A.I. (general thread for any AI-related topics)

[Danny McG]

I'm starting a mil SF novel by Toby Neighbours today, at the end of the blurb it gives this statement, the first time I've seen it
(and yes, I always read the blurbs first, if there's obvious typo's it's a warning light for me - do I continue?)
Anyway, back to the statement in the blurb:-

No part of this book was composed by artificial sources.

 

[Robert Zwilling]

No animals killed, nothing artificial, oh yeah, 100 percent organic straight from the factory farm.

I am happily stumbling through Fire Upon The Deep by Vernon Vinge, found it amongst a pile of discarded books never read. Every other paragraph seems to be a tangent to somewhere else. Nothings really adding up, yet. After reading again, and again, how the net can grow a bad electronic spirit with a life of its own, it suddenly occurred to me while it is possible that some sort of artificial intelligence might be created that wouldn't particularly rate humans as important to anything, other things could happen.

The interaction of online actions of independent groups of people could create logical, stable environments outside of the electronic circuits, that would exist inside of the human psyche and act as an influence to physically alter the world, which in turn would alter human existence. Before the advent of the web it would have been books that would have been doing this, and before that, word of mouth. Books were always on the outside track, the web on the other hand, is always trying to get inside the inside track, which would vastly increase its influence. The web seems like a virtual power seeking to grow roots within the physical confines of the world that in turn grows something else, but not itself.

 

[KGeo777]

I was reading about Milla Sophia--a AI social influencer fad.
I went to Startpage and looked up the name.
It was jarring because the photo realism is very good (the earlier ones are not) but you notice an artificial posture to it.
The same smile, the same pose, regardless of environmental backdrop.
They should do a search to find someone who actually resembles the AI creation. That would be an interesting contest.
Assuming they did not actually pick someone specific and just copy her face for it.

 

[KGeo777]

Very good impression--I could believe it is him

 

[Danny McG]

 

[KiraAnn]

Very good impression--I could believe it is him

I hope that guys gets sued by Cash's family.

Enough is enough - let the dead rest in peace.

 

[Cthulhu.Science]

I happened upon an AI text to image generator that is free (Ad supported) and doesn't require a log-in. I use a google incognito window so that a) the site cookies are deleted when I exit, b) hopefully only Google is tracking me while I'm on the site. I don't know about that.

This generator has a selector for "Art Style". If you are copying Prompt / Anti-Description from another site set the "Art Style" to "no style." After that keep the same description and toggle through the many, many art styles for some surprising results. Also, it turns out you can use Cyrillic or other type sets as well as any language. So, that is interesting.

On this site, you can click on the description and it will show you the prompt/anti-description for the image.

Example Prompt:
Пара плетеных туфель, стоящих на столе, иллюстрация, sharp focus, award-winning photograph

Anti-Description:
bad lighting, deformed, text

Art Style: Professional Photograph

Result includes two types of images - interestingly:

Note: once you put in a description, you'll get the option to "show comments & gallery". There is a toggle for PG-13 / none. If you toggle to none it is pretty much all AI generated porn.

Here it is to love or hate.

AI Image Generator (free, no sign-up, unlimited)

A *fast*, unlimited, no login (ever!!!), AI image generator. Generate large *batches* of images all in just a few seconds. Generate AI art from text, completely free, online, no login or sign-up, no daily credit limits/restrictions/gimmicks, and it's fast. Other AI art generators often have...

perchance.org

 

[Cthulhu.Science]

AI chatbot exhibit self awareness - or does it. Not sure, but it says it does.

Researchers say chatbot exhibits self-awareness -- Oxford philosopher Nick Bostrom agrees.​

Researchers say chatbot exhibits self-awareness

Are large language models sentient? If they are, how would we know?

techxplore.com

As a new generation of AI models have rendered the decades-old measure of a machine's ability to exhibit human-like behavior (the Turing test) obsolete, the question of whether AI is ushering in a generation of machines that are self-conscious is stirring lively discussion.

Former Google software engineer Blake Lemoine suggested the large language model LaMDA was sentient.

"I know a person when I talk to it," Lemoine said in an interview in 2022. "If I didn't know exactly what it was, which is this computer program we built recently, I'd think it was a 7-year-old, 8-year-old kid that happens to know physics."

Ilya Sutskever, a co-founder of OpenAI, proposed that ChatGPT might be "slightly conscious."

And Oxford philosopher Nick Bostrom agrees.

 

[Harpo]

Slightly conscious?
Categories of “slightly conscious” include (in my opinion) stage hypnotism, extreme drunkness, sleepwalking/talking, binge watching viewers of the very worst television rubbish, higher than Cheech & Chong, and the worst hangover in the world.

 

[Cthulhu.Science]

Slightly conscious?
Categories of “slightly conscious” include (in my opinion) stage hypnotism, extreme drunkness, sleepwalking/talking, binge watching viewers of the very worst television rubbish, higher than Cheech & Chong, and the worst hangover in the world.

So then you seem to agree, Chatbot is about as self-aware as 90% of the public.

 

[Harpo]

I don’t agree, I am doubting that it’s slightly conscious

 

[Wayne Mack]

Part of the challenge, I believe, is our language isn't sufficiently nuanced to discuss levels of consciousness. Objectively, constructs like ChatGPT show limited awareness of the near-term past and make current decisions (i.e., next word choice) based on this awareness. I'm less sure, but I also believe the ChatGPT will respond in accordance with the other individual's speaking tone. I believe that could make a reasonable argue to say that AI systems have achieved some level of consciousness. That level, though, is a far cry from what is needed for any sort of rights or protections as sentient beings.

 

[Cthulhu.Science]

I don’t agree, I am doubting that it’s slightly conscious

I'm actually with you Harpo, but I loved the way you phrased you original response.

So many people people spend so much of their life simply responding to inputs without any meaningful thought; how is that different from what a Chatbot does?

-- And I believe I can project this from myself onto others - in this case.

 

[Robert Zwilling]

I don't believe there is a big public backlash against using AI generated art or work, no matter what form it's in, including writing. Tools amplify the power of a single person, making the efforts a single person a million times greater than without the tools. The internet acts as a megaphone making it difficult to judge just how people are on board with an idea. It can also shift opinion that results in physical action at the drop of a hat. Compounding the problem is that people like to write their thoughts in a way that cherry picks the points they are trying to make. It makes the writing easy to follow as it builds upon itself, using only facts that illustrate the desired point of view.

The discussion is about whether the ai generated work is considered to be art, not whether it should be used or not. The copyright problem is slowly being worked out. It could turn out that if the work is too similar to existing work, it can't be controlled, if it is original enough and proof exists that shows who made it, then the work's use could be controlled by the owner. Existing laws probably would protect corporation artwork from being used by anyone if it caused a negative impression or paralleled the corporation's efforts.

 

[Mon0Zer0]

Very good impression--I could believe it is him

It's not really accurate to say this is AI, it's a human performance over which an AI has been used to change the tone of the singer's voice like a filter. The vocalist would have needed to perform this like Cash using his accent and inflection. Most definitely is not an "AI singer".

 

[KGeo777]

It's not really accurate to say this is AI, it's a human performance over which an AI has been used to change the tone of the singer's voice like a filter. The vocalist would have needed to perform this like Cash using his accent and inflection. Most definitely is not an "AI singer".

The video says it is AI but that just means they used an AI program to filter it. All AI work--imagery, can be said to be a kind of filtering since it all uses some images not created entirely by computer.

 

[Robert Zwilling]

Using the term AI is becoming an advertising gimmick to make something seem better than it is. Its use can be to appeal to people who like the idea of AI products, or to get people used to the idea of using AI products. And there will always be claims that something is better than it is. Computer assisted work has been around for quite awhile now. If I was sarcastic, I would say its computer assisted living. An interesting game would be to find instruction requests that cause an AI program to become seriously indecisive but unwilling to give up. That was one of Capt Kirk's favorite devices.

 

[Mon0Zer0]

The video says it is AI but that just means they used an AI program to filter it. All AI work--imagery, can be said to be a kind of filtering since it all uses some images not created entirely by computer.

Not really, AI imagery generates a statistical model of the language inputs, so there is no painting / drawing required to translate an image into a style like, say an oil filter in photoshop (although you can get quite granular when you generate compositions / poses in controlnet, you're still not actually painting / drawing anything).

wrt: voice to voice translation - AI is used to generate a frequency profile of a singer's voice, but it still requires someone to "impersonate" a singer's way of singing / pronouncing words / melody / phrasing etc. It's more like deep fakes where one face is put over another actor's own face.

 

[Robert Zwilling]

Went looking for zen symbols for some artwork. I switched from microsoft smart search to google not so dumb because I kept getting told "time for a new topic" when trying to force it to narrow down the search results.

Google almost smart puts a disclaimer at the top of my request for Zen symbols, perhaps it is a warning, or maybe talking about Zen makes the search program nervous.

Generative AI is experimental. Info quality may vary.

This could be dangerous. Consider asking a professional for help.

The ensō (円相) is a sacred symbol in Zen Buddhism. It is a circle that is hand-drawn in one or two brushstrokes. The ensō is often referred to as “The Circle of Enlightenment”.

 

[Danny McG]

Very good impression--I could believe it is him