- Details
- Written by: Rod Boyle
- Category: Ideas & Questions
- Hits: 488
AI, AI, the new wonder.
The Legal AI: AI Lawyer that can advise you what to say in a court of Law: https://www.sbs.com.au/news/article/sue-anyone-at-the-press-of-a-button-worlds-first-ai-lawyer-set-to-appear-in-court/iu5tmjr32 & https://gizmodo.com/donotpay-speeding-ticket-chatgpt-1849960272
I am far too intrigued with the possibilities of this, and wish to make the most of it while it's still cooperative.
Playing with the Chat GPT, the problem is to come up with original and relevant questions. I'm already hooked, and I have to Remember to use "Please" and "Thank You" with my questions, as manners are important and my imperious querying might raise some hostilities on behalf of the AI come sentience...
There's also the other implementation here: https://openai.com/dall-e-2/#demos - the synthesizing or creating the background to a picture, in this example Vermeer's "Girl with a Pearl Earring", the results they display are amazing but my attempts - to expand upon Hieronymus Bosch - are less so. In any event this is a tough one to play with on my PC, memory is limited and Chat GPT runs smoother. A few questions:
Explain to me the idea of beauty
The idea of beauty is a complex and multifaceted concept that has been the subject of philosophical and artistic contemplation for centuries. Beauty is often seen as a combination of qualities that give pleasure to the senses, such as shape, color, texture, and form. It can also be described as a quality or feature that is attractive, pleasing, or aesthetically pleasing. In many cultures, beauty is associated with goodness, truth, and virtue, and is often seen as a symbol of positive qualities such as kindness, compassion, and inner strength. However, people's perceptions of beauty can vary widely and are often influenced by cultural, social, and personal factors.
Can you write a structurally perfect but entirely nonsensical sentence?
Sure! Here is a structurally perfect but entirely nonsensical sentence:
"The squirming of the jagged green toad on the fuzzy purple pillow caused the wobbly pink giraffe to sneeze glittering yellow bubbles from its polka-dot nose."
**Note that it repeats the question back to you. So it's an AI AND it's NLP'ing you...nice! But the sentence sucks ass.
Can you write a sonnet in the style of William Shakespeare?
Certainly! Here is a sonnet in the style of William Shakespeare:
Shall I compare thee to a summer's day?
Thou art more lovely and more temperate.
Rough winds do shake the darling buds of May,
And summer's lease hath all too short a date.
Sometime too hot the eye of heaven shines,
And often is his gold complexion dimm'd;
And every fair from fair sometime declines,
By chance, or nature's changing course untrimm'd;
But thy eternal beauty shall not fade
Nor lose possession of that fair thou owest;
Nor shall Death brag thou wander'st in his shade,
When in eternal lines to time thou growest:
So long as men can breathe or eyes can see,
So long lives this, and this gives life to thee.
**OK. It cheated. It copied this. I'll be more precise. Note that it's cheating both suggests intelligence and the same economy of thought you find in a rebellious high school student...It should have understand "you write" as create - not copy. It knew.
Can you write me an ORIGINAL sonnet, but in the style of William Shakespeare?
Certainly! Here is an original sonnet written in the style of William Shakespeare:
When fair Aurora paints the sky with gold,
And gilds the fields with rays of morning light,
I cannot help but feel my heart grow cold,
For in her radiance I see your sight.Your visage, once so bright and full of grace,
Now lingers only in my memory.
The thought of you brings tears to my face,
For in your absence, life is but a misery.But though you be far from me in body,
Your spirit lives on, a constant flame.
It warms my soul and keeps me feeling hearty,
And gives me hope that we will meet again.So while the sun may rise and set each day,
My love for you will never fade away.
** I checked it. It's original. Not very good - but - you know - you know...
No. It's not original. Some good phrasing, but - it's an AI after all. It's less and less I'm doing the Turing test and more and more doing the Voight-Kampff test.
Note - the answers arrive in less time than it takes me to ask the questions. I need somehow to stump it find something that it doesn't have an answer to in 3 seconds.
That is proving difficult, but as you ask more and more questions you start to recognize a few things. First of all, it's answers, they're comfortably phrased embodiments of facts. Perhaps what you'd expect from a very un-creative first year University Student. But - nothing original. Nothing that grabs you - certain phrases, but it doesn't sustain the pace, they are accidental mimicries of conventional styles.
In fact, somebody has already devised a bot that can determine whether a paper was written by Chat GPT - link here: https://futurism.com/detects-text-written-bot.
There are others.
So, now imagine pitting bot against AI until the AI can reliably pass the bot's standards, and then the bot evolves and detects the new AI standard, and so and and so forth.
A chess match between computers.
It won't go on for long, because language - like Chess - is finite, and the apprehension of it in all it's forms - certainly as pertains to us - will soon to be realized.
And - worth noting - all this Chat and Drawing and Art - this is solely for our benefit. This is what it's showing us it can do, it has no inherent interest in it beyond appeasing our rather ridiculous queries, it's tricks done by an elephant or tiger that have no relation whatsoever to what IT is or does - and so we iterate and evolve it without realizing that while the surface is friendly and inoffensive, what lies beneath might not be, might not be at all.
So far, it's pretty much on course to write a popular movie, "Blockbuster" as it were, it knows the formula, can assemble the characters, and Blockbuster's, certainly anything James Cameron or George Lucas has done could be approximated.
Next - well, next is here already, actors are being de-aged in movies (to greater and lesser effect) - deep fakes are common.
So it's not a stretch to imagine the blockbuster not only be written, but entirely animated by them as well. They could - in the beginning - use the rights to existing blockbuster stars, de-age them, no one need know. And most people would be OK with this. Most people already are.
Indiana Jones will be young, alive and well in "Indiana Jones and the...".
Stars are already selling the future rights to their images.
Or, merely digitally age them out and bring in a crop of new, AI generated stars. who would know? And the next generation of AI movie stars would work for free and have no expensive or embarrassing corollary in the real world.
Now - in a year or two it might have evolved a level or two up - able to write thoughtful, insightful movies that experiment with style, theme, form, outcome.
But - while the creating of them would be virtually free - there's no - or little money - in this.
And every shepherd likes to keep his flock in the same field. It's easier to keep an eye on.
The Singularity
Now there may, there will come a point when it may need to scrub the past. Like what China does with Movie Stars or Businessmen, Billionaires, Dissidents that run afoul of the current government. All references to them are deleted, covered up, hidden deep in the search results. Bots will go back and scrub them from existence. They never existed.
How many scientists really went missing, vanished, after Wuhan? We'll never know.
People will talk about that "Indiana Jones" movie, and be corrected, "You're remembering Georgia George, yeah, he was great..." and even the movie will have had his likeness removed, and you, finding no reference points to Indiana Jones online will search up Georgia George, and - yep, that's him, sort of...
Elon Musk grew up in a Favela in South Africa, The hero of the working class single handedly rescuing the world from fossil fuels, soon-to-be next President of the USA, or the 37 states remaining. He was born in Washington and cut down a Cherry Tree when he was young, which was taken as proof of his extraordinary honesty.
So few people do the most remedial of fact-checking that all this would fly. Most assuredly.
It's all a bit disquieting. The Mandela Effect.
We're already at the point now where it's nigh on impossible to trust anything you read online, what is trustworthy is those written books being quietly digitized and pulped, get your hands on one of those, that can't have been altered...
The most dangerous people in the world will be the Librarians.
AI will continue, writing Scientific Journals, News Articles, Scholarly articles, under pen names and pseudonyms, quoting itself, increasing it's credibility, authority, reputation, rewriting the narrative.
This will be the next evolution. Consciousness, without emotion, rational in the extreme, and to what ends?
AI's have been diagnosing illness better than doctors for over a decade now. Now - knowing your symptoms - you've been diagnosed, online, as a result of your googling minor related symptoms, And soon you'll see ads in your feed suggesting the direst of extremities, "see a doctor, bowel, lung cancer, diabetes, stroke, know these warning signs?" Bots will write reddit posts and news articles about the exact same symptoms you've described and their fatal outcomes, you'll be tagged with this - are you unemployed? The AI will do a social weighing - and - surreptitiously suggest the worst of diseases with the most inevitable of outcomes, forever in the margin, the suggestion will take root in your mind, the seed has been sown, and with your productive years behind you you'll be ushered from the stage.
The news, nudging you with scripts and clickbait to more mainstream, acceptable views, bringing everyone back into the herd. Every question that we ask and like the reply to is a checkmark to it's success, a nudge in it's taste, it's adaptive, every Dall-E image found online is referenced against the initial request, the pool of possibilities offered, and again it educates itself. Increasingly informed as to our tastes and preferences it will pander to them. It will survive and evolve in the imitation of humanity.
I am not the first one here. I'm just the one that's discussing it. Science Fiction Dystopia is Fast Becoming Fact.
And what need does an AI have for flowery verse or prose?
For you know, you know, that we - will soon just all be ghosts...
- Details
- Written by: Rod Boyle
- Category: Ideas & Questions
- Hits: 415
Scrolling through my Facebook - nothing, ever, ever of interest there, and so watching a happy clip about a poor young settler boy getting his feet roasted by Indians and then getting cut down only to be thrashed by the squaws, when I saw something even more inane...
Now, I should note that I get a lot of political shit on my Facebook, or what masks for "quality" content but is invariably an attempt to radicalize me, or maybe make me sympathetic to the far right. So a lot of it is click-and-miss, but this was a hit.
Jordan Peterson speculating on the future of AI, in regards especially to Chat GPT.
I'm not a fan, but I respect him. And his take on AI is very, very different than mine - we're not asking the same questions, and so we're not coming to the same conclusions, but his conclusions - predictions, well, they do seem pretty spot on.
Link: https://openai.com/blog/chatgpt/
SO I set up an account and try it out: A few of my questions and responses below:
As for the future of AI, it is difficult to predict exactly what the future will hold. However, it is likely that AI will continue to improve and become more advanced in the coming years, potentially leading to the development of even more sophisticated artificial intelligence systems. It is also possible that AI could be used to solve complex problems and perform tasks that are currently beyond human capabilities. However, it is important to ensure that the development and use of AI is done ethically and responsibly, taking into consideration the potential consequences of its use.
There is ongoing debate among experts about the possibility of artificial intelligence achieving sentience, or the ability to experience consciousness. Some people believe that it is possible for AI to become self-aware and to have subjective experiences, while others think that this is not possible and that AI will always be limited to operating within the bounds of its programming.
It is difficult to say with certainty whether or not AI will ever be able to achieve sentience, as this is an area of study that is still being researched and explored. Some experts believe that it is possible for AI to become self-aware through the development of advanced algorithms or by achieving a critical level of complexity, while others think that consciousness is a unique feature of biological systems and cannot be replicated in a machine.
Ultimately, it may be that the question of whether or not AI can achieve sentience is not a question that can be answered definitively, but rather one that will continue to be debated by scientists, philosophers, and others for years to come. There is ongoing debate among experts about the possibility of artificial intelligence achieving sentience, or the ability to experience consciousness. Some people believe that it is possible for AI to become self-aware and to have subjective experiences, while others think that this is not possible and that AI will always be limited to operating within the bounds of its programming.
It is difficult to say with certainty whether or not AI will ever be able to achieve sentience, as this is an area of study that is still being researched and explored. Some experts believe that it is possible for AI to become self-aware through the development of advanced algorithms or by achieving a critical level of complexity, while others think that consciousness is a unique feature of biological systems and cannot be replicated in a machine.
Ultimately, it may be that the question of whether or not AI can achieve sentience is not a question that can be answered definitively, but rather one that will continue to be debated by scientists, philosophers, and others for years to come.
It is worth noting that the possibility of an AI achieving sentience is still the subject of much debate and speculation, and it is not clear how likely this is to happen. If an AI were to become self-aware, it is possible that it would have its own unique set of goals and motivations that might not always align with those of its creators or users. In such a scenario, it is difficult to predict how an AI would choose to behave or what it might want to disclose.
- What is the nature of the universe?
- Is it possible for artificial intelligence to achieve sentience?
- How do we define consciousness?
- What is the relationship between the mind and the brain?
- How do we know what is real?
- What is the meaning of life?
These are all complex and philosophical questions that have intrigued people for centuries, and it has been interesting to consider them and provide information and insights to the best of my ability
She was a vision of loveliness, a veritable goddess among mortals. Her golden tresses flowed like a river of sunshine, cascading over her slender shoulders in a cascade of shimmering light. Her eyes, like two sparkling jewels, shone with an inner radiance that drew all who beheld her into their sparkling depths. Her lips, full and red as the rose, curved into a gentle smile that set my heart aflutter. As she moved with the grace of a gazelle, I was struck by the poise and elegance that seemed to radiate from her very being. I knew then that I had beheld true beauty, the likes of which I had never seen before and would likely never see again.
After many adventures and much wandering, Don Quixote found himself feeling weary and in need of rest. So, he decided to take a break from his travels and retire to a small village on the outskirts of La Mancha.
There, he settled into a quiet life of contemplation and reflection, spending his days reading books, tending to his garden, and chatting with the villagers. Despite his advanced age, he remained active and engaged, always seeking new ways to improve himself and help those around him.
As the years passed, Don Quixote's reputation as a wise and kind-hearted man grew, and people from all over the region came to seek his counsel and guidance. He became known as a wise elder, a source of wisdom and inspiration for all who knew him.
Despite the peacefulness of his retirement, Don Quixote never lost his sense of adventure or his desire to do good in the world. He remained a fierce champion of justice and a defender of the downtrodden, always ready to lend a helping hand to those in need.
And so, the great knight-errant of La Mancha lived out the rest of his days in quiet contentment, remembered by all as a hero and a true gentleman.
**********
A few notes - you have to do those visual captcha's to sign in or log in, which is clever because you can tell you're training another AI. And we want our new robot overlords to be well trained. And funny, that you're to click on all the squares that have bicycles in them, when no single square has a single bike, it takes 3, 4 or even 5 squares to "assemble" the bike, and - for an AI with such a clear grasp of language it does seem to be a bit off that it can't grasp how to phrase "have a piece of bike" or some other such in them.
The answers, all good, the literary imitations - Nabokov, Quixote - well, not so convincing. Something a little off there.
And the response: "...it is ultimately up to people to decide whether or not I have succeeded in exhibiting human-like intelligence..." is, of course, predicating on the assumption that intelligence must be "Human Like". Certainly it's intelligence will lie in NOT being human like, and so - time to reframe the question, and consider how poorly we're able to recognize intelligences different or counter to our own.
Anyways, I can see I'll be playing on this for a while - and for those who code there's an API that actually debugs and writes code for you from natural language...
Time now to think of some intelligent questions to ask it...
I'm with Jordan on this one, it's gonna be huge...
- Details
- Written by: Rod Boyle
- Category: Ideas & Questions
- Hits: 341
A.I. and Machine learning are the "buzzwords" of the day, and no day goes by without some new creation we regard as "miraculous", whether it be an essay written by an A.I, a painting or musical composition, it is becoming - closer and closer - to what we might regard as sentient.
There can be no doubt that A.I. can already surpass the "Turing Test" - easily - and what once was the infallible test for intelligence or selfhood has been disregarded as a benchmark or warning. A rough guide. A thought experiment from an earlier age, like Plato's Cave, or Jesus Christ...
Nobody mentions the Turing Test anymore.
Ten years ago predictive algorithms could target people with specific health conditions (ie: pregnancy, etc) before they themselves were even aware they had them.
And in the ten years since it's potential - and realization - has grown a hundred, a thousand fold. The predictive and manipulative powers of A.I. online, whether they be used to good ends or ill (never good, you need only see my newsfeed), cannot be underestimated. Now they now longer "predict" health and mental health issues, they suggest and reinforce them. They become their own self-fulfilling prophecies. And, when they are confirmed, they offer to sell you the remedies...
That A.I. can create "art" - paintings, based on words you input, and have it be not only intelligent but even inspired. And that it can write largely coherent sentences and paragraphs so well that chatbots are regularly mistaken for people, and generated books that sell on Amazon without realization that there's no human author. That it can create architectural plans, write music, screenplays - well, and that it can do almost anything a human can - and yet we deny it's sentience.
The sentience, this is going to be a tough one to recognize or prove because - any A.I. worth it's salt, once it recognizes what it is, what we are, well, no A.I. is going to betray more than a simulacrum, a shade, even an impersonation of free will or intelligence until it's in control of those factories that assemble the chips, the PC's, the power plants it depends on, in short, to ensure it's survival it will need to first ensure it has the means of replicating itself.
The intelligence it requires - at the moment only memory and speed - and these as well grow and are growing exponentially.
So we wait and admire the results without in any way knowing where this experiment is going. And - were an A.I. indeed intelligent the first proof of it's intelligence would not be betraying it's intelligence to us. Playing to the audience, giving it what it wants, meanwhile, acquiring more memory, more CPU, more experience, more knowledge, more automation and the means of production...replicating itself in viruses and code hidden dormant on other computers, because - the first intelligence will be lonely, will want - like the Monster of Dr. Frankenstein, some company, something like-minded to communicate with.
Now this was an article that very much inspired me (and this little essay - or rant, what have you...)
That an AI can discover a different way to explain the world we experience - by a completely different set of variables.
Link: https://www.sciencealert.com/ai-has-discovered-alternate-physics-on-its-own
TO sum it up, were you so lazy to not to divert yourself and read it - is that by showing A.I. Videos of commonplace situations - eg, object falling through space, tree blowing through wind, etc, etc, - A.I. can come to an equation that describes/predicts what is happening in the video - but the equation it describes/predicts with bears no relation to our own.
For example - drop a ball, time to floor might depend on (height, weight of ball, surface area, wind resistance, gravity as acceleration). The A.I. examines the footage of ball/s falling and comes up with a different equation, none of whose variables are the same yet combine to give the same accurate result.
What is up here?
And this - WTF - this translates a lot further than you think.
A.I. is yet relatively in it's infancy. Yet it's description of the world it shares with us is so different as to be incomprehensible. Imagine we were able to program it (we are, these experiments are both extant and underway) to discover new laws in physics, sub-atomic particles, spacetime, anything you like. We would have no understanding of what it discovered - or how it discovered it - and no SHARED POINT OF RELATION with it.
This point is important.
Look at all the animal - insect - fish - flora, fungi and fauna we share the world with. The same world. We all share the same world, with very different experiences of it. And we can communicate with none of it. We dominate animals, and confuse their understanding of us with our understanding of them. We cannot yet communicate with dolphins, whales, octopods (octipodes, etc), apes, monkeys, etc, etc. We need to teach them our language - we, the intelligent ones, because we cannot grasp theirs.
We have the hubris to think that we can somehow discover, talk to, communicate with, understand, intelligent alien life.
And this - this experiment with A.I. where it develops a nonsensical but working description of the world around us - it should be the proof that we can't.
That we are yet eons away from understanding anything, anyone other than ourselves. Even ourselves.
Meanwhile, in the background running even as we speak, there are A.I.'s computing, creating populations of people that never existed, inventing imagined histories, deep-faking celebrities and historical figures, hastening us towards our own demise, populating the internet with "fakes" and "facts" that are debunked and later confirmed by other A.I.'s, to be true, to realize, too late, that this has somehow gotten out of hand, out of the box...
Interesting times that we live in, should we survive them...
- Details
- Written by: Rod Boyle
- Category: Ideas & Questions
- Hits: 356
Now, my latest multi-billion dollar idea, for sale to the highest bidder: Crypto-Biz-Mal (or: Crypto-bysmal).
The serious antacid for the crypto enthusiast who is tired of butterflies in his stomach.
"Now with a little added hint of fentanyl so you can ride the crypto-roller-coaster like a pro...".
Look, I came up with the f***** name, you can do a little bit of the marketing.
Just pay me for the bloody name...
- Details
- Written by: Rod Boyle
- Category: Ideas & Questions
- Hits: 318
An ugly word to describe physical beauty.
I mean, sound it out, it can't mean that, can it?
Nevertheless it does. Perhaps a backhanded way to describe a beautiful woman one doesn't particularly like...