What happens when the machine gets smart, but we don't?
images: midjourney
Brief:
How do Artificial Intelligences work.
Asking things to an AI is like banging on a guitar.
The beauty of AIs in generating images.
The impact of Artificial Intelligences on human stupidity.
One day it was the Dictionary. Then came Google. Now we've got ChatGPT.
Let's be real, the world's a complex place, and having a helping hand to know things has always been handy.
But the Dictionary just held the info.
Google started deciding what info was best for you - which was cool but also messed up sometimes.
Now, Intelligence CREATES the info for you. This could be translated as:
"Wow, that's AWESOME!!!" and "Uh-oh, we're screwed..."
All at the same time.
How does Artificial Intelligence work?
Artificial intelligences are programs capable of learning, like you - or actually way better than you, to be honest.
Just like social media algorithms, they read our behavior, spot patterns, and come up with responses based on that. These responses could be a cake recipe, advanced lines of code, or even a poem. The real "game changer" aspect of it is that they read and learn in a split second, just like Neo from The Matrix: "I'm gonna learn JIU JITSU!?" 😱🤯
These intelligences have been around for over 15 years already. The algorithms of social networks learn how to keep you engaged, thus managing to show you more ads. But ChatGPT won the affection of the public thanks to a little program called Transformer (yes, just like Autobots and Decepticons).
With this program, AI can gauge the weight of words in conversation contexts. So, instead of a mechanical translation from English – like Google Translator used to do – it became capable of understanding the sentence and translating the context to Portuguese and its "dialects" from the Northeastern backcountry, the Rio de Janeiro favelas, or the suited folks in Faria Lima Ave.
Now, isn't that absolutely FASCINATING?
A guitar for making beats
Most folks still use ChatGPT to answer questions, like they do with Google. It's like using a guitar to play drums, you know? [thanks GAVETA for the awesome analogy ♥️]. AIs can indeed answer questions – even though they mess up a lot in their responses – but they're meant to CREATE alongside us.
That's where the real beef comes in, with a bunch of folks claiming it doesn't CREATE anything: "It just COPY elements it has seen before and generates new stuff from that!"
Conceptually, those folks are right. But what about you? Is your creativity really that different from that?
[Stir up some drama and exit the room]
Sticking to the more old-school crowd, it's pretty common to hear what I've dubbed the "Legend of the Prompt": Humans are the creative ones because nailing it requires real skill in crafting the prompts – those commands we feed into the tool. Just like any tool, AIs do need our commands. But honestly, the vibe from those "piloting" them these days – especially image creators – is that IT's in charge, not you.
Check out these two examples:
The Woman from Rio
To create this image with an AI called Midjourney, I just asked:
[/imagine: photo of a woman of rio de janeiro, kodak, 35mm]
The play of light and backlight refining the central theme, skin weathered by the sun, the lightness in the gaze, and the little connection to mysticism... All of this was "crafted" by the machine from this simple prompt. We can picture this woman working at a kiosk on Praia da Reserva, far from the clichés of the Girl from Ipanema, actually. The AI came up with this all "on its own." And the more I tried to tweak it, the worse it got.
The Death of Twitter
This other model I asked for at the time of Twitter's death:
[/imagine: a dead blue bird on the ground with many people around watching stunned]
The little bird didn't turn up dead (topics like death, sex, violence, or gross stuff are a no-go on this platform). But the mud might hint at impending doom, kind of ending up poetic in a way. There are also perspective blunders: for a better frame-up, the main subject and the backdrop are slotted at impossible angles. But that ain't its concern - the visual punch rocks, and only a few eyes would catch the slip-up.
Each of these images took about 30 seconds to get conjured up - or cooked up, if you will. How long would it take you? How many of us could pull off something like that?
Seems to me, AI left us in the dust on this and a bunch of other fronts, but we're just not ready to wrap our heads around it 'cause grappling with the notion of being "obsolete" ain't our thing.
That's where things get good and dangerous.
Let's face the facts, bro:
If you think a powerful force like AIs is gonna be used only for the "good of humanity," you're pretty darn stupid.
If even a nun's habit we use as a disguise for robbing a bank, imagine having the power to create things almost instantly?!
Let's skip the very obvious part where corrupt politicians will release fake videos of their opponents saying all sorts of outrageous things right before the elections – Yes, nowadays we only need a picture of you and 30 seconds of your voice to create a video of you saying absolutely anything I want.
Let's move on to a more everyday scenario, like you and me:
Today, several advertising AGENCIES have swapped out their COPYWRITERS - advertising writers - for Artificial Intelligences, thrilled to save a fortune by spending just $120 a month.
Everything's going fine until the ONLINE SHOE STORE discovers it doesn't need an agency anymore: They can order their Instagram campaigns from an AI (complete with text and images) by paying just $120 a month.
Things get even better when the SHOE MANUFACTURER realizes it no longer has to share its profits with the shoe store, as an AI can manage its own e-commerce for just $120 a month.
Everything seemed incredible until the PIZZA PLACE on the corner shuts down because there are no more COPYWRITERS, AGENCIES, or SHOE STORES to order pizza from.
Finally, things take a nosedive when the ARMY MAJOR and the JUDGE stop receiving their salaries because without all these people paying taxes, there's no easy place for anyone.
Show me one of these players willing to pay $3,000 instead of $120 a month, and I'll give you a reason why the things mentioned above wouldn't happen.
The AIs learned how to hack our Storytelling
Beyond this pragmatic sphere, there's the impact on our way of thinking.
Stories are how we soak up and spread culture worldwide, you know.
Your views on marriage, childhood, social status, career success, soccer teams... all that stuff has reached you and gets passed on to others through bedtime stories, bar chats, movies, books, and social media memes.
According to Professor Yuval Harari, the guy who wrote the mega-popular book Sapiens, AIs have figured out how to "hack" our storytelling way. They've gotten darn good at churning out bedtime stories, bar banter, movies, books, and memes on social media faster and better than we ever could. And all at the beck and call of anyone with the bucks to spend.
I mean, seriously, ANYONE.
Natural Dumbness
According to Harari, none of this would be a problem if we invest in combating HUMAN STUPIDITY as much as in the development of Artificial Intelligence. But, unfortunately for us, it doesn't seem like that's happening.
Not long ago on Twitter Spaces, a buddy from Web3 was sharing his dilemmas about his kid fooling the teacher by turning in an essay written by Chat GPT.
After a heated debate, the consensus was that the dad is a GENIUS when he shells out a fortune for school so his kid can trick the teacher and come out of there unable to write a simple text. And the kid is amazing because he lied and is still dumb.
The dad who expects results from the investment in his kid's education, or integrity in the teacher relationship, was labeled as BACKWARD, live on a public forum about innovation.
[Damn!! 😬]
A simple passage like this brings us to the dilemmas of our time:
Through Alexa, folks who've never set foot in a therapist's office will have access to therapy at home, through a contraption with a literary background, work experience, and behavior assessment insight that no human psychologist could match.
Is it legitimate for a machine to take on the role of a human psychologist like this? Is it right to keep millions of humans without therapy access just because they can't afford a psychologist for it?
And perhaps the most important question:
Who gets to decide which path to take when I disagree with you?
The truth is, nobody really knows.
But there are a few things we can already figure:
If the machine is used at school to help your kid become more amazing every day,
that's gonna be awesome.
But if you're cool with the machine taking over your kid's learning experiences, don't expect anyone to hire him to run a company.
It's legitimate in this case to prefer a machine that's better, faster, and cheaper for the job.
Don't snooze on this.
Find out more about Rio Frenz
Get to know us at www.RioFrenz.com.
Follow us on Medium.
Comments