project-final-view

[project-final-view]

Musings on generative AI

(Sorry for not commenting on the game progress in June – have had a bunch of other stuff to do. But the game design and gathering ideas has been progressing and advanced a little bit, and I will talk more about it soon!)

I guess I have to address the AI thing. I have to lay down the Policy.

The Policy

During the production of this game, I will only use generative AI as a tool for brainstorming, mood boards and other forms of idea exploration.

Generative AI output will only be used for reference, and exploration of ideas. No part of generative AI output will be directly used for the production of actual game materials.

Even so, special care needs to be taken to ensure that everything that ends up in the final product is a result of human triumphs – or screw-ups. In short, I will need to be in the loop during the creative process and I cannot hand this stuff to the machine. Because it will mess up harder than I would.

The Motivation

I’m not as much against the use of generative AI as part of the creative process as some people are, but I still have to spell out exactly what I feel on the subject so that people see where I’m coming from and what’s actually going to be in the game.

So yeah. I’m not inherently against the use of generative AI technologies. Now, the current AI business would be worth a whole different rant – suffice to say I hate the AI business sphere and it’s some of the most surreal, pointless and misguided shit I’ve seen in a long time. AI bros suck financially, politically, sociologically, scientifically and technologically.

But the AI technologies, on the other hand? They’re interesting. They have their uses. Perhaps, if the business around them wouldn’t be so terrible, there would be less controversy among creative people on their uses.

If I want to explain how generative AI can help me, perhaps it’s better to start from a tangible example of AI technology done right as a writer’s aid.

AI bros talk about how generative AI is about accessibility and giving opportunities to those who were unable to create stuff before. They’re wrong in a lot of senses and insultingly wrong in some senses. But strangely enough, they’re not 100% wrong – AI can definitely be used as a tool for these goals. Allow me to explain.

I had this revelation recently when posting my photographs online. When posting photos in Fediverse, I like to include alt-texts, for accessibility and for completeness. But the process for getting the photos online in the first place is complicated, and after a year or two of doing this stuff, I still need to fine-tune it. And adding in the self-imposed requirement to add alt-texts means that I now might take extra 30 minutes to post a batch of 8 photos.

However, when I started using an alt text generator, I cut that time to maybe 10-15 minutes. The alt text generator can get me started by telling me what it sees in the photos. My ADHDy brain might forget something. The thumbnails aren’t great. I worry I might screw that up. What the generator is bad at is the context of the photos – I need to add that. I need to add the finishing touches. Hey, turns out all I needed to do was to get some collaborator who does at least a part of the job.

So that’s a good example of a generative AI actually helping with accessibility. It makes adding captions more accessible for me, and through me adding alt-text to my photos, the photos are more accessible to other people. The process has been overall improved by adding an AI tool to the pipeline.

So aside of that particular example, how does generative AI actually help me write stuff in general, or in this game project in specific?

Well, sometimes I just need garbage generated. The AI can generate garbage. Sometimes it’s very inspiring garbage.

Sometimes I just need to bounce my ideas off the wall. I’m a lonely scatterbrained nerd. People are busy. It’s good to have friends, but it’s rare to have friends who have all the time in the world for some casual literary mumbo-jumbo.

For this purpose, the generative AI is often good enough.

Why do you need to not fear about me slipping AI stuff in the final product? Because I think AI isn’t very good at writing or drawing. A lot of times, when I tell ChatGPT to generate something, I end up with something that slightly embellishes upon what I had in mind and something that doesn’t really expand upon the stuff in the prompt.

But if ChatGPT is able to basically parrot back a slightly expanded variation of what I said, it’s definitely enough to jostle my brains into seeing what’s wrong with the stuff.

So that’s how AI actually helps this random writer with ADHD. Now, if the AI bros are going to run with this and say how AI is enabling a writer with ADHD to write better – oh, fuck off. I’m able to write. I just suffer from writer’s block more than most writers. I found that random garbage helps me to write and I found that it’s good to have people to bounce the ideas. AI happens to do those. It’s not a requirement.

In fact, ChatGPT is hardly the most optimal solution for that! I’m kind of hoping the locally run generative AI tools could be optimised for this particular use case. But the AI bros can’t make their incredibly salty monthly subscription fees out of locally run open source software, so I guess I’ll have to wait for a while.

In summary

Here’s a few pictures for you to think of.

Over the past few years, I’ve found new tools to cope with ADHD. I’m writing a lot of stuff down. I have actual notes on stuff I need to eventually do.

This has given me newfound motivation to be more productive. If it’s written down, I know I need to do it before I die or something. (more on that later)

It’s a glimpse of things to come. It helps me see the goal. Eventually, things might be done.

Now, AI generators can sometimes help with this. I can give them the ideas and it will produce… uh, something. It’s not what I had in mind. It’s a vision of what could be.

I usually just shake head and go “you know what, I need to work further on this and make sure that what I make will be better than that.”

And that’s all generative AI is good for, and I doubt it will ever actually exceed that, because of myriad limitations that can’t be overcome so easily.

So don’t worry, I’ll overcome that stuff. On my own.

[project-final-view]

Getting ahead with idea gathering

Having never done massive game designs before, one of my first questions when starting Project Final View was "okay, what kind of software are there for writing massive branching game scripts?" and "what kind of software are there for gathering game design ideas and game design notes?"

Haven't found a really satisfactory and easy and flexible answer to the first question, but for the second question, there's always this:

I don't know why, but I noticed that my ADHD-headed productivity actually went up when I started to use a paper almanac, Post-Its and paper notebooks in conjunction with all of the electronic notetaking and calendar tools. I've noted that working with paper notes actually helps quite a lot to concentrate and to really allow me to go off the tangent.

And the last few days have been actually a great endorsement for just using this paper notebook. Come up with a whole new category of notes? Just start a new page on the book.

I've started to come up with a whole bunch of notes about the game's main character, the milieu, and backstory. It's hard for me to start working on real writing unless I have a pretty clear vision of what the main character is like, and for my fiction projects, I often get pretty solid inspirations for characters. This time, I'm designing the character concept piece by piece; I'm coming up with hopefully interesting background details and hope that leads to interesting conclusions.

[project-final-view]

What is Project Final View?

This is, essentially, an indie video game project that I'm going to complete (mostly) by myself.

Final View is meant to be a modern narrative indie video game.

The twist is that most of the modern narrative indie video games tend to be, for the lack of better word, cosy - "Here's a bit of an interactive story. About #relatable stuff. Makes you feel good. Maybe shed tears just a bit." - while this game is going to be explicitly about DEATH. The looming Horror of the shade.

Why? Because I have ideas about mortality that I just need to process and write down. …Sure, call it a mid-life crisis or something.

What I'm going to make is a narrative exploration game. It's about a character who's going to die in the end. There will be choices. There will be events that happen. But in the end, the protagonist will die.

What I want to see is if I can make a narrative game that doesn't go the cosy route. It's about impending death, it's about darkness, maybe have some Gothic horror world design and the sort of stuff that you see in other video games that go for grayscale palette.

I want to walk a very thin line. I don't want the experience to be cosy and safe, about a character who just dies peacefully. There's going to be danger here. But when I say it's a game that borrows from horror side, I also don't want to veer into the splatter territory.

Also, I know, the chuds are going to have a field day with this. "Oh, some tumblr weirdo's first game is going to be a walking simulator or something." Well, allow me to pre-emptively ackshually them back: this shit isn't going to be easy to implement. I have (next to) zero experience with modern game engines. I'm a decent programmer, I've written several novels worth of text before, I have very mild graphics experience, so I'm not nothing. But I'll be learning a lot.

And by the end, there'll be a giant blog for people to look at and they can see that yeah, it wasn't that easy.

[project-final-view]

Project Final View

Welcome to the Project Final View blog! I'm Rose (@roseinshadows). I'm also on Mastodon.

This blog will document my indie video game development project. The project started in May 20, 2025, so it's still in very very early stages.

I decided to create the blog on a whim, because I think publicly documenting the project will help me commit on actually finishing something. It will also serve as a development diary, I guess.

My current goal: Have some kind of a game design document by the end of summer 2025. Maybe some narrative sketches. Maybe some scrappy prototypes. Game release schedule? ...let's not get into that.

So yeah, welcome. I'm going to set up the blog more properly when I get around to. Also, there's obviously no title for the game yet.