Ramblings

ULTRACOMFY's personal homepage.

User Tools

Site Tools


artificial_intelligence

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision
Previous revision
artificial_intelligence [2025/07/23 14:38] ultracomfyartificial_intelligence [2025/07/23 16:45] (current) ultracomfy
Line 7: Line 7:
 <WRAP centeralign>Computer Science/\\ <WRAP centeralign>Computer Science/\\
 <fs xx-large>Artificial Intelligence</fs></WRAP>\\ <fs xx-large>Artificial Intelligence</fs></WRAP>\\
 +My personal definition of Artificial Intelligence (AI) encompasses software that uses machine learning to dynamically acquire and then execute a skill. This specifically means the use of a neural network to mimic human learning through teaching and iteration (as opposed to fixed, preset algorithms). When put to a task, a neural network seemingly is capable of breaking a problem down into numerical values, according to which it will then act to maximize a score. The score is often called "fitness" and is meant to describe how well the AI did. An AI will chase the maximum score it can get.
 +
 +====== Modern Day AI Proliferation ======
 +The world seemed to implode for a moment when ChatGPT was released in 2022. Just briefly, society was under the impression to be getting a glimpse of the future and the arms race towards AI was on. Since then, everything is now AI, even toasters are hard to get without it. The problem is that it takes a lot of processing power and time to train AI, so most of the AI you will find in pretty much every product that didn't need one is total garbage and does things awfully or in a way that is almost useful but would have been easier to do yourself.\\
 +AI now exists for virtually any task you could imagine - visual art, music, tv scripting, programming, video editing, full video generation, interactive video generation (there is an AI that lets you play a very trippy version of minecraft), and, the one that kicked the latest round of hype up, text generation and interactive conversation. For example, ChatGPT is a program designed for holding genuine conversation, and oftentimes it will also get popular facts right. For that, it is really good. It seemingly has also found useful applications in creative spaces where facts aren't necessary, brainstorming for creative endeavors for example.
 +
 +====== Anti-AI sentiment ======
 +Especially in the field of art, the proliferation of AI has found itself a crushing reputation. It is despised by pretty much everyone - training is based on what is essentially stolen artwork((Until it's not eheheheheh! Fuck you, users of Instagram, your art now belongs to Facebook! ~Message proudly sponsored by Facebook (They changed their TOS to make you implicitly agree to have your artwork used for AI training when you use the service, making it non-stolen as you "agreed" to it.))), it wrecks the environment, has a bad influence on society, it's problematic for artists who make a living off of their art((They are now competing with AI "slop" that can be generated with just a few clicks in a few seconds, outcompeting them in speed and price in a way that is so convenient as to almost negative the poor quality of AI-generated art.)) and ultimately benefits the big corpos who own and run the AI systems. There's a lot to dislike about AI, regardless of the type of application the AI was trained for.((To be very clear, critics like this speak almost exclusively about AI products for sale - not research AI used for research purposes by computer scientists. NASA uses AI to analyze trends and pattern in research and exploration data, from weather observation to the moon and mars missions. The critics I am talking about do //not// typically refer to these uses of AI, as they consider them to be good and genuinely useful applications of AI.))
 +
 +And while I can understand agree with these points as a concept, it does not lead me to the same conclusion. My then-bestfriend is staunchly Anti-AI, whereas I felt like AI wasn't the problem and that AI is //good//, it's just the way it's currently implemented that is, admittedly, bad. Like, for any of the points there is something you could respond: Yes, the artwork is stolen, but your brain too is trained on the things it sees around it/"if you put stuff on the internet, anyone can use it for anything"((ie. = It's your own fault.)). Yes, AI is costly on the environment, but that wouldn't be a problem if our energy came from climate-friendly energy sources. Yes, artists get fucked out of a market, but that's just the market - don't sell air if there's air all around us. Yes, AI art looks shit((Does it really? I'm not an art person, or anyone, really, for aesthetics and the much.)), but that can and probably will change over time. Yes, AI benefits big corpo, but they //are// offering you a service - we don't go hating on bikes just because they benefit big corpo.\\
 +\\
 +However, I don't think this is a good way to argue either of our cases. I don't think I even properly agree with any of these rebuttals, even though I //did// bring them up my then-bestfriend. It just misses the point. It's not even that this is a discussion about "I am defending AI as a concept, you are talking about the implications of AI //as// they are currently implemented in our world", because AI will always be a problem for real-world artists. The concept of AI is //always// a threat to human artists, this problem //will// not be solved. So from what I gather, the proposed solution seems to be to abolish or severely limit the permitted use of AI.\\
 +
 +====== The Comparison ======
 +So, let's talk about cars. Cars have a problem. Primarily, cars are terrible for the environment. They require infrastructure that makes our cities look ass with all the roads, they reduce the walkability of our cities, they disincentivise the creation of tightly-knit communities, their infrastructure is expensive, they are deadly, constantly cause congestion and so much more. What's the solution?\\
 +"Just walk" is //not// a solution. Not only would this cause inherent problems because our infrastructure isn't planned for carless people, it's also that there are tangible benefits to cars or similar machineries that can genuinely advance us as a species. We wouldn't be here if it weren't for the revolution that cars are.\\
 +What we need to do is to **provide alternatives**: Public transport (trains!), bikes and urban planning that enables car-less day to day life (keyword 15-Minute Cities). This would reshape our society and our living places in such a way as to //harness// the exciting things that cars can bring us while minimizing the problems that come with them.\\
 +\\
 +"Banning AI" as a solution to the concerns raised about AI is like suggesting "just walk" as a solution to the concerns raised about the car. It's not a solution. AI is too useful to be just banned away, it //will// stay around, in one form or another. What we need to do is to provide proper alternatives, proper ways to use it, such that we can actually benefit from the invention. We need to reshape our society to minimize the casualties of the proliferation of AI. This could be more or less extreme, some would say we as a society just need to go full [[https://en.wikipedia.org/wiki/Post-scarcity|Post-Scarcity]], [[https://en.wikipedia.org/wiki/Post-capitalism|Post-Capitalist]] Fully-Automated, some say strong copyright laws are all we need - to me it matters that in the end we find a way to integrate AI in a way that benefits us as much as possible.\\
 +\\
 +One such example of shaping the world around new innovations was the electronic calculator. Teachers fought against it because it would be detrimental to their courses. And it was, the calculator genuinely hurt the courses. But because of how useful electric calculators were, they were there to stay and eventually it was the courses that were reformed to adjust to the benefits that calculators would bring - less tedious, long handwritten calculations, more focus on theory, formulae((Even though we have formula tables for these as well, hah.)), more //specialization// as the "manual labor" of calculating was largely cut out of the job. Overall, students learned more, better and deeper mathematics than anyone before them.\\
 +I fully believe that a similar transformation can be done with AI (regardless of whether I think that such a transformation actually //will// happen).\\
 +\\
 +<wrap lo>And sorry to the human artists out there, but I //really// don't think you have a moral claim to be without competition from AI. AI will compete in your space, and it will win, and I'm sorry that this is a reality but I don't see why it shouldn't be.</wrap>
artificial_intelligence.1753274334.txt.gz · Last modified: 2025/07/23 14:38 by ultracomfy

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki