https://social.trom.tf/display/dbc8dc44-1464-2853-620a-b69239507976
Out of all the things I've seen and read about AI so far I like this one the most. This is a more balanced take on AIs in my opinion - don't hate or love it but just understand how it works and continue to have a healthy amount of skepticism.
@tio I see you haven't lost your skill of coming up with witty acronyms, haha 😄
AI is ASS
NOTE: This post contains many images and text formatting, if you do not see this via our Friendica page then please use this link to read it fully https://social.trom.tf/display/dbc8dc44-1464-2853-620a-b69239507976Back in 2015 I wrote a book for TROM about humans and the machines they have created. It is called the Human Machine and you can read it here.
Back then it was one of the first hypes about A.I. And as you may know by now, AI stands for Artificial Intelligence. Back then Google and Apple started to promote their AIs: Google Now and Siri. You could talk to these bots and they will provide you a lot of answers from many different domains. Ask how many dogs are in the world, ask them to set up an alarm for you, how's the weather, you name it. It would integrate with a lot of apps and devices.People were asking these bots existential questions like what is the purpose of life:
I remember the Google supercomputers generating these weird images and the news outlets calling them "machines dreaming". Meaning, how machines dream:
And that Pepper robot that you could talk to...
These came after a few years of another hype: the self driving cars that were about to revolutionize transportation. In the book Automated Autonomous World that I wrote for TROM in 2013 , I partially fell for that hype and wrote in a very non-skeptical way about them.
To be fair with my past-self, I was focused entirely on how in a saner society the tech-part of the world could look like. So I presented many prototypes and unproven tech. I talked about how we have to change the infrastructure for these to work, and so forth. However, what I did not realized as deeply, was how misleading the world is. You see I naively thought that the big tech companies may not exaggerate as much about their tech. I sourced their own claims from their own engineers. But 10 years later I could say for sure that the vast majority of the tech I showcased in that book, was BS. For the most part at least. That means, the words that came out of the mouths of these engineers and "specialists" that work for these big companies, were mostly crap. Marketing.That book may be removed from TROM when we do a cleanup this month in preparation for the upcoming TROM II documentary.
Back to AI.
In 2016 I started to be more and more aware of how the tech world is full of bullshit and bullshitters. For that reason I chose to present tech that was already there, working, and when it was just in testing phases I would present it as such and link to scientific papers about that tech.
When it came to AI I already smelled the bullshit. The stupid fear that AI will take over the world, and all of that crap. But I said let's educate myself. And so I said the best way is to understand a bit how these things work. I took several courses and read a bunch of stuff. I learned how AIs can categorize objects:
How it interprets text and speech:
I learned how facial recognition works:
I looked at how AIs learn to play games, drive cars, do science research.Of course I only scratched the surface of these since they are complex. But I got a solid understanding I think. I realized that these are Statistical Software. Added with a bunch of filters and rules. They are complex, cool, interesting, but nothing WOW in terms of how they work. I learned that data and computer power is what makes the difference. I realized how non-magical these AI look when you have a clue about how they work. The end result is mesmerizing only for those who have a very dim understanding of these pieces of software.
I then talked about how they compared to humans. And of course the conclusion was that AI is just a complex set of statistical software with filters and rules and a lot of data, and has nothing to do with either artificial or intelligence (whatever that means).
Of course there are some applications for these pieces of software that I touched upon in the book, but mainly it was a grotesque hype.Anthropomorphizing
That's a new word that I learned while writhing this book. It means "to ascribe human form or attributes to (an animal, plant, material object, etc.).". That's what AI are mostly.
The present.
7 years later and the AI "revolution" has resurfaced again. This time it seems even more mesmerizing, although not so much for me. I fee like I've seen this before...
ChatGPT and the like, Stable Diffusion, DALL·E, now they have different names and do similar things. They are fun to play with. Can create some weird images out of text, something I was showcasing in 2016 in another book about Language.
And the craze is starting again with people thinking these AIs will revolutionize the world, or will take over, and then no more jobs, no more artists, no more life. If you look again at how these work, they seem a lot less mesmerizing. The companies that push them use anthropomorphism to make them sound like they "think" and are like humans: neural nets, learning, hallucinating, and so forth. In fact they are still the same statistical machines that produce probabilities and that results in interesting outcomes.
Without going into specifics I recommend this funny video that explains the entire situation quite well: A.I. is B.S. ->>> https://ytb.trom.tf/watch?v=ro130m-f_yk
Also read through the sources of that video.
When I asked chatgpt about TROM I got this:
Which is not bad.
But when I asked who made TROM I got a very bad and totally wrong answer:
No idea who Catalin and Andrei are, but it is totally wrong. The project did not start as a research project, whatever the hell that means.You see although it says some true things, it also says some entirely false things. Because this AI only knows how to put words together. Has no clue about anything else. And it will always produce well written sentences that can be totally wrong. Thus, this chatbot is entirely useless for factual information. And if you think it will get better and better, for one let's wait for that and then "hype" it, but second it may never be able to due to how it works (statistical prediction of what words go with what other words).
When I tell chatgpt that the project was not created by those two dudes, it invents new ones:
And so on it goes:
AI is ASS!
AI is actually Anthropomorphized Statistical Software, or ASS :)
We may find good uses for this ASS, maybe when they only rely on a main source of information like some specific research papers and provide exact quotes and sources. They are fun to play with for a bit tho. They may help us create more creative content or video/photo/audio effects. But "they" are as much as "they" as websites are "they". Depends what websites we are talking about. Same here, these ASS are diverse and made for a diverse set of purposes.
But what it is clear to me is that the HYPE is just that, a HYPE. Pushed by companies who want to make these ASS more appealing and sell them, by the bloggers and youtubers who jump on the hype train for clicks and views, by the media that does the same. Combine that with the fact that we live in a dumbed-down society where humans are just workers who have no time to learn/digest any of these things, and we are faced with an Idiotic society that has no clue about any of these.
Our trade based society is at fault here again, for incentivizing humans to hype these things all the time, in order to trade them. And for the fact that it keeps us all busy idiots for the sake of trading to survive. This has created a Land of Confusion.
Welcome to Earth.
Tio
in reply to Rokosun • •like this
Roma and Alexio like this.