Hello there! I’m DiamondGotCat. Feel free to call me Cat :3
I’d like to integrate AI into Zorin OS!
I’m hoping to collaborate with everyone on this—what do you think?
I think all current "AI" tools are subject to criminal levels of false advertising. The term "AI" has specific connotations thanks to decades of sci-fi portraying it as near-human or surpassing human intelligence, while our current LLMs are essentially predictive text algorithms. They are not even close to being an echo of a whisper of what most people think as AI. In 50 or 100 years, when / if we actually make something that could really be called "AI", do you think those entities will look back fondly on this period in their development?
Apple themselves, despite their upcoming "AI" tool being a core part of their recent showcase, have just published a scientific research paper to remind the people working on LLMs that they are not, in fact, AI - they cannot reason, they have no wisdom, and they can't even really admit when they don't know something or are wrong about something. Somehow, people were actually shocked by this "revelation".
Yet, almost every single company peddling "AI" these days markets them as Artificial Intelligence and pushes them into places they should never be, putting nowhere near enough emphasis on their shortcomings. This leads to otherwise intelligent people like teachers and lawyers believing that our current "AI" are far more competent than they really are, and using them in ways they should never be used.
My own experiences with testing "AI" tools within the company I work for has seen shocking results. They regularly cannot follow simple instructions about how to output their responses, and include information about USA fire codes when I specifically ask them for UK building regulations. At best they give correct, but generic and bland responses to the point of being only truthful due to not being complex enough to contain a lie in the first place; certainly not helpful, though.
The fact that there are actually such tools being marketed towards, and adopted by, hospitals and healthcare providers terrifies me. The people involved should be sent to prison, but I fear that even when mistakes are made by and due to these "AI", they'll dodge accountability thanks to clever contracts and our laws not changing fast enough for this emerging threat.
And that's without mentioning the obscene privacy and copyright violations inherent to how these "AI" tools are trained and operate.
What do I think? There are indeed some specific use-cases where specially designed and trained "AI" tools are well-suited. But those are few and far between, and still have a tendency to go very, very wrong.
I think you are, unfortunately, free to make LLMs and market them as "AI" for purposes the technology is demonstrably unsuitable for, including one tailored for Zorin, but integrating such a thing directly into the OS like Microsoft is doing with Copilot and Recall? Nah. Nope. No.
That'll be the death of Zorin, or just about any other linux distro to try it outside of the likes of Red Hat and Canonical with their business-oriented releases. Because businesses are still blindly jumping on the "AI" bandwagon like lemmings jumping off a cliff.
People come to linux, for the most part, to escape the likes of Microsoft, Google, and Apple as much as is reasonably possible while still engaging with modern society. The others do so because they have old hardware that is struggling to run modern Windows or Mac OS, or at least is no longer officially supported. I doubt many of them would appreciate you forcing the same garbage on them that they just, purposefully, left behind or that their computers aren't powerful enough to run anyway.
The people who still want to use our current fake AI to be lied to by their computer for some crazy reason? They can install it themselves, or just open up ChatGPT in their browser.
Thank you for pointing that out.
It's true that it may often give wrong answers.
Right now, I actually have an AI I made myself, but it's not very accurate.
I won't touch on this topic.
Why would we need AI ? lets keep that stuff at Microsoft or Apple
When you do that on Your own System, have Fun with it.
What I think is pretty easy and short: I don't want AI on my System - not online or local.
This whole AI Stuff is mostly totally overhyped. It is a new Fashion Word for a new modern Time. But at the End it only steals (more) Data and Energy. Feel free to use it on Your System; that is up to You. But when this would land in Zorin OS ... To be honest, I would go away from Zorin.
@DiamondGotCat FYI: There is quite a number of posts relating to AI infiltration on PC's and other Operating Systems here. The future of Windows
Basically, the main players of A.I. is just a search engine as pointed out above that searches all available information based on the search request. True A.I. is machine code that writes itself that cannot be intercepted by Humans.
I have already voiced my concerns regarding AI being used in customer service positions, or anything related to running on your operating system, where your security becomes in question, with AI running on it, and snooping on everything you do. We already have enough problems with AI being forced into our cell phones as it is!
I now have a new example of bad AI customer service, that I have personally got to unfortunately experience yesterday. I was making an order for specialty coffee, direct from manufacture as I've done before. I wanted to use Amazon pay, but it forced me to use Shop Pay for some reason.
Because it had been some time gone past since I last used Shop Pay, I didn't remember that I had used my credit card with it, instead of debit. Because I didn't notice it right away, I made the order. OOPS! Then realized later what I had done.
Turns out, it was after business hours, and their not operating during the weekend either, so I couldn't get ahold of a person. And since their site is nowhere near as good as Amazon, I couldn't cancel the order myself, using the site tools.
So what did I get, chatting with AI, that I swear is not even as good as CHATGPT. It didn't matter how many different ways that I said it, to that stupid dumb version of AI, it would not cancel my order, and instead refereed my case to emailing the staff, for which they aint gonna see until the following Monday, tomarrow.
Not trusting at all that the AI was worth the code that takes up space on their server, I decided to personally email them as well. I had to get the email address from Google though, cause I coudn't find it on their simplified site. I emailed my case, requesting immediate cancellation and refund to my credit.
So, what do I think of the AI interaction? I considered it an absolute freaking nightmare, and I absolute hate that AI they are using 100%. I don't want more being charged to my credit card! But AI is so freaking stupid, it couldn't cancel my order, so what good is it? It was clear that it wanted to only handle simplified cases with simplified answers.
So here I am, still waiting until Monday morning, when I hope and pray the customer service staff will see my email, or the AI refeered email, to get that dang order cancelled. Yes, in the end, this is my fault, I didn't double check the card attached to the Shop Pay account.
But its not my fault that the company uses a piece of garbage AI, that makes everything a freaking nightmare for no reason. I have to suffer now, until my order hopefully gets cancelled.
So, why am I not a fan of using AI in our everyday lives?
( 1 ) I don't want AI snooping on my entire life, to make a corporation get richer off my data, and to sell me endless adds I do not want, just to complicate my life having to reject each and everyone of them, every single time.
( 2 ) When AI can't do a simple job that a human can do, and frustrates me so much, that it raises my BP, and makes me nervous as heck, cause it won't handle my case, and do the very simple thing I am requesting of it.
AI is not ready people, its not ready. STOP USING AI!
A.I. (pure A.I.) would disagree with you!
Hello @DiamondGotCat it is nice to meet you, I enjoyed looking at your project. You put a lot of hard work into it. I'm on a forum for game making enthusiasts and there is a lot of interest in generative AI, as you can imagine. There have been some clever examples of its use in games from members there. I imagine it must be pretty interesting to make your own generative AI, though I have no idea what's involved. Is it expensive to train? How are you getting your datasets? I ask because under help you say "extend the dataset". What part of the project roadmap are you working on now? Any way I think it's a great tool for those that like to use it and understand its limitations, problem being as already pointed out in this thread, that currently marketing for the word "AI" has been used to the point of being misleading to the public. And that companies are releasing some that are frustrating customers, as pointed out in a reply on this topic; which is poor form, they really should be testing it better to ensure it does the job they want it to
My initials are A.I. (no, really). I don't mind being integrated into every Zorin OS install. :trollface:
On a more serious topic, I want to give you Kudos @Ultrabenosaurus for your awesome response. LLMs are useful and have their place in the world, but I just can't wait for this hype cycle to be over.