discussion How long before AI is able to make a full Renpy VN/games like the ones we play today?

[TittyCommittee]

Citation needed. What do you mean by Advancement?
AI generative technology "entrepreneurs" do not care about feelings, but they're subhuman anyway.
Advancement is just a word to describe something moving forward.

By advancement, I mean having major progress in AI. which is reaching its plateau currently.
Honestly, even if it does work. I don't see it being viable for personal computers
Do you think you are able to run a multimodal AI model that is able to create games on a 4090? I highly doubt it.

 

[HandholdDegen]

By advancement, I mean having major progress in AI. which is reaching its plateau currently.
Honestly, even if it does work. I don't see it being viable for personal computers
Do you think you are able to run a multimodal AI model that is able to create games on a 4090? I highly doubt it.

Which is why I hope people will stop simping for it soon.

 

[RawDawg]

Im thinking a few years. The first few will be okay but you can tell. Then it will transition into being good and even more exactly what you are looking for

 

[TwistedKnyte]

Musk just came out with possible robots.... and chatgpt is getting closer everyday to perfection... SKYNET IS HERE!!!!

 

[lcflex]

No, that is not what that means. Did you think that through before you responded to it?
Please think about it a bit longer.

You're going to need to explain that one, because it certainly sounds like you described AIs doing exactly what humans do when we create things.

 

[WillAnderson]

99.9999% of creativity is taking existing things and reconfiguring them into a different form.

Not exactly. Humans concatenate previous written sources to some extent, sure. But we also concatenate life, which is too chaotic. At a certain level of abstraction, pretty much every life has been lived before, but reasonably, every life is unique and every author or artist has unique experiences that go into what they do. That is the major difference. The sum of all fiction is a huge library to cut and paste from, but life renews trends in fiction. What the artist has seen, both art and the world, matters.

As an example - William Gibson is credited with creating the cyberpunk genre. This is a bit reductive, but Neuromancer is a different work compared to what came before it, though it draws on other sources from before it. Nevertheless, the genre didn't really exist before... 1970 at least, in any real way. Suppose you train an "AI" on the sum of fiction from before then. Could it have written a cyberpunk novel? I don't think so. It took a bit of inspiration, a life, a certain odd combination of optimism and pessimism about the way we are headed. Of course, now ChatGPT can churn out cyberpunk novels all day, because there is something to draw from and combine with other works. But there aren't ideas there. So yes, there is a significant difference between a creative human mind making an artwork and an expert system producing a prompted piece. They could both be useful things in their own way, but they are dissimilar in very distinct ways. I think it will probably be interesting and enlightening - not to mention quite surreal - to read a book written by an AI that has its own experiences to draw upon.

 

[TittyCommittee]

Which is why I hope people will stop simping for it soon.

Hmm Hype will die.

 

[lcflex]

Not exactly. Humans concatenate previous written sources to some extent, sure. But we also concatenate life, which is too chaotic. At a certain level of abstraction, pretty much every life has been lived before, but reasonably, every life is unique and every author or artist has unique experiences that go into what they do. That is the major difference. The sum of all fiction is a huge library to cut and paste from, but life renews trends in fiction. What the artist has seen, both art and the world, matters.

As an example - William Gibson is credited with creating the cyberpunk genre. This is a bit reductive, but Neuromancer is a different work compared to what came before it, though it draws on other sources from before it. Nevertheless, the genre didn't really exist before... 1970 at least, in any real way. Suppose you train an "AI" on the sum of fiction from before then. Could it have written a cyberpunk novel? I don't think so. It took a bit of inspiration, a life, a certain odd combination of optimism and pessimism about the way we are headed. Of course, now ChatGPT can churn out cyberpunk novels all day, because there is something to draw from and combine with other works. But there aren't ideas there. So yes, there is a significant difference between a creative human mind making an artwork and an expert system producing a prompted piece. They could both be useful things in their own way, but they are dissimilar in very distinct ways. I think it will probably be interesting and enlightening - not to mention quite surreal - to read a book written by an AI that has its own experiences to draw upon.

That's a good example, but not the way you think. If you trained a LLAMA on pre-1970 work and asked it to produce a cyberpunk piece, it couldn't do it, because the term didn't exist and nobody had mashed those particular pieces together in that way before. But if you told the AI, "here are the things that make up cyberpunk. Go write cyberpunk", it could. So yes, someone has to think of the arrangement like Gibson did, but that is overall a tiny piece of all creativity. Gibson didn't invent any new pieces to build stories out of. Most authors, artists, and other creators will go their entire career without even encountering a genuinely new piece, nevermind actually invent one. The ability to be creative is not dependent on the ability to invent something that is completely new.

 

[MattShizzle]

To be fair, there was a story that sort of predicted the internet LONG before Neuromancer. A Logic Named Joe - March 1946.

 

[mustafarxxx]

I think people glorify tools like chatgpt too much. They are basically software that predicts the next word. I think we are a long way from the desired level of AI.

 

[MattShizzle]

Chatgpt can write original stories and such already...

 

[Bigfoot_13]

with the speed of ia innovations, I think it'll happen sooner than expected.

 

[silver577]

Its already doing, but some humans are acting as minor assistants

 

[kokokokokokoko]

I think it should already be able to create a basic VN in Ren'Py given it's trained for it, but we are far (2y? 5y? 10? is it actually possible?) from being able to create an actual GOOD VN

 

[s3gfaul7]

I think people glorify tools like chatgpt too much. They are basically software that predicts the next word. I think we are a long way from the desired level of AI.

I disagree. ChatGPT is fantastic for what it is. It will never reach AGI level, but that's not really the point, is it?

As it stands, a competent creator can optimize their workflow a lot with 4o's help. And that's where the misunderstandings come in - AI creating a game doesn't mean that the AI decided to make a game and did it. It means a creator used AI heavily to help them both with writing and even scene composition, not to mention coding. Just being able to offload some of that workload is amazing and these tools are only getting better. And the better they become, the easier it makes it on the creator to get what they need from them.

 

[TwistedKnyte]

I think it would depend. AI(In this case not true AI, theses are just rudimentary learning programs with very strict parameters) if given enough information on what to do, and what to avoid then I think it can become pretty compelling. Just remember unless the planet goes pop, things will only improve

 

[Squeeky]

Let me rephrase the question. How long before AI tools become good enough so that developers can create the completed games they want within a few weeks or days? Instead of the weeks or months it takes for just an small update? (Small gameplay wise. I'm sure the work involved is great.)

This I think is a much more practical goal. For an AI to spontaneously create an adult VN, it would need to "want" to which doesn't seem likely, but for a developer to use a set of AI tools to generate story, images, and program logic based on prompts, I agree with some here that's 5-7 years out. AI can make some amazing images today, and while they are assembled from other works, it's such an intricate assemblage that they appear original. It's like saying people can't be different because we're all made from essentially the same chemicals, and yet there's a huge variety of people. (And I'm sure you've noticed some people look just like some other people you know.)

I've actually looked into doing an AVN with AI, and while the image consistency is rough, the toughest thing is to get something close enough to what I want, especially scenes of a person interacting with another person. That will get better, and probably faster than I expect.

The "good" news is that many advancements of IT are driven by improving the access and quality of porn. So we are in the sweet spot for where AI will be pushed.

 

[herelooking027]

This is unlikely to ever occur for many reasons, but the simple way to explain it is that "AI" is basically just a word salad machine and then humans put sorting bins under the word salad bowl and code and recode the robot's algorithms until it mostly sorts correctly. Human input is fundamental in making and altering the algorithms not just in prompting the machine to do work. AIs will never be able to do general tasks without training for specific purposes and stacking purposes on top of each other will only enable a certain level of results. As an example, AIs are horrible at making new code for purposes, but are great at suggesting code based on what a human is actively doing and how that compares to code the AI has been trained to recognize previously.

Edit: AI would be able to make cookie-cutter games pretty quickly if it gets a lot of data about the same kinds of games. Claude can remake a minesweeper game for example. But AI won't make new stuff or things that are unlike a lot of what it has already been shown.

 

[badplottwist]

It would be difficult. You can't just take an AI and say "go." Unless you put in a lot of guardrails, it would be the most generic story, poorly written dialogue, 12-fingered characters, directionless nonsense. AI chatbots don't really have object permanence for basic conversations, and you're probably looking for a game to keep track of hundreds of variables. We're probably a good ways out from an AI being able to write a kinetic novel, let alone a full VN.

 

[AbbaDoo]

Quicker than most people think. Within a year perhaps.