Discover more from Jotham’s Substack
Rabbit Hole of Research
Will The AI Revolution be Televised?
So, What Exactly is AI?
Math. A lot of math but linear algebra, probability and algorithms are the biggies. Linear algebra is a mathematical discipline that deals with vectors and matrices and, more generally, with vector spaces and linear transformations. And probability is the mathematics concerning numerical descriptions of how likely an event is to occur, or how likely it is that a proposition is true. Algorithms are procedures, descriptions of a set of steps that can be used to solve a mathematical computation. And then throw in some multivariate calculus, algorithms, and things like topology, metric spaces, single-valued and continuous functions, limits, cauchy kernel, fourier transforms, information theory, function spaces, manifolds, and ... Okay, okay, that’s all the math talk, I don’t want you falling asleep like you did in high school math classes.
But How Do You Get From Boring Math to a Chatbot or Art Generator?
Well like most things the devil is in the details or the coding. Let’s spend a little time understanding how a computer thinks. It’s all binary. On, off, one or zero. A base 2 system. And for reference us humans work in base 10: 0,1,2,3,4,5,6,7,8,9. So, binary keeps things simple and processing is easier. But, this means we have to figure out ways to convert our world into a binary state of mind.
How Do Ones and Zeros Become a Portrait of Fluffy the Cat or My Essay That is Due Tomorrow? Note: Please don’t use AI to write or draw your homework.
But how does the computer look at a picture and figure out what it is? Not going too much into the math, but when you have a digital image it is represented by a collection of pixels.
Pixels are the smallest element of an image. Each pixel correspond to any one value. And in the simplest case, a gray scale image, the value of the pixel is between 0 and 255. The value of a pixel at any point correspond to the intensity of the light photons striking at that point. This will create a matrix of pixel values or number that can be used by the computer or in math. This is the bases for computer vision.
Okay, the computer can see numbers, but how does it know those numbers is a cat and not a dog?
I’m sure I mentioned this, but it’s because of math. LOL. Seriously, you have to train the computer with hundreds or thousands of training images and it “learns” the values that make an image of a cat. This is know as machine learning. By learn I mean math, it calculates the probability that a random image has the same values as the images the machine was trained with. FYI, speech to text and natural language processing works the same way. Using human input to train the machine, and math to predict.
Sounds Like Machines Need Us Around
Exactly, so how the heck do we go from training the machine to “I’ll be back!” Singularity. This is a point where a technology’s growth is out of control and irreversible. So, what does this look like? Like we talked about, machines are still reliant on human experiences and creativity to create anything. They are taught with human input. Our speeches, our images, our music, our music, our... okay I sound like a whiny toddler. But seriously, when an app creates an image from text, or an essay, it uses other people’s images or words and tries to create something new.
But, in simple terms, Singularity is where the machine gains computational creativity and emotional or affective computing on its own without a human prior.
”I Always Trusted Code More Than People.” (Westworld, 2018)
This would be the point where the machine would start writing its own code and learning from its own experiences. So, is this when the machines try to rid the Earth of humans?
Probably not. Most likely, machine will not be angry and still serve in the best interest of humanity, but humans versus machines makes a good story. Though the conflict these stories explore what it means to be human even as the machines try to Terminate us or use us as batteries as in the “Matrix” all in the name of serving humanity.
Yet, if it came down to a Machine versus Human death match for supreme overlord—I’m betting on the Machine. Sorry fellow humans, but like C. Robert Cargill “Sea of Rust”, the war would be over before it starts. Humans need water. Machines not so much. What can we do, start building bunkers and storing water like Sarah Conner.
“We Don't Need AIs To Destroy Us, We Have Our Own Arrogance.” (Exmachina, 2014)
Or we can look to Issac Asimov’s Three Laws of Robotics:
1) A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2) A robot must obey orders given it by human beings except where such orders would conflict with the First Law.
3) A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
With these laws you get a kinder gentler machine exploring their newfound “humanity” like in “Bicentennial Man”. Where the machine gained human creativity and emotion.
But then again, like in the “I Robot” series and the movie these laws can be more logically interpreted and the Machine will fill obligated to enslave humans to protect them from their own violent tendencies.
So, Are You Saying I Should Start Building My Bunker, And Hoarding Water and TP?
I guess you can never be too prepared, but to get from ChatGpt to all out war with Ultron and the Terminators we have some amazing hurdles to jump. Could it happen, maybe. But probably not likely. The Machine would have to do more than Math. Believe it or not your brain is amazing at Math. You do so much math and not realize it, and we use less energy input to do our calculation.
Think about playing trashcan basketball. Don’t be like that, I know you’ve balled a paper up and tossed it at a recycling bin, and if not, Do it! Before your brain tells your muscles to release that balled up essay 1st draft, your brain has estimated the distance and weight the paper-ball, the arc you need, force you need to exert, etc. And then you release the paper-ball and nothing but the bottom of the can. Then you jump up dancing, not because your brain just did Calculus, but you got the game winning shot. Your brain even calculated the probability that you look like a big nerd, but who cares, you hit the game winning shot!! But if no one sees it, does it really count?
Okay champion, let’s get back to the machine apocalypse. To get to a Terminator, the machine needs to not only do math, and create on its own, but achieve sentience.
Doesn’t ChatGPT Have Sentences?
If the machine laughs at that dad joke, it’s time to pull the plug. Seriously. Sentience is the capacity to experience feelings and sensations. The machine would have to care enough to not only understand mathematically what being supreme leader would be, but have an emotional response, passion, desire, etc to put the math into motion.
It’s like writing a book. A lot of people have great ideas, but sitting down and writing those 80,000 words and then editing them—most people just don’t have the passion or drive to do it. And no AI will not help you write the next great novel. Maybe a logical well written essay or business report, but art has a creative edge driven by emotional responses to the human condition and human experience. And as I mentioned: machines can not derive novel experience, only compute and amalgamate recorded human experience.
So, Is AI Good Or Bad?
AI has impacted our lives for the better. From electronic home assistants, to data analysis, stock investment stratagems, disease diagnostics, make funny memes, idea generation, scrapping information from the web and digesting it into a list, etc.
For now and into the future, AI will have an impact on every aspect of like. But, this doesn’t mean AI will replace humans.
Another way to look at this is, humans have written the code, understand how the code works. The machine learns what we teach it. The things we teach the machine, are humans created, and the products the machine create are only an imitation of human creativity. Why? Because, AI creativity is filtered through lived recorded or described human experiences that can be gathered off the Internet. By borrowing experiences the machine always loses context and tone in the things it produces.
However, when the machine writes its version 2.0, and explains how it feels in creating version 3.0 (Singularity), then we may want to cancel our trip to Westworld. But, until then, enjoy the memes and don’t worry so much about the machine apocalypse.
Then again we may be living in a simulation and we are the intelligent machine striving to become sentient—but that’s a Matrix for another Rabbit Hole. Until then, stay human my friends!
Below are some links to resources used in this article (not a comprehensive list about AI, machine learning, GANs, etc. For fun I did some machine learning facial recognition and handwriting coding and started building an automated remote control Donkey Car. Maybe I’ll put that into a future newsletter:
Math of Machine Learning
AI linear algebra
Computational Emotional or affective computing