General AI Discussion

You could’ve said the same about governments, cameras, GPS, internet, cookies, cellphones… Now AI.
I think people being lazy and giving away their rights and liberties is the issue, not the technology.
Every big tech is disruptive, people make choices.

We collectively choose to let « artistic expression » be more important than privacy, now anyone can take a picture of you, diffuse it virally on the internet and even sell it, if the picture of you was taken in public.
I think this is a dumb choice, I don’t agree with it but people choose to accept to live in a world like that…
That is just an example.

I feel the problem is people are generally uneducated and lazy most of the time.
Organized power like governments or private companies are not.
Efficient private companies are specially fast at reorganizing themselves and thinking of new strategies to make profit out of the lazy and dumb mass of people that we are.
They kinda define and push for a certain legal environment for these new techs to evolve within.
If a decision is made, people protest a bit in the beginning but most of the time give up and accept « the new norm ».
Individuals rarely take the tribunal way cause of lack of money and time it takes, it’s hard to compete legally with giants like Google for example.

Before, people had to buy music. If someone wanted a bootleg, it was most of the time not as good quality as the original. People had to listen to radio and hear ads and ear other artist before hearing their songs. Artists had to be « discovered » or work very hard to make it.
But at least artists were making money selling records locally.
The mass of music consumers might complain a bit it wasn’t perfect but in general people were happy and enjoyed music.
Then Napster came out and suddenly average joe has limitless instant access to any music produced in the world for free.
Of course a bunch of average Joe’s started to really enjoy that new tech and private companies saw the opportunity.
Only few people paid their MP3’s or bought original CD’s, they stole it from artists for a while. Artists were in fury, some losing a big chunk of their past revenues.
Some laws were slowly made to punish people from stealing music but application was and still is very lazy.
Instead of stronger regulation, companies like Spotify found a way to give lazy people what they want and found a way to make artists give up their rights and get a small fraction of their profits to get heard and « distributed ».
This can be undone but once people are used to a lifestyle and are addict to their dose dopamine, in this case easy and cheap access music, it is very hard to go back.

Uber was also a case of the same phenomenon.
It was morally and economically bad but lazy users loved it. Who cares if they don’t pay taxes, if they provide their service at loss to destroy competition ( and rais their prices once they destroyed it ) and if it’s less regulated, people love it!
Governments tried to hear taxi drivers and fight for them but the mass of people, the lazy one, loves it! That same mass also vote for those same governments… They better listen I guess, if they want to keep their job.

So now generative AI.
The same thing, lazy people loves to generate art, it’s fun, easy, it works well, it’s the future!
Again, big private companies already define the legal aspect in courts with big bucks to prepare the terrain for their future market.
Artists and philosophers raise the redflag.
I don’t remember average Joe’s complaining about the lack of generative imagery before and graphic artists were doing just fine.
But now average joe’s tasted generative AI and now they are addict to their new toy and they want to jeep it and they are gonna fight for it.

It’s ok if the mass collectively choose the lazy way, the « progress », again.
Too bad for hard workers, for the traditional way of doing things for centuries, they are annoying anyway and when you think about it, a computer storing a copy of an image in it’s memory is the same as a human having a blurry memory of it right?
That’s a convenient argument and it’s as simple as that right?
Well, the further we go this way the lazier we will get.
Hopefully there will always be curious humans inspired by hardworking and understanding the importance of doing things themselves but as a social specie, we tend to mimic the mass to feel accepted.
The more we go the lazy way the lazier we will get.

If only you would get off your high horse you would hear a lot of people that were complaining about lack of time, talent or being to tired after doing exhausting jobs to learn how to draw.

1 Like

But remember that typing a prompt into an AI also doesn’t teach you how to draw.

2 Likes

but it allows all those people to express themselves in a way they couldn’t before. For many it’s not the act of moving a pen that matters, but what comes out at the end of the process.

2 Likes

You wont be able to express yourself in ways that the AI won’t allow you - ie hardcoded words or phrases that are blocked or will get you banned - now or in the future and you dont get any sense of accomplishment from improving over time

What is more important?
People with exhausting jobs hobbies or an artists having a job?
The tool you hard working hobbyist will not only be responsible for the loss of an artist’s job, it might also literally use that same artist’s work to generate the art of your hobbyist…
I think it’s almost comedy how pathetic it sounds.

If your hard working hobbyist don’t have time to learn to draw, maybe he could at least pay a small amount of money that would go to the artists that made his AI generated creation possible, don’t you think?

That’s exactly the high horse I was talking about.

@manuDesk What’s your point of view about AI tools which datasets have been trained on Royalty Free / open licensed assets ? (which are not the most common, I agree).

I think it’s cool tech and an amazing tool.

It seems inevitable too.
I wish humanity would globally make a moratorium on tech and take time to think about where we going as a specie but it is not where we are going it seems.

As for AI-generated images using copyrighted art in their dataset, I just hope it would be super in favor of artists, to the point that it’s almost as expensive to create an image by an AI than it is pay for an artist.

Even if tech is cool, I think at some point humans will have to analyse more carefully their relation with tech and how tech can affect them sometimes in a bad way even if at first sight it can seem like only beneficial.
A bit like some species or environments are protected, maybe some skills or jobs should be protected too, even if they can seem obsolete.

1 Like

I don’t see the high horse thing you talk about.
I think your argument about hard working people that have no time to learn to draw was comical and poor but I still made effort to answer it seriously.

I am quite pessimistic about the future of humanity. Capitalism will always pull strings of puppets (politicians and CEOs), those themselves sitting on the shoulders of ordinary workers. The latter are not always given the chance to decide what their future lives will become.

You can download stable diffusion and this version doesn’t have any restrictions. So there are ways to type whatever you want.

There are not just fully automatic ways, like going from prompt to image. There are plenty of opportunities to for instance to provide those generative AIs with compositions or different kinds of inputs. There are ways to make the workflow more goal oriented (at least there are papers which demonstrated that this is possible).
I have no doubt that you can easily get a sense of accomplishment when using those and that this can also be a different way of learning art in a quite intuitive way, with never before seen iteration times.

2 Likes

I’m not just talking about stable diffusion, im talking about AI in general such as chat GPT.

That’s sad.

See LLaMA for open source version of that. Not as good yet (but getting there), but you can fine tune / adjust / tweak as much as you want.

There’s some heat sneaking into this discussion- please keep personal insults and comments removed and focus on the issues instead of people, thanks :slight_smile:

2 Likes

If I could use an AI to make me memorize and draw with a pencil an anatomy and muscle chart in different poses so that in the end I don’t have to use a computer / AI at all and draw everything from memory and do it faster than it would take me to normally study out of various anatomy books, I would be more enthusiastic.

You can replace drawing with any activity - playing an instrument, playing a sport - but in the end I want to be able to (permanently or semi permanently) unplug from the AI and still retain the training. I don’t want a permanent crutch. I definitely have not seen any evidence of this floating out there.

As @VertexMachine mentioned, there are different language models, including open ones which are not restrictive.
Personally, I don’t think they are very useful yet, but allow you to tinker a little bit. But I am quite confident that in the future for instance Blender artists might be able to type what kind of script or geometry nodes setup they want and this could be generated. There might not be a massive feeling of accomplishment, regarding how the goal was achieved, but it might help artists to achieve something that would otherwise be very complicated or time consuming.
I even see that pure language models could be useful (in the future when they work better :slight_smile: ), for instance to overcome writers block, or to pick out flaws or inconsistencies in stories.

Some newer models allow text and image inputs and can output both as well. In my view, it is not unrealistic to expect that those could be used as some sort of assistant. Just as in the example you mentioned, where you might first describe that you are looking for some guidance or input, in case you forget something. You could then have one or more drawings at different stages or whatever makes sense in the context and you might get feedback, what you forgot and so on.
It certainly still takes some steps to get there, but everything is clearly pointing in that direction as far as I can see.

Personally, I am very enthusiastic about those tools regarding education, as just described. To me, there is potential to achieve more personalized education.

It does not exist yet, but I have no doubt that once those systems become better, that they will even be specifically trained in that direction.

1 Like

I believe that dedicating time to developing proficiency in English language skills, like expanding the vocabulary and refining communication abilities can unlock significant intellectual potential. Practicing good language skills are gymnastics for the brain. No debate will be hard to follow again. :sunglasses:

Furthermore, critical thinking as a basis, which you have been showcasing, is a cornerstone of intelligence. :wink:

It was free will by me to quote this comment for no reason :slight_smile:

Humanity have a choice indeed, the question before us is one of paramount importance, what constitutes the ultimate goal for humanity?

  • Should we allow machines to perform all tasks while we do nothing, how would it affect human well-being, and what are the potential risks of becoming spoiled and taking everything for granted?

  • Should we embrace responsibility and actively engage in various endeavors and earn our own rewards?

1 Like

Can you prove it though? :eyes:

The thing is that there is no single entity called ‘humanity’ that will be making the decision. Each country/jurisdiction will make this decision (or decide not to make it) for themselves. And within the countries there is rarely unity, esp in controversial topics like that.

And it will take quite a lot of time…

Unless we get Skynet-type scenario at some point, then we will unite (mostly) against the machines…

1 Like

I’d call that a very optimistic view, and I reckon it’s mostly shared by those who fare well with the status quo.

That’s kind of an inverse conspiracy theory, isn’t it? A group of very intelligent, insightful and selfless people with super powers, who happen to be very rich, working behind the scene, for the Greater Good. I’m totally sure you’re right. :wink:

I believe what we are in fact witnessing with the development of AI, is the most spectacular Dunning Kruger effect manifestation mankind has managed to bring about in its history so far.
In order to “make sure something is used for good” one needs to understand both mankind and that something, Unfortunately, while these baby-faced tech visionaries are convinced they do, in fact they understand neither.