/lounge/ - Lounge

Yup, I sure do.
But AI doesn't seem to be sentient yet, let's hope it stays that way!

Humans have a tendency to create the technologies that people have been writing cautionary tales about for 10 trillion years.

>>325701
No. I honestly don't think about this at all. I've never even used ChatGPT or anything like it, and I am completely uninterested in whatever direction the technology is going in. It's intellectually lazy, makes you retarded, doesn't even do a good job at what it's supposed to do, and I'm not even convinced it's a real thing.

>>325702

I have mixed thoughts about whether the current methods will ever produce true intelligence/sentience. Some of the things they are doing with attention and whatnot mimic the natural processes that the human brain goes through to produce thought, so it's not inconcievable, but I think it may be that they're barking up the wrong tree and may never achieve it through these methods. Intelligence and sentience exist in the first place because of the process of natural selection. I think that in order to produce true intelligence they will need to simulate the process of natural selection and create pressures that produce intelligence naturally. Essentially I think they will have to create little simulated versions of the world and grow alien species from the beginnings instilling basic instincts first, then having them evolve and compete over time and eventually intelligence would be produced. Then you can take your little simulated aliens and give them senses in the real world. As long as you make your simulations close enough to the real world in the first place, they will then be able to act intelligently in the real world.

In a scenario like that there is no telling what could go wrong. You could introduce humans and english language to them with your role being that of a God since you can control everything in their simulated world. Then we would basically have created little aliens that worship us as Gods but then if we let them out they would quickly realize we're not Gods. Possibly even before that if we just explain things to them and start giving them sensory data slowly.

Problem is if we're designing them to be more intelligent and capable than us that could go really badly in the obvious sci-fi kind of ways.

>>325703
Assuming you refer to the image generation, try novelAI then.

* You feed it tags,
* you feed it one "artist:(name_here)" tag atop of it
* and add "lowres" and such into the negative tags,too

Which is, in my opinion, beats CHatGPTs average style of generating "nice things"

Capre diem.



>>325704
>Problem is if we're designing them to be more intelligent and capable than us that could go really badly in the obvious sci-fi kind of ways.

There's still a frontier to hold onto: a healthy human body is "more energy efficient" from the point of Moravec's Paradox: a human body is incredibly well-optimized when it comes to use tools and sweat, as opposed to forming thoughts that make ideas. And a healthy human body's, well, regenerates; as opposed to maintenance-requiring springs, cogs, gears and servos.

>>325701

civilization is just cruel people using violence and slavery so a few violent psychopaths get their way and everyone else is expected to participate under threat of violence.

of course ai is supposed to be powerful people's slave, everyone is. in the system of violence there is only slave and slavemaster.

>>325708
>try novelAI then
I just said I'm not interested in it you tard

The name "artificial intelligence" is just a marketing trick they used, inspired from sci-fi to arouse people's curiosity.

All the AI stuff like chatGPT and Gemini, its just some dumb algorithm scrapping google and wikipedia and mindlessly rephrases the information it finds. This thing is not intelligent at all.

>>325714
I really don't understand how you and other people still believe current AI models are just stochastic parrots.
Frontier LLMs are currently producing novel insights in long-standing previously unsolved problems in mathematics and other fields, and sometimes manage to solve them autonomously. I don't want to say that current AI is truly intelligent but the assertion that they just mindlessly rephrase stuff is flat out wrong.

>>325714
i once heard someone say it is basically just autocomplete like google tries to guess what you are searching for based on what you already typed

what philosophically disturbs me is how many 'entities' write or say "lets be honest".

they all say "lets be honest" but then when they are done being honest they never say "ok lets stop being honest and be liars again".