine, its more like a cheese board. It can produce lots of content, but its going to lack depth, at least at first. Its a way to produce a lot of content quickly using AI Artwork and AI Writing, that will need refined with a human touch.

Well food comparisons aside, I started to concept out a workflow, that would be broad enough to allow for several different game themes, but narrowed down enough to keep your interest and guide the experience. It would need to be flexible enough to build on but interesting enough to want to play through.

Perhaps one day LIW will switch over to this engine, but I don't want to jump the gun and commit to anything yet.

It's kind of funny that this happened to release during other gaming expos haha. Good timing I guess!

I think what you'll find are some interesting things throughout this long post, and hopefully it will fill you with excitement and you'll fill out the survey and leave us feedback! 😊

🏄‍♀️On the surface the concept is fairly straightforward, we basically use Generative AI to write and imagery to imagine it, and a game engine like Unity to produce it. But what's happening underneath is a bit more complex.

🎳Rather than just letting the generative AI do whatever it wants, we can contain it to a formula and continually add to it. Its sort of like bowling, sure we could say do whatever you want, but you would get a ton of gutter balls. If we put up the guardrails, you'll see a lot more strikes.

By separating both the art and the writing into their own modular parts, we can effectively create games and also have someone working on the game engine independently. It's a bit tough to explain right now, so we'll come back to this.

The project is entirely PRE written and PRE imaged. When you chain together writing, descriptions, and information, then you can guide the Writing models and achieve better results. Which brings us to...

A lot of people have heard of ChatGPT. It's the talk of the town, and it might be doing your homework for you. ChatGPT runs on something called a Large Language Model, or LLM for short. LLMs are much more powerful than you might realize. They're more than just a chatbot, they can write, classify, identify, extract, etc. all pretty much because of emergent behaviors. The most powerful ones are capable of logic and some level of reasoning, which are all important for Generative AI.

What's neat is we can start from a single topic, lets just say "Datable humans in the Big City". This is our foundation, and the entire game will use this as a parameter moving forward. This is key, because we can change this to generate entire stories, situations, NPCs, Houses, etc. for any other games as well.

📀As you can see above, we use the theme and core idea of the game to generate content such as names, occupations, locations, personalities, mannerisms, hobbies, whatever you'd like to include. The only real limitation, is how many tokens you submit (more or less, similar to characters) but this only seems to improve as tech also improves...

🧵Now since I'm the first person to see doing this, I get to dub it right?! I'm calling it "Narrative Chaining". We can string together random lists of information such "Lisa... friendly... accountant". Then we can ask the AI to reflect on this information, use a list of pre-generated locations and classify them a location. In this instance, the AI chooses "Delta Bank". Accountant at a bank? Makes sense!

Then we feed this information "Lisa... friendly... accountant... delta bank" and ask it what type of clothes should she wear? The AI chooses "Black Skirt".

We can continue this process all the way into narrative, creating unique one of a kind stories, because of all the variables that have been introduced.

By introducing Lisa's attitudes, fears, hopes, dreams, mannerism etc etc, we can craft a story that has a more meaningful or satisfying outcome.

This is also great when it comes to the sex of the main character. Ideally if setup correctly, we can flip a switch and have scenes that used to be for a male lead, into a female lead.

Obviously with this method we can get as granular as you could possibly want. Body Type? Cup Size? Likes? Dislikes? Freckles? Pubic hair? Height? Favorite Sex Act? Bedroom Mannerisms? Breast Shape? Nipple Color? etc. etc. etc Anything you can think of.

AI writing is pretty great, but there's something irreplaceable we can do...

🦹‍♀️Now alternatively, and more excitingly as well, we can let patrons create the characters as well! Either entirely or flagged appropriately. This opens up a huge number of possibilities. Entire characters you can meet, greet and... well interact with 😳! All driven by letting your imagination go wild! This will likely be a survey or online excel sheet that patrons can access that contains all that juicy info like personality, hair color, eye color, clothing, you name it!

Most games previously cant really offer this level of interaction with the community, normally limited behind high paywalls, but with this framework, we could let patrons with even the smallest of pledges, to the biggest and boldest, create their dream characters.

🎊Imagine creating and then meeting your perfect waifu!? Bring your fursona to life in a furry dating sim!? Breed the hottest purple spider angel-demon monster girl that you helped design!? The possibilities are truly endless and pushing the envelope of creativity is something I would love to see!

We can throw character information into our existing story generators, and you get to experience, what is basically, life in another world! 🤯

This would be more than just leaving a name in credits, it would be about leaving your fingerprint itself on a project! (It may be that not all of the games will contain user generated content, but likely most of them will allow it!)

Which brings us nicely into our next topic...

🎨So writing and characters are all very good, but we're not just making a text adventure game right? Well all the artwork on this post is AI generated and I can barely believe it and I've been following AI art since before Dalle-2 last year.

So you might be wondering how AI imagery will work exactly. Basically we can use AI writing to condense the already written short stories, locations etc, into Stable Diffusion prompting, combined with our own prompts.

We then import those prompts and use python automation to run through each prompt. The resulting images are renamed so that they better fit.

🎭What's interesting further, is that we can even swap out styles or models to support multiple art styles. During the game you could switch graphics to any of our list of pre-chosen styles. Don't want stylized graphics? No problem, switch to Realistic. Not a fan of realism? How about Cartoony?

Of course each style would double the image generation time as well as the storage and would double the clean up efforts in regards to quality control, but it would be a great feature to support as a project grows.

I don't know about you but all this AI talk has me a bit overwhelmed, which brings me to my next topic...

So by now you might be asking a few questions,

Is Generative AI even ethical? The AI model trained from humans that created images and human written words.

It's definitely a hot point of contention. The project itself is definitely standing on the shoulders of giants, and I wont pretend that it isn't, but at the same time such is the techy life a lot of us live right? This tech is available to everyone equally and thank god for that. I'm just simply trying to use it as best I can.

Even when I was a kid I wanted to make games, I was enamored with everything about them. I had to get into pixel art and programming, it was hard stuff. I realize now I kept at the programming parts mostly out of necessity, not entirely out of passion. To me it was sort of an ends to a mean so to speak. I know not everyone feels this way, but to me these were, at the time, essentially hurdles to overcome.

Now an unfathomable technology has stepped in that can do art, it can write, and I'm sure before you know it, will be the lead programmer leaving me in the dust. Eventually even this engine will be outdated and will be even easier to create. Or perhaps 3D latent space will reign supreme.

--

What about all the artists and writers you're putting out of business!?

Well first of all, let me be clear, we're still doing business as usual for Life in Woodchester.

Also we'll need actual artists and writers to come in and tweak, fix, enhance, detail and re-prompt existing artwork. Nothing is really perfect the first go around, especially AI art in its current state.

Writers too, the writing choices need to make sense, and the intro story will require human steering for certain.

So to be honest, I don't really have much extra money even now to create these types of games. What I'm getting at is that these projects wouldn't even start to exist without this kind of technology. So I don't feel like I'm taking away "business" from anyone.

As far as artist protection, I don't intend to rip off or abuse the styling of any particular artists or written works of a particular author that may have been caught in the training. No artist prompting.

What's most interesting to me is that this technology is available to everyone. I honestly see the current situation as leveling up existing artists, not taking away their abilities. Same with writing, you have the power of an editor and writing enhancer at your fingertips, go for it! I know even as a programmer its been quite helpful for me.

--

🤖In Summary, I just want to give people a taste, however small, of the impossible. A view into the world of latent space and a tiny view of the future!

Speaking of a bright future, how exactly do we maintain quality you might wonder...

🌟A long standing meme has basically been that AI can't draw hands and neither can you, but as models have merged and data has piled on, its getting a lot better from every aspect.

Again, combining human fine tuning on top of our workflow will ensure a quality controlled game at the end of the day.

One technique is to run an AI Image model several times and choose the best one afterwards via human eyes.

Also while images like the above are quite impressive, training data starts to wear down at a certain point for various poses or acts and positions. That's why we'll likely be focusing more on a cartoony, anime, flat, anime/realism blend etc type of style (as of writing this post) where we have a more forgiving environment and don't cross far into "Uncanny Valley" where things look kind of creepy because they are not exactly realistic.

But alas, grab your string cheese and let's get back on the tour bus! This ain't over yet!

☁ So cloud storage is something that we currently have and are testing with. The idea being, you could have thousands upon thousands of images all stored in the cloud, and only grab the ones you need for a specific scene.

You could also adjust on the fly, sort of like a streaming service does.

It could make for "infinitely" saleable games with many characters, theoretically even hundreds or thousands of them and worst case we can just compress the images and do standalone releases, for those of you always on the go.

But I don't know! Some people have really crappy internets, or don't want to always be online, so feel free to let us know your preferences :)

"Wow", you might be saying, due to witnessing the real power of cloud gaming. Or perhaps instead you are wowing at how its even possible to drag on a Patreon post this long.

At any rate, if you would walk this way we can get your right back to the core of the game framework, the engine mechanics...

⚙ So great, we have setup a complex web of generative art, big woop. Without the "game" engine its not really a game is it?

Great point! So I figured I'd run through roughly how the game works, some of the design philosophy and how its built in a modular, future forward way. The following is the actual UI screens, images will change depending on game.

The first thing you might notice is the appearance screen. Here you'll be able to select from several avatars rather than just one! Here you could pick an edgy wolf in a furry dating sim game or a nerdy geek in an Isekai game! Because of generative AI, you can choose from several avatars, and those avatars could have stats or characteristics that some hate, and others love!

The game engine works similar to the core LIW experience. You have a central hub/house/apartment/room and you're able to take actions. Those actions will cost you time and days. Depending on if we've added the main story or not, you'll only have so many days to accomplish your task! Or perhaps its a more laidback sandbox type game, and you can spend the days as you wish. Likely, we'll be including both so achievement hunters have something to look forward to, and those of us who just want to relax, can do so!

[Above is a mockup of how an image forward design can help drastically change the vibe and mood of the game]

As you play the game you'll learn more about the characters you meet, their likes, dislikes, occupations etc, etc. You can use this info to help you make better decisions and stay informed.

You'll visit locations around the game that trigger various character stories. For now, you'll choose from 5 choices on how to handle the scenario, each with their own ways of handling things.

Eventually, at the end of their story you'll have to make a choice that forms the major branches of romance. These branches will all have unique scenes only available on that playthrough, encouraging both player agency and multiple playthroughs to get alternate endings. Each branch will also have a relationship level that allows you to unlock more options.

- 💗Romance: This option will set you down the path of a traditional romance style relationship. Go on intimate dates & enjoy the world together as a couple. Experience traditional romance and sex, as well as some secret naughty things that character will want to share with you! (Could be dom, sub, or anything based on their character profile)

- 😈Domination: This option will set you down the path of a domination style relationship. They'll likely call you master as you make them do your bidding. You're the one in charge now. As your domination relationship level increases, you can have them do more and more naughty things. Public Sex? Time to get embarrassed! Fang Inspection Day? That's everyday now!

- 😨Submission: This option will set you down the path of a submissive style relationship. They'll likely call you names as they make you do all sorts of things. I hope you like being a worshiping servant! As your submissive relationship level increases, they'll have you do more and more naughty things. Lets hope you can hold your breathe for a while...

So these are the basic pillars I've come up with for romance, I'd love to know what you think. These are also expandable.

- So take domination for example. We could add in later, if you get to say level 20 domination, they'll go to a slut house for public use.

- If its romance, perhaps you'll build a house and design it together.

- Or if any relationship is over say level 10, you can get a pregnancy to happen. Or you could attach this to a usable item.

At any rate, the mechanics are made in a way that lets them be modular and expandable, which brings us to our next topic...

🧩 Another great thing about having the art and writing separate from the main engine, is that as we expand the game engine, adding things like Items, or events, or stats, those can all be integrated into the games almost seamlessly.

What this lets us do is have a person or team adding to the core engine, while others can continue to refine actual games. This is great because it means we can release multiple games and update existing ones with relative ease.

📓 Phew! If you read through the entire post, congrats, you can skip this section. If not, I don't blame you! Here are some of the bullet points.

• I'm working on a new game engine called Linkia, designed to use Generative AI for writing and producing games.
• The engine will create game content like names, occupations, personalities, and locations based on a single theme.
• Using a process I call 'Narrative Chaining', we can generate unique stories and game experiences.
• Patrons can contribute to the creation of characters for more interactive experiences.
• The artwork for the games will be generated by AI, with the ability to swap out styles and models.
• Generative AI raises ethical questions, but I see it as a tool that's accessible to everyone, not as a replacement for artists or writers.
• I plan to ensure quality control by fine-tuning the AI-generated content with human intervention.
• The game could potentially use cloud storage for a large number of images, but I'm considering user preferences for internet usage.
• Because of generative AI, we can continue to add to the engine, in parallel with making games and updating them.

💡Phew. So up until now most of the talk has been about the engine, but what about the actual games!?

Well that's why I need your help! I want to know what kind of themes you'd love to see. Now obviously I expect patrons to choose things similar in taste to LIW, (which is totally fine, and something we can do), but keep an open mind as to what's possible 🎉

Here were some ideas I had gathered based on trends in the market, popular subreddits, personal taste 😜, etc. All names are subject to change as well as details about them.

Also try to keep in mind, although some of these have some art styles, all art styles can be swapped or even added at any point. So its less about visuals here and more about the substance. Though most of the screenshots are at least somewhat in the ballpark of what I was thinking.

📝 We will eventually try to aim for a male and female protagonist for all games, and with multiple sexual orientations as well. Likely though, at launch, we'll have to focus on just one though. Optional Futanari could also be something that attracts a wider userbase. Again, we would love to hear what you think!

Each game will come with the 3 branch system of Dom, Sub, Romance, and eventually inherit the other systems we're working on. All games will allow patrons to deeply customize characters or in the case of Universe 34, vote on them.

• 🌟Universe 34 - A rule34 parody, dating / brothel sim. This idea revolves around allowing you to become more intimate with your favorite characters! Will you let the sparks fly with Lara? Turn the princess into a submissive? Players will vote on characters they want to see in game.

We can even do gender swaps on characters. A female Masterchief? A Male Tracer? And/Or adding optional Futanari.

• So I had already started some Logo material, and templated out the writing cells, but we'll see what the survey says and also your feedback!

-----

• 🗼Lumenlove Skyline - A city of scandals and naughty fantasies. This idea is similar to Life in Woodchester, with the main differences being, you're on your own in the city and patrons will get to design the characters down to the very last detail! Will you find love in this busy concrete jungle?

-----

• 😻FurFable: Lovebound - A city of anthromorphic wolves, dogs, foxes, birds & many, many, more! Also inspired by Life in Woodchester, with the ability to both play and create NPCs based on your favorite fursona! Including all the details from floppy ears to pointy beaks.

-----

• 🐻Mythic Meadows: Monster Breeder - Explore the world to find, catch & breed lovely Monster Girls and Guys. Sell their products for money and keep them happy in their own preferred ways! Create your very own monster girls with other patrons and experience them in in all new world!

-----

• 💦Harem Fandom: Exclusive Access - Play as a marketing expert, you see all the unrealize potential in girls have in creating "fan content". Convince girls around college campus and beyond to start creating lewd content for your new app. Give them tips and even offer to "help out" with some of those high earning scenes. Get more money as the fans pile in!

[Considering using a more realistic model for this one as the default? Remember though any game can have multiple models if that's what the players want.]

-----

All very exciting ideas for me. Of course I'm excited about all of these ideas, but If I had to use some of my predictions...

• I think the popularity vote will go to 🌟Universe 34 simply because of the many characters you can already relate to and become intimate with.
• I think the most intriguing would be 🐻Monster Breeder. With generative AI you could do things like making offspring of two monsters? Could get some interesting and unexpected results...
• My underdog thought is 💦Harem Fandom, as college girls, onlyfans, nudes, are big subreddits, but having the content range from softcore to hardcore, and having text conversations/manipulations with life-like girls could be up a lot of peoples' alleys.

Below, are some other ideas based on some popular fetishes, subreddits, games, etc.

• MILF Dating Service
• Teacher and hundreds of students (could play as foreign teacher?)
• "Landlady" & "Roommate" stories only. (Same protagonist, different timelines and all different scenarios and ways the stories can unfold)
• College Life & Drama
• Giantess
• Expansion
• Weight Gain

Again these were more like secondary ideas on my radar, but let me know what you would want to see by voting in the survey!

👏 As the engine develops, we'll be creating a new Patreon, but even still I want to make sure I'm giving back to you guys who have brought us this far!

So while the poll will be open to anyone, if you have pledged any amount to LIW (Life in Woodchester) so far, you'll get 2x voting power on top of their current tier's voting power.

Anyone who is a past, current or future Producer or Creator will get a guaranteed character slot for all future games. (We'll probably limit this to top 30 for now, with producers and current subscribers having the highest priority)

Next in line would be VIPs and then Patrons after that. So less waiting time!

The voting will be on everything from the game type, to the next engine focus, and so on.

Its possible we may open up a limited, very high cost tier, that guarantees we make your game, but we'd have to make sure its TOS enough and talk through the details. Not sure, we'll see.

So if you haven't already, please fill out the quick survey that lets us know more about what you want from your game experiences :)

⌚ So In theory, if we have the machine well greased, prompts are setup, macros working, scripts all check out; I think we could develop one of these games in about 1-2 months. I

ts a few days, maybe a week of altering our setup. Then generative AI it really would take less than about 72 hours of image generation and the writing is probably about 48 hours or less. Then we comb over it with people to see parts that went wrong or need resubmitted.

Now as for when we're ready for the first game, that'll depend on the game chosen and some other variables. I'm thinking maybe 2-3 Months until we're ready? We'll have to see since I'm terrible with timelines 🙃.

For now, we'll keep Linkia info on this Patreon, but expect it to move to its own Patreon.

Thanks for coming to my ted talk. 💙 Lastly, go fill out the Google Forms Survey! Do it! :)

See you soon with an update on the poll, and the next LIW build.

Best,

DirtySock