FA+
Shop
June Update: Machine Assistance
2 years ago
Whoops, forgot to post one for May, anyway, here's June.
AI - Artificial, Abomindable, Abberant, Augmented Intelligence, Ignorance, Idiocy, Ingenuinty - has been a point of contention since...
Well pretty much since we built machines. Even in the age of steam, when industrialisation first became a thing, people like Jules Verne and H.G. Wells were already dreaming up fully automatic steam robots, somehow given intelligence by little more than gears and pistons.
In 1770, the world was amazed by what is called the Mechanical Turk, an automaton that was capable of playing a strong chess game and could even solve chess based riddles.
Of course it was nonsense, the supposed automaton was little more than an elaborate box in which a real human sat and did all the actual playing.
Nowadays AI, like ChatGPT, is very much the same. There is no true intelligence to be found here, even though it may pretend to otherwise, due to the clear language interface it uses. You can "talk" with it, which instantly humanises it, so it totally gotta be intelligent, right?
Its little more than one of those awful automated answering machines which get so often used by insurance companies, or any other kind of company, who have no actual interest of talking to you. All responses are canned and prerecorded, and it only reacts to very specific answers, otherwise it will break down.
ChatGPT is at about the same level, except that it can respond to more general queries, but it is still little more than an Expert System: able to provide information, but unable to put it into new context.
Stable Diffusion, the art "AI" is similar. It is little more than an expert system which throws color at the wall in patterns that are dictated by its knowledge database. It will not create anything by itself, because it can't. It isn't intelligent, it isn't capable of creating new context. It is merely a tool, nothing more, nothing less.
It still requires a human to create any kind of context, and like any tool, it requires a certain level of skill to do so well. Of course the quality of the tool is also exceedingly important.
Even the world's best surgeon isn't going to be able to do a good job if you hand them a potato peeler instead of a scalpel.
Although apparently there was a surgeon kept as prisoner of war in Russian gulags during world war 2 and its aftermath, which did absolutely miraculous feats of medical care using little more than a pocket knife and strings of wire, but that is a grueling story I don't want to get into now.
You know how easily I get sidetracked, especially once I'm this far in into one of my ~~rambles~~ journals.
Funnily enough, thanks to repeated mosquito attacks to my face, I have absolutely lost my train of thought now.
So uhhh... what was I trying to get at?
In my opinion, the worry a lot of people seem to feel about AI is vastly overblown, in fact, I am very certain that it is Artificially Inflated (hah) by those Silicon Valley tech grifters to make their latest scheme seem more impressive than it really is, and to get people who don't know better to buy into it "for fear of being left in the dust", like with so many things that they produce over there.
Though I'm not gonna argue that it is free of issues either. There are ethical and legal concerns that need to have regulations slapped all over them sooner rather than later. The ability to impersonate people by deep faking their voices and faces is a tremendous danger. It can easily destroy lifes, or even cause national incidents.
But that is humans abusing a tool to get an edge on their competition/other humans, and that is an story as old as time. I'm willing to bet money that the very first stone ever knapped by human ancestors was used against a fellow human within 24 hours.
And yes, this is also coming from me having spent much of this month fucking around with both chatgpt as well as picture generation via machine learning.
This stuff, in its current and near-future state, isn't going to steal any jobs. However, it will make some of them easier, and even create entirely new ones. For example, when properly prompted, ChatGPT can easily assist me in setting the stage for my stories. That way I won't end up blowing all my creative juices on the set dressing, before getting to the actually horny stuff that we all are here for.
I have to figure out how to properly integrate it into my workflow, but it should be a good boon.
For now anyway, until they put it behind a paywall.
That said... oh boy... AI is a topic I could talk about for hours. I do consider it a very, very dangerous technology indeed, though not for the same reason that most others think. I don't think that it will go SkyNet on us and try to purge us in nuclear hellfire.
I think it will instead go full Wall-E on us and coddle us to death. It will make so many thinks so easy and convenient that we will unlearn to do even the most basic of tasks because the machine takes care of it for us, and then one day, when the machine inevitably fails, mankind will die.
There is a Great Filter for intelligent life, and we have not, in fact, passed it.
Its called Convenience.
That will be it for now, thanks for coming to my TED talk.
AI - Artificial, Abomindable, Abberant, Augmented Intelligence, Ignorance, Idiocy, Ingenuinty - has been a point of contention since...
Well pretty much since we built machines. Even in the age of steam, when industrialisation first became a thing, people like Jules Verne and H.G. Wells were already dreaming up fully automatic steam robots, somehow given intelligence by little more than gears and pistons.
In 1770, the world was amazed by what is called the Mechanical Turk, an automaton that was capable of playing a strong chess game and could even solve chess based riddles.
Of course it was nonsense, the supposed automaton was little more than an elaborate box in which a real human sat and did all the actual playing.
Nowadays AI, like ChatGPT, is very much the same. There is no true intelligence to be found here, even though it may pretend to otherwise, due to the clear language interface it uses. You can "talk" with it, which instantly humanises it, so it totally gotta be intelligent, right?
Its little more than one of those awful automated answering machines which get so often used by insurance companies, or any other kind of company, who have no actual interest of talking to you. All responses are canned and prerecorded, and it only reacts to very specific answers, otherwise it will break down.
ChatGPT is at about the same level, except that it can respond to more general queries, but it is still little more than an Expert System: able to provide information, but unable to put it into new context.
Stable Diffusion, the art "AI" is similar. It is little more than an expert system which throws color at the wall in patterns that are dictated by its knowledge database. It will not create anything by itself, because it can't. It isn't intelligent, it isn't capable of creating new context. It is merely a tool, nothing more, nothing less.
It still requires a human to create any kind of context, and like any tool, it requires a certain level of skill to do so well. Of course the quality of the tool is also exceedingly important.
Even the world's best surgeon isn't going to be able to do a good job if you hand them a potato peeler instead of a scalpel.
Although apparently there was a surgeon kept as prisoner of war in Russian gulags during world war 2 and its aftermath, which did absolutely miraculous feats of medical care using little more than a pocket knife and strings of wire, but that is a grueling story I don't want to get into now.
You know how easily I get sidetracked, especially once I'm this far in into one of my ~~rambles~~ journals.
Funnily enough, thanks to repeated mosquito attacks to my face, I have absolutely lost my train of thought now.
So uhhh... what was I trying to get at?
In my opinion, the worry a lot of people seem to feel about AI is vastly overblown, in fact, I am very certain that it is Artificially Inflated (hah) by those Silicon Valley tech grifters to make their latest scheme seem more impressive than it really is, and to get people who don't know better to buy into it "for fear of being left in the dust", like with so many things that they produce over there.
Though I'm not gonna argue that it is free of issues either. There are ethical and legal concerns that need to have regulations slapped all over them sooner rather than later. The ability to impersonate people by deep faking their voices and faces is a tremendous danger. It can easily destroy lifes, or even cause national incidents.
But that is humans abusing a tool to get an edge on their competition/other humans, and that is an story as old as time. I'm willing to bet money that the very first stone ever knapped by human ancestors was used against a fellow human within 24 hours.
And yes, this is also coming from me having spent much of this month fucking around with both chatgpt as well as picture generation via machine learning.
This stuff, in its current and near-future state, isn't going to steal any jobs. However, it will make some of them easier, and even create entirely new ones. For example, when properly prompted, ChatGPT can easily assist me in setting the stage for my stories. That way I won't end up blowing all my creative juices on the set dressing, before getting to the actually horny stuff that we all are here for.
I have to figure out how to properly integrate it into my workflow, but it should be a good boon.
For now anyway, until they put it behind a paywall.
That said... oh boy... AI is a topic I could talk about for hours. I do consider it a very, very dangerous technology indeed, though not for the same reason that most others think. I don't think that it will go SkyNet on us and try to purge us in nuclear hellfire.
I think it will instead go full Wall-E on us and coddle us to death. It will make so many thinks so easy and convenient that we will unlearn to do even the most basic of tasks because the machine takes care of it for us, and then one day, when the machine inevitably fails, mankind will die.
There is a Great Filter for intelligent life, and we have not, in fact, passed it.
Its called Convenience.
That will be it for now, thanks for coming to my TED talk.
WALL-E is just the Di$neyfied version, a kid-friendly version of Skynet.
What we need is to make robotics symbiotic with humans.
Getting a proper symbiosis going might be difficult or even impossible, though it highly depends on how this AI finally comes to be as well, from what foundation it emerges, the military or civilian sector.