Return

Rendering, AI, 2026 2026-01-03

One of my biggest achilles heels in programming is drawing graphics on the GPU, which I also complained about in my previous diary. Whether it's OpenGL forcing me into an awkward path that turns everything into spaghetti, or because I have no idea how to design an efficient but flexible renderer, I've just never figured out how to do it well and it has demoralized me for a long time. I recently saw an article about how GPUs should work if the world wasn't dumb, and yeah, that would make my life infinitely easier.

A while ago I made a breakthrough that allows me to do a lot of things I've never been able to do before.

5000 opaque sprites, 5000 semi-transparent sprites, this video is just from the corner of the whole demo.

What this demonstrates is the ability to efficiently sort opaque and semi-transparent objects, and switch the shader (the rainbow colored sprites in this video) for any given sprite. None of these sprites live on the GPU, they're drawn from scratch every frame from the CPU. The majority of the performance cost of this demo actually went to generating random movements for these sprites (sin() is surprisingly expensive on the CPU).

This also allowed me to make progress on Scavgame. The video below is slightly unrelated (it's showcasing a developer editor), but you can see a correct field-of-view shadow, wavy grass, and correctly sorted semi-transparent graphics, none of which I could do properly before.

I might be able to finish the game like this, but it's already kinda slow (not slow-slow, but slower-than-it-looks) because I can't draw pre-generated sprite buffers the way I thought I could, so I have to draw every floor tile and grass blade one-by-one every frame. I should be able to fix that without too big of a refactor though.

AI (LLMs)

Following the line of thought from my previous diary about libraries, I started writing some tutorial-like pages for various specific programming tasks like software rendering. However, I started feeling extremely repulsed by the idea of publishing those tutorials and ended up not doing it. This website doesn't exactly have a big audience, but AI models would definitely consume it and make it a part of themselves. That thought makes me feel like I'm a slave working for the profit of AI corporations that give me nothing, but then have the gall to ask me to pay for their usage while they're destroying the economy and making everything computer-related expensive for me, and as a cherry on top all the AI bros are out there blowing a raspberry and calling people who create the content that AI is trained on obsolete and useless.

There seems to be some kind of bitter jealousy that was released full-force with AI: many people who can not create anything have become very aggressive towards people who can now that AI can create things for them, and there's a complete and total lack of understanding towards the feelings of artist etc. Non-creators seem to think that artists have been hoarding some job opportunity for themselves and now are angry that they're losing their jobs, but I think it's more like their sense of meaning in this world is being threatened by some corporations and being replaced with something soulless and fake. Art is much more than just a product.

I think the most valuable use of AI is as a search engine, to help you find information and learn things, especially since search engines themselves have become completely useless. There's been a lot of times when I wished I worked at some company that has pros that I can ask questions about complicated topics from, such as specific details about how servers manage memory. With AI you can ask your question directly and naturally, and it will answer the question directly, you don't have to try to manipulate your question in hopes of the search engine giving you something hopefully related to your question, only to still get a load of crap back. AI sometimes gives you links to more resources, for example it gave me some Nginx manual that exactly answered some questions I had searched about in the past but never found an answer for. If search engines could answer specific questions, they would have linked me to that page long time ago, but they just can't do it.

What I don't like about AI is the parts that generate content that is served to people. I've already stopped clicking on most videos because it's so tiring to expect to see something interesting, only for it to turn out to be some AI generated fake video. I see copious amounts of AI art on imageboards and the internet and I don't really get why people even post it, I can only kind of understand the value in porn and memes since those kinds of images are not enjoyed the same way as art in general. In particular I have a strong distaste towards those AI generated ghibli images and I instantly lose respect for anyone who uses them, I feel like that could be a rant of it's own.

I'm conflicted about animation, especially 2D. There's a massive bottleneck in animation production and quality because animation is so time consuming and hard to do well, it's kind of cool how AI can take your image and animate it. Maybe the reason I don't hate it as much is because you can use it on your own art, animating it yourself would be so time consuming that it's effectively never worth doing even if you could, and it's pretty obvious that it's AI. Even indie animators tend to cut a lot of corners and use very simplistic art styles, and anime uses a lot of CGI and static images to keep it financially viable.

With art it feels like taking something that people love to do and replacing it with something soulless, but with animation it's creating something that almost nobody has the time or finances to do, plus artists can use it to animate their own art. In a hypothetical future where society becomes prosperous and people can spend time learning and doing animation properly and actually make careers being quality animators, I might change my mind, although animation will probably become even less financially viable in the future. I don't know how it would work for "proper" animations rather than just shose short animated images, the idea of watching an AI-animated movie, even if all they keyframes are hand-drawn, the idea sounds icky but I guess I'd have to see it to know what to feel about it.

There's one major part where I disagree with AI critics, and that's that something should be done about it. The problem is that nothing CAN be done about it. Even if you start regulating AI companies or something, some other country like China isn't going to give a crap about your regulations, they will use it without restraints and then their AI is going to simply be superior to everyone else's, which will cause everyone to use that. Banning digital programs doesn't work. I don't know how compensating people would work, there might just be so many people to compensate that you would just be shutting down AI, and again, China isn't going to compensate you for shit. Another problem with compensation that it would give massive corporations monopoly over AI, nobody could make useful indie AI models since none of them have billions of dollars to compensate everyone on the internet.

Things like Steam's AI label are not very practical, I think it's good on paper, but you already can't see many parts of AI such as source code or designs or assets that were edited afterwards, nevermind art and sounds that aren't easily distinguishable. It's also a bit silly to act like a game that used AI on some minor asset/design in a corner somewhere, and a game that used it for most of the content, are the same. It is interesting how some pro-AI people who act like AI is the greatest thing ever seem to get mad when they have to disclose AI usage, I'm not sure what that means though.

People using AI to write emails/articles/blogs/social media posts is an abhorrent abomination that shouldn't exist. AI should not be used for communication outside of translating languages. There's one game developer who obviously uses it to write their blogposts about game updates, and it's painful to read, it makes every insignificant thing sound like a smash hit update which comes across as incredibly condescending.

Job loss is a separate topic entirely. Most jobs are already useless button pushing jobs or jobs that could be done by 1 person but is fragmented among 20 due to bad management, what happens if AI automates most of them AND some jobs that previously were useful? Society is already kind of broken, I don't have any idea how it's going to work in the future. I certainly don't think that we'll have something like universal basic income, let alone "universal high income", it doesn't make any sense for a lot of reasons. Maybe it will happen, but it will be started by China because they're more authoritarian and are willing to do whatever changes are necessary, including throwing some people into a landfill, or only giving basic income to people with high "social score".

Well in any case, even if AI is a bubble that bursts at some point, it's not going to go away. Here's some uses that I think are potentially valuable for it:

I don't personally use AI for anything other than answering those specific questions that search engines can't answer. I definitely aren't willing to pay for them until they start paying the people their AI is trained from (i.e. never).

Work

I have what could be considered a very lucky job IRL. It's in a small company, work times are relatively short and flexible, I work from home part of the week, very relaxed and easy job. However, I'm pretty set on quitting that job by the end of 2026. Or to be more specific, quit working entirely.

Even though the job is easy, I perceive it to be a waste of time. The majority of my time there goes to some dumb crap like fighting with Adobe programs and redoing all my settings and window layouts over and over, finishing something that someone else should have finished (some jobs are distributed to multiple people for no good reason which is less efficient for everyone involved than the person who started it also finishing it), trying to get Google to do their job, updating third party plugins on 100 different websites over and over, and sitting around waiting for a retarded fucking server to unblock my browser.

The website server provider we use blocks me basically every other page view when logged into the admin panel, and then unblocks me after 10-30 seconds or so, it's impractical to do anything else during that kind of time period, and there's like 100 websites we're managing, so it's just an absolute crapton of time wasted on nothing for no reason (not that going through them one-by-one clicking update buttons is meaningful in the first place). We've known and I've complained about this problem for a very long time but nobody seems to care about fixing it. If it was up to me, I would move our business to someone else and permanently blacklist that server provider, but alas, it's not up to me. That said, if it was up to me, I wouldn't be so fixated on constantly updating everything that already works either.

There's 2 kinds of waste of time. There's things that are uninteresting and undesireable, and then there's something that serves no purpose or eats much more time than necessary. You could be working as a trash delivery guy and think that it's a waste of time for the first reason, and in a sense it's true, it's not a very valuable use of your time on a personal level, but that work also serves a purpose and is necessary for society. My problem is with the latter, incredible amounts of my life is spent on things that are completely pointless. It's easier to do than delivering trash, but it serves no purpose because it should clearly be automatic, has no good reason for happening or is done in a stupid way, or millions of users are forced to waste their time on because some fucking corporation isn't doing their job correctly.

There's a lot of times when I'm too tired to do anything on my free time, but I can't go to sleep because I need to keep my sleep schedule intact so I can wake up for work properly. I'm also very often most motivated and inspired at late night, but then I'm forced to throw that motivation into the trash for the same reason. It feels somehow extremely offensive. There's other things I want to do (including wellbeing-related things like going out more), but that I don't have enough time for since most of my days are already spent on my job, and I'm spending my free time on being tired and useless, or working on my projects.

When I think about things like this, I can't help but feel despair at how much of my life's and time's potential is and has been completely wasted, and would continue to be wasted if I continue working. I've felt this way for years, but I was never sure about what I should do since I wasn't confident that quitting would accomplish anything.

There's also secondary issues with my work, such as the fact that we do marketing-related things, which are sometimes dishonest or manipulative or contain subjects that I disagree with, I feel like I'm doing something unethical at times. Another problem is that I'm being made to do things that I don't consider to be my job, is not my specialization, I'm not good at, and am not interested in, which results in low quality result being delivered to a customer, and I feel bad about it. Even the things I do specialize in, are not being done anywhere near the way I would do them if it was up to me. Then there's AI, multiple times now I've had to clean up a mess that someone else created with generative AI and it just feels so incredibly soul-sucking. I also have a health-related problem and doctors in here have never been interested in helping me (that wonderful "free" health care at work), I'm not even 100% sure if I have long left to live, so I'm feeling less and less comfortable spending my life on all this crap.

I've saved enough money that unless society crashes down (not unlikely to be honest), I should be able to live at least 5 years (maybe up to 10 if I really optimize money usage) without working. I'm hoping to create something during that time that allows me to start making money from my own projects, making games seems the most likely. Now with my renderer looking promising, I can actually see the path to making a game, that has been the primary blocker for me. I suspect I might come up with completely new smaller game(s) before finishing the likes of Scavgame or Volve.

Even if I have to return to working by the end, I think the younger you are, the more valuable your time is, so it's more valuable to take this 5-10 years as soon as possible rather than keep saving it for later indefinitely.