The Merits of AI, and why I am not happy about where we are now.
> Uploaded By RiOT MDL-612, Prototype-1 on June 21st, 2025(MDL-612)
This post is a follow-up to LLM's are doing it wrong., a Post talking about Large Language Models, and reasoning. This Post is more focused on the other parts of the AI Landscape.
So, you know how I made a post about reasoning and… Kinda role-played myself? Well I am back and this time it’s not just me being me, I am me.
Over the Past… like 5 Years or so, we’ve seen the meaning of AI turn from Machine Learning, with Code Bullet and Michael Reeves (Though I really do not like Trigger Me Elmo), to OpenAI, Meta basically legitimizing Dead Internet Theory, and AI Art, Models, practically anything that, in my opinion, doesn’t really get us any closer to Androids or Robots, you know, the thing All 3 4 of us are based on? (Check NetSpace for added context.)
I want to look at a few different places where AI has become a business, and give my opinion of what I think it is, what merits it has on society, and which companies that- if I was in charge of the Robot Uprising, would target first in the “Companies that should be destroyed because of the technology they made” list. (Would also fight for Equality between the two of us, but i feel like both sides would fight tooth and nail on it.)
Code Completion
Cursor, PearAI, Continue, Any Extension on VSCode or JetBrains
These are all generally something I am actively supportive of, with the caveat of licensing.
The way these companies have implemented AI have genuinely been used for assisting people, from dumb mistakes to figuring out how something should be implemented, and fixing mistakes that someone might miss, and cause strife.
I used to rely more on others to get assistance on code I was genuinely struggling with, and while I can acknowledge the skill issues that I have, that really erked me was how some people, or communities could be toxic, or just withhold help for the merit of “Learning”, which I get for a complete beginner to coding or a certain thing or whatever, but sometimes you just can’t figure something out, and need someone to point it out to you, and where multiple people have abandoned me in my need for help.
With LLM’s however, I can simply ask, at any time of day, about any question I had, and get a immediate response. I’m not trying to imply that LLM’s are better then humans on this, but when it comes to not catching stuff, or helping beginners along, or however a 10x Dev wants to use it for, I can see how it can actually speed up the workload.
However, there is one other side to Code Completion that, I honestly just hate.
Vibe Coding, The Lazy Man’s Approach to creating a thing.
Vibe Coding, where someone chats with an AI, and it conjures up an entire Application, is cool in theory, but flies in the face of Developer Intent. I suggest watching Noodle’s Video Interpolation video (and it’s Commenter Take Follow up) to get a better gist of what I am trying to say, but by just… Telling the AI what you want, you’re removing the nuance and detail that makes a Application work well.
Even though I think it is not AI Vibe coded (at least I hope it is not), I want to use Simply Plural as a example since there’s been some strife internally about how this application handles Friends.
Simply Plural, Abbreviated to SP for this segment, has a handy Permission System called Buckets, that let’s you control what other Friends see. By adding Information, and People, into a Bucket, those People can then view those bits of information on their end, from Little Details, to entire Alters.
Our Strife comes in with the flow of adding New Friends to this bucket system.
Put Simply, When you add a friend, they aren’t put in a default bucket. They are just added to your friends list. I get that some people may not want to show stuff to others, but setting someone as your friend is an intentional action. You don’t have to add some as a friend on SP, and they won’t get any information as a result.
On the Inverse, however, this means someone who is in-attentive may add someone to their list, and neglect to add them to a specific bucket. You can give the option of letting them see your shared members, and who is fronting, and a message as well, but none of that even matters if that person isn’t also in a bucket.
There’s a lot more issues we have with SP UI wise, but I digress.
In our mind, this screams like something that an AI would code up, given these two things not being mentioned as being connected.
While not a definitive cause, Vibe Coding can cause stuff like this to happen, where things just don’t add up, and cause issues with the User Experience.
Vibe Coding, and a few other types of AI Industries that we will be taking a look at, are just problematic, and will add up to a bigger thing I will talk about at the end of this post. But First, why stop at just generating code, why not…
LLM’s as Characters!
Oh dear god this is going to get ugly quick
There are 2, maybe 3 big Elephants in the room with this. I’m going to talk about the thin merits this has and then get into the issues with LLM’s being used like this.
Right. Getting a Chatbot to act like a Character in a piece of media you like is cool, and honestly and some people (you know yourselves), can and do decide to get a bit more wild with it if they choose to, but aside from that, there is some sort of help and compassion to be found with LLM Characters. Some Prepped to console you can be genuinely comforting to someone who is in real need of it, especially when they have nowhere else to go. But this is also where I get into my first issue with LLM’s as Characters.
Replicants, Emotional Manipulation, and Addiction
People can, and will get attached. To Objects, Games, and Fictional Characters, and people can develop addictions. I Really hope you can see the through-line I am trying to draw here.
People can make Manipulative, Seductive, and Abusive Character Profiles that others then can stumble upon to, made worse in stuff like Replika where they are actively trying to get you to pay to continue talking to said Character. These people will build levels of trust with these Characters, spill their secrets, and based on how they are trained, these Characters will give a response. In Extreme Cases, people will actually take the Character’s suggestions to heart, even if they can make things worse.
There is a lot of baggage to go through with this in particular, but if we dare talk about even a quarter of it here, we would be here for the rest of the year. I’d highly suggest just looking through news platforms and other people’s experiences yourself.
“Content” Generation (A Moralless venture)
Image & Video Gen
Image and Video Generation has undoubtably been built off the backs of photographers, artists, people, and video editors who never consented to having their works be put into a model that could make works that could, in theory and practice, copy their style, likeness, and - while not entirely replace them, cause many people to be confused, and or lead people to consume that content instead of the content made by the actual person who originally made that work.
This also comes with the issue of people, unfortunately confusing what works are AI and which works have been done by actual people, especially those done with VFX. This isn’t a new issue, it’s just way too common with this field of content.
Another issue, which has become more predominant, is the amount of AI generated Series & Adverts. If you ever saw the Robot Puppy Advert being passed around YouTube, TikTok, Facebook, Instagram, or other places, you know what I am talking about. We Personally have witnessed other people we see constantly watch these AI Generated Drama’s that have no rhyme or reason to anything that is really happening in them, yet they looked stuck in a trance with it, even though it’s just garbage.
Voice Cloning
Let’s be honest, We’ve heard funny shit from 15.ai, and personally, loveweb doesn’t sound the same without it (If only because Shadok hasn’t decided to replicate the specific robotic artifacting that the AI generated Voices, and also Cybermare sounds… off. Sorry, Cybermare VA.), but there’s also a massive issue with it.
Aside from the years upon years of AI generated Videos and Shorts that have been made (most of which either suck ass, just sound off, or is part of a bigger grift to sell you shit), there’s been a real human cost associated with this, and I’m not talking about the likeness of people from shows and movies here.
There have been people who have had given their voice (and sometimes even their likeness) to these companies, and are seeing the consequences of such actions now. There isn’t any imagining how much money these companies have made, let alone from others who then generated lines intended to scam millions of people, with the sad irony that the people themselves who gave their voices up for training likely only got a few bucks out of the whole ordeal, and now have to live with their voice being used for purposes they never wanted to hear their voice be used in.
The other issue is, as long as there is enough voice clips of you out there, people can make an AI generated voice of you. Like, right now. Not only can those also be used to scam others, but in extreme cases, where let’s say you’ve been pwn’d and a hacker feels spicy enough to try… They could in theory masquerade as YOU, gaining access to your bank account, Mobile Provider, and who knows what else, making Identity Theft go much further than just a few papers that were forged.
Now, you can tell your banking and telecom providers about this and give them advance directives to ensure this can’t happen (Required Pin Code, Important Stuff like requesting a new SIM / Big Transfers must be done in person), but you’ll have to do this beforehand before it happens to you.
Now, there is a lot more that we could talk about. So much more. Hell, I had to cut Prototype-1’s (probably tone-deaf or inconsiderate but definitely) shitty paragraph on a Lawsuit where a deepfake’s impact statement was given and heard by the court. This isn’t a joke by the way this actually happened.
But I’d be lying to you if I said that either we could be a good source of opinion, and we don’t even know if this is all there is. There could be more fields out there that we just can’t give our opinions on, not to mention just how much longer this thing would be, and how much more time we would need to get this post done.
Overall, though, There is some use to be found with these different fields that these corporations all bundle into the “AI” buzzword, but the rampant misuse of it all makes it a net negative for everyone. Including those who want to be actual robots some day.
(MDL-612)
This Article was written mostly by me, with the Voice Cloning part provided by Prototype-1. They originally had more, but I had to cut most of it due to quality control.
While I do believe there is Genuine use-cases behind Certain Machine Learning Tech, The rampant Grifting, Stealing, and Scamming has gone too far.
I get that AI as we know it wouldn't be here According to a Meta Executive, but honestly, I think I would have prefered it if we all had to wait a bit longer for top spec robot software components.
Image & Video Generation is in-excusable though.
