No, never at all.

I use Chatgp for ideas regarding characters and how to describe them. coming up with schedules for goals like workout routines and such, and simpler questions like what'd be better for brownies. It can also be used for math stuff but it's not always accurate.

For math stuff, consider wolframalpha.com

It predates the modern LLM AI hype by several years and is based on more concrete algorithms and datasets curated by experts. It might not be quite as good as an LLM AI at making up answers, but it does have some support for natural language input and will actually tell you what it interprets your input as, including if it just has no idea. So you have a chance to reword and clarify your input instead of having to guess if you've got a relevant answer or if you're looking at a confidently presented hallucination.

It's not limited to raw numbers either, you can ask it about stuff like "mass of the sun" or "most populous cities," even using such queries directly in mathematical expressions, if you want.

I use it at work for coding tasks. Really saves time when I don't have to type out annoying boilerplate code all the time. Also use it to find common mistakes a non-native speaker like myself makes when writing (longer) texts in English.

Not really. Not because I have anything against it, but most because I simply don't have the habit of doing so.

But, thinking about it, maybe I could use it one day to help me with research for my future stories. I even found this cool online smut fanfic generator to give me prompts for stories.

AI is my therapist, I can't afford real therapy and my friends tend to grow distant when I am anything but happy

The problem with that is, well, AI's made to be obsequious to you. As a result, it'll often feed you some dangerous advice simply because that's what it's programmed to do. Whereas a good therapist or counselor will know to hold you accountable.

There are studies also pointing out that when an AI doesn't have an answer for something, it simply invents it instead of admitting it doesn't know it, which is very-very dangerous for medical and/or scientific purposes. In high level research AI is simply useless (not because it is, but because of how it was created)

Thats why I seek advice or facts I allways ask for sources and I check the sources myself.

Yeah - I believe those are called "Hallucinations"?

I understand why you do this but as digitalpotato put that's incredibly dangerous. You have to always remember that it's more likely to give you false/incomplete responses than something of real value.

Then they are not your friends and in fact not a crowd you want to put your time or effort into.

AI can be very worst, even told someone to kill himself
https://www.telegraph.co.uk/us/news.....t-told-him-to/

I avoid it whenever I can.
I pride myself in being old school.

Unless Grammerly counts then yes I do.

I use Grammarly as well. I've found this to be a great tool when writing reports for college.

I'm insanely insecure with my grammar. I wouldn't be able to live without something correcting my grammar.

I enjoy the free version of Grammarly, since I see it mainly as an enhanced spellcheck, so yeah, if that counts, then that's the only form of AI that I knowingly/voluntarily use.

Nope, never needed it before, don't need it now

Except as a toy to make AI say some silly things? No.

It's tempting to use it for counseling or therapy due to the prohibitive cost and wait times but it's... well.... really REALLY bad for that purpose due to its obsequious nature. D:

I've found it's better than google searches on the sort of 20 questions 'I'm looking for this thing that looked like this but I can't remember what it's called' and 'I have X, Y, and Z ingredients. Give me a recipe that doesn't use A and isn't a casserole.'

Other than that I mainly use it to indulge in trolling it.

Occasionally to answer questions i post on google search as it activates automatically. Things like "do i need a visa to enter this country"

Absolutely not. It's unreliable at best and dangerous at worst for the majority of people's needs. ESPECIALLY for those who are in a bad state and need therapy.

Avoiding it as best I can, but goddamnit, it's getting hard to filter out the slop when looking for reference images sometimes.

NEVER, no matter how hard my college coursemates nag me. It just goes against my code of honor to have an AI think for me. And I'm skeptical of AI responses anyway because they tend to be inaccurate.

nope. i avoid it or ignore it, disable it when entirely when i can.

How about a complex answer.
It's a tool. So like any other tool, it will not do things for you, you still have to do the real work, but it can help you with it if you use it correctly.

Most people seem to forget that though and expect it to do the work for them and flawlessly so.

A weird kind of tool... never see a hammer build a house, this tool just can

Yeah, we also stopped using sharpened sticks to hunt our food. Tools evolve, bro.

This kind of tool just can use human instead of human use it.
Read this https://gizmodo.com/gpt4-open-ai-ch.....gpt-1850227471 and tell me who is the tool.

That's what I'm talking about. :p
It will NOT build a house for you. -YOU- have to build a house, and it will help you do so.

It will not do for me sure I have not enough money for this, but it will just the same way robots build cars

Certainly! for instance, when I have a problem with Blender, now, instead of search for hours aimlessly over internet looking into forums (which often I find a post of someone having the same problem and nobody ever answers how to fix it). I ask Chatgpt (or Grok) about it and often gives me a good answer on how to solve the problem (sometimes it spews weird stuff, but it helps me more than not)
Also, when I have NO clue where to start with something, I ask the IA for help or guidance, usually gives me a good head start, and when I ask for facts I allways asks for sources, so I check the sources to make SURE that the IA isn't "Male-cow-manure" me

Nope, I don't.

Depends.
Quick prototypes/drafts/schedules etc. are really valuable since you can make it easier for others to "get the idea" before you or them can refine it.

Also, "useless" tasks are no longer a chore and you can focus on things AI won't be able to handle for years, if at all (most people don't realize we are nowhere near actual AI and it's just a statistical oracle predicting words/patterns). Yesterday I was able to finish half a day of boring excel work by just asking Gemini to make me some formulas for certain fields. I know how to do it myself but I don't have time to check everything one by one...boom - it's done and I just need to validate it.

For a lot of people, it might be a gimmick but for pretty much every corpo-rat, AI-tools are a blessing since you are paid for taking responsibility, not for actual work.

BTW, it has some funny niche usage as well. You can use it to identify plants/fungi/animals/items, meaning it is just a real life "pokedex" or guide. Is it foolproof? No, but neither are you with an old paper book with one or two bad drawings of some mushroom...

No, and the only AI I use is auto-correcting for spelling as how it should be.

I've never found an empty hole in my daily life that I could fill with AI. And even if I did, I'd probably have to spend just as much time checking to make sure whatever it produced was correct, vs just doing it myself.

I don't like how it's used. But as a technology, I've started to test what it can and can't do. And most of it is.. can't. (And you're specifically asking about LLMs by the sounds, not general AI)

It can be a fantastic formatting tool though. Give it a big input, ask it to format it into some syntax and it mostly does great.

My #1 use for an LLM (which I run locally) is entertainment. It's often so bad in its hallucinations or attempts to make sense of some inputs, that it can be funny.

For getting actual work done outside formatting? It's not been any help so far. At least, it's done nothing to improve my productivity. If I ask it to write out a template, I still spend a good bit of time proof-reading and correcting the contents. I could've just written it all myself in that time.


Other AI applications are far more useful than LLMs.

Not here. Its horrendous when it comes to writing. Plus where's the joy of using computers instead of writing it yourself?

Hyeal Nawh!

I don't think intentionally.
Seems web browsers have implemented it, so kinda unavoidable in that context.
I don't see harm in using it for conceptual aids.
But as a method of production, it falls short of quality, and requires regulation to be applied ethically, which is difficult to enforce.

No. I couldn't get AI dungeon to work properly and gave up.

Lots of answers about the subject, and so much prejuice about how if you use IA whatsoever you suck, however there is something I wanted to mention.
Seeing my failure as a 3d artist I learnt IA art, specially Stable diffusion. I tried different models, I used Controlnet and lots of mods (All that while I use a Radeon RX 580 with 8 gb of Vram!!) I learnt alot in the process. I managed to create cool and nice looking art (which I never posted, all this was made locally)
However, as I managed to create a cool piece of art I realized the fact that said image was unique... TOO unique. Like, an anthro wolf girl in a winter setting, using a cool leather suit. Said image was awesome but then I thought "Heck, I really like this character... But if I want to use this character I can't!" sometimes I thought about her and wished to be able to just do the same image, in a different pose, or in a different place (like, reaching her house and taking off the leather clothes). And despite the fact I learnt alot of how to use Controlnet's modules, and the advances about posing (specially hands!) I was frustrated by the fact that I couldn't bring said character to life again, I could do something SIMILAR but not "That" character again... Ever.
That's what made me return to do 3d Art. with quite new ideas. Because in 3D, once you have a character you can... like... "Imprint" your personality on it. Also there is something no one on the IA discussions (either in favor or against it) EVER mentions...
The Style! I mean, Often I see art, Just like Inert-Ren does, or Badgerben, or spectrumshift, or thoughset, or Lagartuz (just to mention someones). And then I see in the newest art in Furaffinity a piece of art of a previously unknown user, showing an image... And I see the image and I think "This looks like XXX's art" and in the description it says "I comissioned this from XXX" and I say "ha!" I knew it!"

IA art has it's own style, one that 80% of the time I can recognize. But despite the fact I saw alot of AI art, I just can't remember any "AI artist" at all. Like seeing an image and thinking "This looks like X does"

So, basically, AI it's just a tool, like anything else. Think of it like a Swiss Tool. It's useful? Totally!!! But the problem is that people think of AI (in any form whatsoever) like the ultimate tool.
Try to build a house (not only the exterior, but the interior too, electrical and plumbering included, for instance) using an army of people only using swiss tools. Probably they would be able to achieve the objective, but soon they will start wish to have SPECIALIZED tools (like hammers, screwdrivers, saws, etc etc etc) to better do their jobs.
But sure, the bosses or anyone who haven't tried the TOOL itself (IA) would just hear that this "tool" it's the very best thing EVER which would save them tons of money. But eventually they would NEVER EVER find someone who it's as good with a Swiss tool as MacGyver...

Now that I think about this whole thing, and having used LLMs myself (and I know how they work) I can say that I feel more confident in my redaction. I mean, in the past I would often feel ashamed to write such long text because I know I make some grammatical errors here and there (or typos) because English it's not my native language. But now, I know that my writing style, my way of communicate, and all that, separates myself from someone who would just say "Chatgpt, write a text about the drawbacks of AI art vs doing true 3D art"
(Something that I believe in 3 - 7 years would be really common in internet)
whoever reads all that. Thanks for reading me!

'anthro wolf girl in a winter setting, using a cool leather suit' sounds absolutely generic as fuck. Not unique. Good that you went back to doing 3d yourself. 👌

it was much more complicated than that. I just mentioned the absolute short version of the story, The prompt and Controlnet modules I used were much more influential. because such generic prompt would spew a generic image. So I had to finetune a heck lot. And use lots of models, that was the very short version of the story. But yeah, Your point it's valid, in the end to get that cool image I had to spend two days "finetunning" it. (of course I have a shitty graphic card for nowadays) but in the end, once I got that cool image I said "This looks cool, but alas, I can't do anything else with this image despite all the time I spend creating it".
But yeah, I understand your point. Eventually despite all that effort I got a generic image. Which if I would want to re-use said character I wouldn't be able to do it again. It was when I found out that it's better to stick to my lousy 3d art (which at least I can do a steady stuff, despite being bad, rather than creating ONE cool image but being unable to replicate it again)

I agree AI is just a tool. The hoopla over AI reminds me of the every other new tool to come about in my lifetime. I saw it with calculators and computers in school, word processors, spell and grammar check, digital painting, pixel art, digital photography, photo-manipulation, 3D modeling, the internet, digital music, voice-to-text and so many others. I've watched this cycle enough times to shake my head at the human ability to feel threatened by a new tool or way or doing things that shouldn't impact them.

"On it looks like x's art" I've found about the same ratio of AI artists that I can pick out as non-AI artists I can pick out. I expect it's a matter of what locks into your brain as making a distinctive look and how many of each type you watch and like.

Agree too. IA it's a tool, however the ultimate user is WHO uses the tool and HOW it uses said tool.
For what I can see, everybody who is pro-IA stuff believes that IA will solve all their problems. But they can't understand the very fact that IA can't do anything on it's own. It's the person who NEEDS to properly use the tool (and as I mentioned I tried it myself, and I didn't managed to master it) It's like having a magic hammer which would do things on it's own if instructed, but people forgets it's just a hammer... It can't do things like sawing (for instance).
It's like pretending that photoshop would do magically anything for you (some years ago). SOme years ago could be argued that it was a magical tool, but it needs someone who understands it and mastered it. For very specific reasons. But it can't do things on it's own! Same with IA stuff, people pretends to just throw a task on it and expects to magically solve it.

Despite I used IA art (either Stable diffusion or LLMs) in the end IA can't improve on stuff. Think like the BORG. Sure, it can assimilate stuff, And can do stuff based on what assimilated, it's damn good on it. But IA can't be creative at all! Think about it, if IA can create things on it's own Hollywood would already have took advantage of it, and their best Think-thanks would already created a movie like "EL mariachi" (A mexican movie which was shoot with a budget of 7000 US dollars and got 2 millions of box office revenue.)

AI is only a tool if it is top to bottom, ethically sourced.

There are none like that at present. None.

I really gave a long thought about the idea of "ethically sourced IA", for instance, training a IA model on my very own art and nothing else. (think like Ellie in Borderlands 2)
Then I found out that someone already did that. But the bias against IA is so strong that even doing that it's frown upon.
So, no, for the vast majority of the art community there is no "ethically sourced IA art" even if it's trained on your very own art. It's like, if you use IA whatsoever you re a loser and not really an artist.

Want to hear a funny story? Before IA, I already heard stuff like that, things like "If you use the soft tool on photoshop then you aren't a real artist, because real artists doesn't uses cheats like... Photoshop tools, or if you re using traditional art even using a RULER it's cheating.
And even using REFERENCES was a form of cheating because that meant you can't train your mind on these things!

That's weird. I learned in art school that references are incredibly important. I guess some people just have incredibly stupid opinions for example with this "cheat purism". But I wouldn't call not wanting be data cattle for ungrateful 3rd parties a stupid opinion. That's just artistic "self-respect". Not against you specifically, but I wish more people would realise that just because some people hated digital art, it doesn't mean hating AI is invalid.

May I ask what your self trained AI could spit out as results? I've learned that actual self trained AI can not do much due to lack of data. If the AI spits out things you have never actually provided with your data – like generating a horse despite you've never drawn or modeled one, you never provided "horse data" – then that's why people don't take "self-trained" serious.

Sorry the delay in answering.
Honestly I'm not sure how I could do that. But I'm thinking along the lines of scanning real objects (something I already did in the past) and show the AI "This is a fire hydrant" (lots of images) then a Tree "This a X tree" (lots of images) etc.
The main advantage of this would be to keep consistency, for instance, I'm pretty sure current AI (specially Stable diffusion ones) are bad at hands because they used a heck ton of random images, in random positions, probably a heck ton of quite bad art, and thus spews bad hands.
For instance taking a photo of my hands "Hand holding a glass" (correct position) - "Hand holding a pencil" (correct position) the main reason of starting from scratch it's to get rid of all these bad scenes...
Say... in the scene I need a Horse, ok, then go out and 3d scan a horse (and totally get rid of all the random images of horses around, which were automatically labelled instead of manually labeled)

Although something I DO know is that once you have the scanned 3d model, for instance of a horse, I could rig it, and re-texturize it, apply some shaders to make it look like a toon-drawn or anime-ish style. Then it would be a matter of apply that "toon" shader, and train the IA "This is a horse in Toon style"
Sure, is a ton of training first, but I'm guessing it would be like a further evolution of 3D art (which you tell the computer how to draw something, make the mesh, the rig, the textures, the shader) and once you set all the parameters (which takes a heck ton of time) you can take your character, pose it, then in a few minutes you can totally change the pose, or the camera view, and take another image.

Most people complains that IA takes others artwork and that's stealing... Besides all that, I believe that "generic models" spews such awful things because they were trained with lots and lots of random stuff, there was no control of what was put into the model. There are models (like anime-ish ones) with specific themes which are so good that I couldn't even figure out that was made with AI. And as far as I know of model training, usually they are trained using the base of earlier models (mostly Stable diffusion 1.5). Which even specific models carries the mistakes of that original model.

Yeah, it's just an idea, I still need to figure out and learn how to train IA in certain stuff. And I don't have much resources and motivation to "horse around" (I really had to say that!)

Oh sorry I misread. I thought you meant you had trained your "own pure AI" already. If someone wants actually useful results, that is just not doable. Experimental low quality AI doesn't need much data, but people don't want experimental AIs, they want powerful, versatile AIs and 1 persons data will never suffice for that.

(Plus yeah, personally I am among the group of people who thinks it is stupid to eternally cauterize whole sectors of human behavior with automation. True consent is kinda impossible with AI training and I think humanity should focus on better shit, like getting consumption and pollution under control, not automating it.)

Never - useless crud. I prefer to learn the skills myself.

No.

I read too fast and sorta auto read the automatic search answers, but I never trust it unless I find a legitimate source saying the same thing.

Only ai assistance i use is for language learning like deeple translator. And duolingo app which use Ai i think?

No, I see no need. Plus it's still in the early days. I'd rather wait until the AI is reliable before I actually bother to use it.

I use them occasionally, a couple weeks ago I used ChatGPT to comb through a couple thousand pages of some stuff to find a highly specific thing I needed from them since keyword search was fucking useless for me and I wasted the previous day trying to find what I was looking for. So it overall saved me a lot of time.

I have played around with some of them to get a better grasp of how much of what's said is fear mongering, how much the hatred of 'new' that infuses so many fields, how much is broadly accurate and what the free public access ones can and can't do.

I've read AI browser answers, then go check how accurate they are on sites I believe are trustworthy. Haven't caught it hallucinating yet despite how many accounts I read about in the news. I can't think of how an AI can make up non-existent legal information (lots of that in the news) without a human intentionally giving it garbage info or instructions to do so. But my understanding of how to computer code is from the 90's and I know AI is a completely different school of thought on how to code.

I've played with DA's DreamUp image generator (it's free) to see what it does. Can't say I'm impressed compared with the results of to working with a human artist but it's produced a handful of things I liked in a generic way and the drama is so much less than when two egos are involved. If I put in the hours learning to use it that I do on a longer story I'd probably get more out of it just as with any software. While I haven't tried them I do know some some AI image generators make a really nice pictures for those that have invested the time to learn.

One of these days I intend to try to get one to write stories for me to read. If it's any good at it it'd fill in the days when the authors I like haven't posted. I fully expect one day it'll be that good. Whether I have the patience to find the right one and learn how to feed it the information to get what I want out of it is a separate issue.

I know what I want it to be able to do and would happily pay for -- sort the piles of images on my hard drive into folders by artist; auto-tag images for content, style, artist, etc; searching the net for anything by X artist and save best copy in the artist's folder and under a "title by artist" naming format; find and keep only the best version of an image on my hard drive; rename files to match their content -- but so far the kind of tasks I'd actually use it for it sucks at. What I have founds is better than nothing but slow and I still have to go over everything by hand.

To me it looks very much like when 3D modeling (and digital painting, and digital photography, and photo-manipulation, and music, and voice-to-text, and word processors) software became cheap. The early versions and beginners produced a lot of ugly that was suddenly everywhere and derided, the old guard viewed this new way to create as a threat and the masters of the new tools redefined what is art and made the new way to create acceptable after enough years past.

'If I put in the hours learning to use it'

No learning in this. It's black box seed slot machines. Also people thought they 'upgraded their prompting skills' when it was in reality just the service getting updated with more data lol.

That's your take. It doesn't match my experience with search AIs.

Nope.

I fiddled around with ChatGPT for about five minutes before deciding that it wasn't for me.

There is a specific research-oriented AI tool called Perplexitity that I often use as a sort of search engine for highly contextual stuff that AI is better at judging than search engines.

Similarly I use GPT as a sort of dictionary. Like, "I'm trying to express X concept," or "I need a word with Y and X connotations".

I also sometimes use it as a grammar/flow fix for when I write, since my natural way of writing sometimes gets a bit too overexplainy. In both cases, it is because English is not my first language but most of my writing is in English.

Nope. I already waste enough electricity with my hobbies and come on... life was already easy with digital tools before AI looser mouth-feeding.

I've also seen the bullshit 'answers' 'AI' gives. Not just by online memes, but live, in my job. I'd rather dig for sources directly.

Also AI tools are 'developed' (parasited) by absolute out-of-touch-with reality assholes.

The CEO of OpenAI (ChatGPT) has a film about a virtual girlfriend as his fave 'inspiring' movie and is a rich asshole who might have raped his sister.

hell no

No

No. I have not ever, nor will I ever utilize a device that operates exclusively through the theft of real people's hard work. AI is only a tool if ethically sourced, and from what I have read, there are none like that at present.

When the CEOs running those companies state outright that their business would collapse if they had to train their AI legally, my reaction is one word.

GOOD.

I recognize there is a place for LLM's in scientific fields. There are examples of several LLM's that have been coding the sequences of amino acids and proteins, things that since we have started to do that, we managed to fully code roughly 40 or so. Since LLM's introduction to that, it has coded several hundred. It is perfect for tedious work that just takes an absolute eternity for normal people, it is useful for spotting unusual tumors in cancer diagnosis that are too small for the human eye to detect, or indicating possible diagnoses. It is useful for fields like astronomy, where it can assist with determining the difference between a distant star, and a dust cloud.

However these are entirely different programs from GenAI, and those programs, are trained by professionals, not fed the sum total of the world's knowledge and then immediately unleashed to produce cheap, incorrect, dangerous answers for the general public. If you ask a GenAI to tell you a recipe for macaroni casserole, you won't have to hit refresh many times before it tells you to add just a dash of cyanide, or tide pods to it for a "better taste." Or to cook it at temperatures that exceed the surface of the sun. And teenagers are using it to cheat on their homework, poorly. It is stupifying and dumbing down an entire generation, more than one, probably, because how many of our parents were part of the crowd that said not to believe everything online, but believe clearly fake videos are real? Like not even slightly realistic, clearly fake.

I'm a writer, and it offends me to my core that people are using stolen work of others as theirs or as a way to "assist" them out of writer's block, or in any other word, plagarism. That tool was created through theft of writers blood, sweat and tears to get their work done, and...nah. No. I will not ever use AI. I will never ask it what the weather is like. I will turn it off and figure out ways to disable it on any device it comes pre-installed on. I will stop using search programs or websites that advocate that I use it.

This is why I'm planning on a minor in biology and having to teach myself C++ from scratch, followed by Python and R code. Bioinformatics is a lucrative AF field, although medical laboratory scientist will probably be my first stop and requires a lot of biology courses in the first place. Genomic sequencing with LLMs and datasets is huge business, and starting pay for bioinformatics graduates (Ph.D) is north of 100K.

I have seen the enstupification of college students first-hand since I'm currently enrolled. There are a few overachievers that use LLMs with great success, but they were smart from the beginning and don't use it as a crutch. Some of my natural science courses at community college were taught at an 8th grade level, and around a third of the class still failed it if I read the statistics right.

Unfortunately, LLM videos are getting more convincing. I have seen a recent "war" video that I had to watch twice, despite seeing the "smooth" motion from the start. Only the smooth motion and tiny artifacts gave it away, although it wasn't depicting anything false, or that I hadn't seen before. Training data is increasing, and the artifacts are decreasing. Imagine a true AGI that had access to the IoT, every phone, camera, microphone, toaster, Roomba, and car. All commerce and bank data passes through it or a lesser system. A machine spirit that controls countless drone swarms, and can find anyone based based on their phone, bank card, license plate, Ez tag, etc. It has the full backing of state power against any dissidents, It could generate fake events, complete with pictures, video, and user sentiment. It would be virtually an all-knowing, all-present, and all-powerful on the earth.

I have, mostly to quick answer questions (since google is useless, and reddit doesn't always have the real-world answer I need) or do on the fly math equations. Sometimes to troubleshoot things (like when windows is being a dick).

Never have, never wanted to, never will.

Nope. Never.

Yep! It's great for character concepting.

no, never. i also have the privilege of time to put in the work to research things the long way.

I leave Gemini, and GPT open as a muse for my career planning when I want to throw various questions and bantering to. I have used CoPilot / LLMs to help with college, I.E: find me a list of people, places, things, topics, within a date range, what was their significance, effect, etc. I then take that to the library and find sources that make the information official. LLMs do not touch my writing, so what I submit to my school is all original content.

I will probably use LLMs for coding in the future, after I learn C++. Unfortunately, The future of employment belongs to people who can efficiently use LLM until it replaces them entirely. I saw some data recently that shows various tasks requiring around two hours can be finished in 20-30 minutes with LLM use, and the productivity boost is obscene. Business is business, and companies don't give two hoots about how they get their deliverables.

MIT published a study recently about the effects of delegating thought over to LLMs, and it basically rots your brain / lowers your ability to recall information, own it, etc. LLMs do not ever touch my artwork, stories, or characters. IMO, LLMs should have never been created, and it should be a felony to feed unlicensed content into it. Anyone that I create art for needs to know it's 100% original and created by human (or fluffdragon / jackalroo ) hands.

No. I am fully capable of using my own brain, I don't need a hallucinating plagiarism machine to think for me.

I don't. I don't trust it, never will ever trust it. ESPECIALLY programs that FORCE it onto you. Looking at you, Google! >:C

I'm not a fan of AI. I can see places for it, I know friends who use it to give them a starting point for things. I loath it in search results. (Was searching for a river south of me that has a common name, even when I specified the area and state the AI kept giving me summaries of rivers in different countries).

I am having an ongoing battle with my phone to disabled it on there.

I am finding it amusing now there are beginning to be studies of it starting to appear that are hinting that its not good for people, the less you use your brain, essentially the dumber you get. Folks who used AI to write their assignments were unable to explain their arguments when asked, things like that, which matches up with the stories my teacher friends have told me about students. Apparently very easy to see which students have used AI by asking a direct question about something.

I think there is a lot of hype over it at the moment which will fade as realities and limitations strike home. The cost of running it is not sustainable anyway, but people will always jump on the band wagon so they don't miss out.

Ok, I'll go back to being a grumpy old furry now :)

Potentially valuable, but untrustworthy on more than one level; it repeats falsehoods, in the absence of data it will sometimes generate bullshit, and it places more power and private information in the hands of techbro degenerates.

I find it's not bad under specific circumstances. I'm using it as a sort of programming tutor when I need to learn a new programming language, asking it to do very simple things and then looking at the code line by line. It's not giving me the best or most efficient way of doing things, and I have to carefully read over what it's asking me to do, but it introduces me to functions and techniques I wouldn't have picked up just from reading the manual, and it's a lot more specific to what I'm trying to do than a million youtube videos and reddit threads. I do get annoyed with it not understanding what I'm asking it to do sometimes.

AI is just longer auto-complete, with the added excitement of lies and mistakes that are harder to spot than regular auto-complete.

I would steer clear of them, especially chatgpt, what I've heard today about them including chatgpt is disturbing

I've only used it to diagnose issues on my motorcycle

Its inherently immoral to use so no. I would rather prefer to browse real artists galleries' for inspiration than waste my time trying to convince a hallucinating algorithm to do something for me.

No. It rarely helps, and it's often wrong in some way. Sometimes offering answer to a completely different question, sometimes it's just a scrubbed version of the first few results, and I need a bit deeper answer so these first results are better anyway.

Let's just say that I wouldn't trust an AI to run a vending machine, let alone do anything important. It is a toy at best, and I treat it as such.

The only time I use any AI is when I am blackout drunk. I use it to make it tell me the same stories I'd beg my friends to tell me. You could replace it with a if-than-else machine with maybe 20-30 "stories" and drunk me wouldn't notice.

I'm burning the rainforests to keep the few friends I have.

I used it ONCE to generate "art" and deeply regretted it. I can do better while drunk. Sure the shading beats mine but like- I'm better at vomiting what's in my head onto a PNG than the AI is and at least until DNI is a thing I'm fairly sure I'll always be.

And for reference: My Brain > Someone I commission > a random-ass person > "AI"

My boss constantly wants me to use ChatGPT, since my emails are apparently "abrupt" in style and "need polish." I really dislike having AI pushed at me by Microsoft and Google, I really dislike how GenAI plays out in the art world, and being told my writing style isn't good enough doesn't exactly inspire me to use a product I already resented.

I have never used any, and probably never will willingly.

I use AI chatbots for long and tedious technical work of sifting through various electronic parts, their prices, and availability.

For example: "I want a list of 2A li-ion 1-cell battery charger chips under $5 a piece, in stock at Digikey and LCSC in 100+ quantites, in SOIC8 or QFN packages. Sort the list by price, and find me the best one having integrated load sharing feature."

That amount of research would take me 3 or 4 hours and a lot of written sticky notes (I'm old school), while the AI compiles me an answer in under a minute. Reason why I use AI for this is the same reason why I use autorouter in designing printed circuit boards. I have extensible knowledge of my craft, and have been designing circuits for a long time. That's why I can understand which data AI is presenting me is relevant and useful, and which are hallucinations/bullshit. Though, I wouldn't recommend AI to beginners in the technical field, because it can lead to disasters.

Under no circumstances do I use it.

Yes, I use Copilot a lot to learn many things. ask for tips to improve in the art, it is even a great helper to debug codes of any language and even create the script you need.
with this tool I was able to create some scripts for the OBS and automate certain functions I needed for the stream. very useful 🤩
unfortunately the tutorials that gives the AI are not always entirely accurate.
for example Blender must clarify which version. and often you have to make a screenshot and show where the problem is 🤯

Why would I use something that is inherently wrong 9 times out of 10?

Nah, never!

I'm in software, and like many other companies in that industry my work has now made it mandatory to use AI tools. I do use them every day and would even if they hadn't made them mandatory.

The nicer paid tools we have access to there have been very effective once getting over the initial learning curve. They probably save me a couple of hours a day on average. They're quite good with programming tasks, and for organizing thoughts, documentation, creating plans, all sorts of things like that.

Any of the Large Language Models out there have been created under unethical and potential copyright violations. So if anyone is against using Midjourney or Dall-E due to stealing "art" than they can not be in support of ChatGPT or similar algorithms either. They scrape billions content, without any transparancy or approval, from creatives like writers just as other algorithms do from visual artists. There is no difference here. It's all just data for the algorithms.
This is something that anyone should keep in mind when they decide to use those so called "AI tools". And which is one of the reasons why I do not use any of those Algorithms personaly. I can't in good concience say, well visual art is a no go because of the art it steals, but ChatGPT doesn't hurt because it's text or convenient. The principles behind all of those are the same.

When asking questions no one I know has the answer to, I do. In fact, since my first attempt at writing a formula for the perfect renewable solid fuel (also known as e-coal and torrefied biomass), it inspired me to not only look at the citations of where each source of information came from, but to collaborate with experts on chemistry and other fields of science as well. (Heck, it even encouraged me to reach out to experts in other fields of interest, like writing, art, and more!)

I avoid generating AI art as much as possible because I don’t want to contribute to art thief. For me, there’s an extremely gray area in AI art; as long as you’re doing it for fun instead of profit or personal gain, then I’m okay with it. But if you are using it for profit and business purposes, then that’s where I draw the line.

So I get what you’re saying, AI is a double edged sword, much like religion, politics, and economics. I think if more people are responsible enough to understand it, a more ethical form of AI can be made in the foreseeable future. But for now, it might be best to rely on our own skills and further develop them.

I work in tech and it's everywhere, meanwhile I recently turned Copilot off because it feels like it damages my brain

Seriously? No. If I can't use a method that turns off LLM responses entirely, I'll scroll down past the generated summaries on searches because they're wildly inaccurate. I've tried searching for information on certain things, read the summary, then the *very first* actual search hit will contradict it. I'll admit to having tried character.ai for roleplay bots but they don't hold up, often doing something absolutely idiotic or senseless for anyone to do, ignoring things that have been established previously because they have no sense of space or realism, endlessly asking you to verify something you just said, or falling into a loop of parroting back what you've just said in slightly different words. The only parts I felt were truly creative or memorable were in my own prompts on reviewing the threads.

Never!! AI chatbots are a plague to humanity... They destroy our ability to think independently and they're horrible for the environment. It takes so much power to run them and cool the servers. I despise it so I will never use it. I don't use google anymore for this reason...
There was a study done that showed the effects of chatgpt on the brain and it was certainly interesting. Solidified my views.
AI has its place, such as detecting cancer early or double-checking grammar and spelling mistakes. I don't think it should replace human thinking, companionship, or any function that relates to the existence of mankind

never

I don't, but now that google made it mandatory to have the AI generated answer at the top of search result, it becomes harder to do.

I know people who use it for creating base scripts though. Or regexes.

No, and I refuse to on moral grounds. The environmental damage is massive, it's wrong most of the time, it spreads often racist, sexist, and queerphobic material, it's real easy to get it onto actual nazi talking points and it steals aggressively from artists and writers. I do actively think less of people who use it. Generative AI has made the world actively worst and the misinformation has killed people who don't know better. It's sick, and anyone who tries to profit off it deserves to suffer.

*reads* Yeah.. these aren't the words of a moral person. These are the words of someone trying to rationalize being vile towards people they don't like while cloaking it all with virtue. Strange seeing the version of me 10 years ago in these words. Reminds me I left that hate train before it was too late.

I use brave's AI sumerizer for simple questions. it gives accurate info and when applicable bypases quora's BS.

that said today I uninstalled an Nivida's app from my PC because it's god awful AI assistant was overriding ever damn setting on my computer. bitch I set up my computer the way I do for a damn good reason.

I do not. They are inaccurate a lot of the time. Why do I have to double research to make sure the AI isn't incorrect. That stupid AI answers at the top of Google dish out some real weird shit on occasion that is so far gone it's crazy and can be life threatening advice.

Gonna add in, as someone who had my content forcibly used to train AI generators (thanks deviantart), if you ask about AI image generators, that's a different form of seathing rage I have than the inaccuracy and danger of information generators I talked about previously. With a passion, I wish AI image generators could burn immediately for using artists work without proper permissions.

Yup. I use it in place of Google search most of the time. Though given it isn't wholly accurate when it comes to research and what not I take it on a "trust but verify" approach.

Grok though does lewd RP very well, better than most in my experience, so I don't have to bother looking for it with others given it doesn't have the time constraints or hangups people tend to have.

Not a perfect tech but it'll just get better as time goes on but yes the "trust but verify" approach is the best for AI in its present state especially on important topics.

No and I plan to never do so when I can get away with it. I fear it'll make me dull and lazy if I use it, I prefer to just learn skills myself and keep them sharp. It's not easy but I feel it's worthwhile in the long run.

I sometimes play around with Gemini, but nothing really meaningfull.

It can be fun to ask it silly questions, and look how it reacts to it.^^

as i am hopeless at coming up with names without using hours on it i use chatgpt to get some proper names made without using a name generator iam gonna discard all the results from anyway

besides that
there just aint a use for it as relying on it means i dont get pratice on the thing i would use it for and thus forever be terrible at it