• Due to ongoing issues caused by Search, it has been temporarily disabled
  • Please excuse the mess, we're moving the furniture and restructuring the forum categories
  • You may need to edit your signatures.

    When we moved to Xenfora some of the signature options didn't come over. In the old software signatures were limited by a character limit, on Xenfora there are more options and there is a character number and number of lines limit. I've set maximum number of lines to 4 and unlimited characters.

Anyone else already sick to death of hearing about AI

I'm just waiting for the day when I walk into a restaurant and the waitress asks me if I want to try their new "AI platter."
 
I am really sick about all the really bad art that is being generated by AI.Now any idiot who can type a few simple commands at a prompt thinks he is a great, creative artist.

Seems rather a strange displeasure - which galleries, which museums, which exhibitions are you referring to? Or is it somewhere else you are being forced to view the output of AIs?
 
Last edited:
So far "AI"'s main use case seems to be as a plausible verbose garbage generator.
:rolleyes:
Other than flood risk analysis, drug synthesis, cancer diagnosis, predictive analysis, improved remote medicine, mental health bots, enhanced agricultural yields, smart power grid management, pollution tracking......
All those insignificant little things that most people don't notice.
 
Yeah, as several here have noted, predicitve/recognizant AI does have several tasks it's already very good at and in those applications it will only get better.

But for the generative content stuff it's being employed to do at us, it's mostly seriously crap because nobody who's using it for that is vetting its training materials for crap, and it absolutely is not yet good enough to vet its own training materials.

Not to mention the many low-or-no-human-effort generated webpages that are only there to farm clicks/adviews/clickthrus/browser hijack opportinuities/blackhat SEO macguffins
 
I am really sick about all the really bad art that is being generated by AI.Now any idiot who can type a few simple commands at a prompt thinks he is a great, creative artist.

It's really just a more sophisticated form of stencil, isn't it? Some people will be happy to decorate with stencil art, and it may indeed look nice afterwards if the technical skill of the stencil designer and the practical application of it by the user is good enough. It's not going to be more than merely decorative, and it's not going to exceed in quality actual meaningful art done by a good artist. But if all the user actually wants is the border in their dining room to look nice it may suffice for that purpose.

Sometimes you have Michelangelo on call to do your chapel ceiling, and sometimes you have twenty bucks on a gift card to the craft store. As long as which you're using is appropriate for the task, what's the harm?
 
It's really just a more sophisticated form of stencil, isn't it? Some people will be happy to decorate with stencil art, and it may indeed look nice afterwards if the technical skill of the stencil designer and the practical application of it by the user is good enough. It's not going to be more than merely decorative, and it's not going to exceed in quality actual meaningful art done by a good artist. But if all the user actually wants is the border in their dining room to look nice it may suffice for that purpose.

Sometimes you have Michelangelo on call to do your chapel ceiling, and sometimes you have twenty bucks on a gift card to the craft store. As long as which you're using is appropriate for the task, what's the harm?

That's my call. There is also how generative AI is being added to tools that artists use, for example photoshop. I've been using it's "AI" to do things like to tidy up a picture by using the "remove" tool, it isn't doing anything I can't or couldn't do it's simply quicker (most of the time - sometimes it doesn't work so I have to do it "by hand" anyway) to get a result that I want. I used it to colourise a couple of black and white pictures of my grandmother and grandfather and composited them together - again I had to do "manual" work to get to the finished image but the "AI" did the bulk.
 
It's really just a more sophisticated form of stencil, isn't it? Some people will be happy to decorate with stencil art, and it may indeed look nice afterwards if the technical skill of the stencil designer and the practical application of it by the user is good enough. It's not going to be more than merely decorative, and it's not going to exceed in quality actual meaningful art done by a good artist. But if all the user actually wants is the border in their dining room to look nice it may suffice for that purpose.

Sometimes you have Michelangelo on call to do your chapel ceiling, and sometimes you have twenty bucks on a gift card to the craft store. As long as which you're using is appropriate for the task, what's the harm?
And soon robot Michelangelo will be available for a nominal fee.
 
I am really sick about all the really bad art that is being generated by AI.Now any idiot who can type a few simple commands at a prompt thinks he is a great, creative artist.

I get how some of art through history is just having the idea first.

What peeves me is AI images being shared uncritically as if they are real, amazing things on our planet.

Things like black roses, and huge purple Elephants’ Ears plants, and huge realistic wooden “carvings”.
 
I get how some of art through history is just having the idea first.

What peeves me is AI images being shared uncritically as if they are real, amazing things on our planet.

Things like black roses, and huge purple Elephants’ Ears plants, and huge realistic wooden “carvings”.

Not following you? Do you mean people (I'm excluding scammers and scammer-adjacent click-baiters as non-people) are trying to pass of fantasy art as being something real?
 
Not following you? Do you mean people (I'm excluding scammers and scammer-adjacent click-baiters as non-people) are trying to pass of fantasy art as being something real?

I'm seeing a lot of images on Reddit that are purportedly shared uncritically on Facebook of various "inspirational" images that are clearly generated by AI-- the Reddit posts are mostly mocking those who fall for such images because they feature physical impossibilities in the matter of human limb number and arrangement and other obvious signs that the image is not --cannot-- be genuine.
 
The line is getting blurred. First the image generation is getting better. Second there is a lot of image processing models now, for upscaling and noise reduction, which basically work as AI image generation, except you start with real image and only let the AI add the details. If you overdo it, or start with really crappy image, you'll get all of the AI artifacts, but on real photo.
 
Not following you? Do you mean people (I'm excluding scammers and scammer-adjacent click-baiters as non-people) are trying to pass of fantasy art as being something real?

Friends on Facebook sharing fake nature AI art or fake sculpture AI art and saying things like “Isn’t nature wonderful” or “Isn’t this sculptor clever”. I.e. Believing it’s real.
 
Last edited:
Friends on Facebook sharing fake nature AI art or fake sculpture AI art and saying things like “Isn’t nature wonderful” or “Isn’t this sculptor clever”. I.e. Believing it’s real.

Thanks - get what you meant.

I suppose that is inevitable as it has always happened, I can see how social media makes it more likely. But is that the fault of the AI?
 
The line is getting blurred. First the image generation is getting better. Second there is a lot of image processing models now, for upscaling and noise reduction, which basically work as AI image generation, except you start with real image and only let the AI add the details. If you overdo it, or start with really crappy image, you'll get all of the AI artifacts, but on real photo.

Yeah - this is one area I've been concerned about with my own art. Because of the hardware I have I tend to produce my digital art at no more than 1500x1500 pixels, which is fine for electronic displays but not for printing at any great size.

I've tried several "AI" upscalers and some alter what I've produced by adding details rather than only upscaling what is already there. Thought I'd show an example:

166715025ba088.jpg


The top image is a small digital sketch I created - about 600x600 pixels, the middle one is the "AI" upscaler I've settled on (and then downscaled), it will "add detail" but only such as anti-aliasing upscaled lines to get rid of that stepping effect you see if you simply double up the pixels and I don't think that alters my image, the bottom image however is one of the newer "AI upscalers" - now it hasn't done anything wrong, the image is indeed a render of what I created, but I think that alters the "my" part of my artwork and I wouldn't claim that was my artwork. It's horses for courses, used as a tool I can't see anything wrong with generative AI in artwork.

ETA: hadn't noticed until I was checking it had posted correctly that the "add detail" AI upscaler fixed a few of my mistakes in my initial sketch. Cheeky blighter! ETA 1: But it made a few itself!
 
Last edited:
Thanks - get what you meant.

I suppose that is inevitable as it has always happened, I can see how social media makes it more likely. But is that the fault of the AI?

It’s creating more deception in a world where I think there was already too much deception.
 
LLMs seem to be a collosal effort for very little utility.

Compared to that A.I. systems are doing a good job in medical diagnostics and basic research.
No surprise there: if you have high quality data, you get good results.

Internet discourse is too messy and too biased.
 
Last edited:
LLMs seem to be a collosal effort for very little utility.

Compared to that A.I. systems are doing a good job in medical diagnostics and basic research.
No surprise there: if you have high quality data, you get good results.

Internet discourse is too messy and too biased.

It's part of the bubble (looks at Nvidia's share price), the big players are having to show they too have one of these new fangled LLMs, otherwise they will be dismissed as not keeping up with the pack and their share price will suffer. The issue has been finding the right use for the LLM AIs, and being search engines isn't one of them. They should be UI focused tools or used to rewrite/expand existing material. Which to be fair is how MS is now positioning its Copilot based on OpenAIs LLMs. Now when used as a search tool copilot returns so many references it's starting to look like old fashioned search engines results... Who'd have thought!
 
Back
Top Bottom