SIIIEII Artificial Intelligence: Emergence From Nothingness
Hello world. The news is, uh, dramatic, at this point in time. The Russian invasion of Ukraine continues to be horrific. It seems we are still not over the genocide phase of humanity, as the soviets genocided Ukrainians through the hunger famine of the 30s and in ways large and small through the years. Uh, what else. There’s a still ongoing global pandemic, food, gas, and energy crises, and humanitarian and human rights issues across the world.
And what a time to be a law student in America. We’ll see if Roe v Wade is for sure overturned in a few weeks, but if it is…well, for better or worse this will essentially become my generation’s fight for civil rights - specifically, the civil rights of women to bodily autonomy in America.
On a personal note, I’ve had some pleasant events in my life recently. An episode from this podcast was featured in a Norwegian university for a philosophy class syllabus I believe it was. I try my very best to be academically rigorous, it makes me very happy to have contributed in some way to the lives and education of others. I do admit, the quality of this show gets better over time…My, uh, more recent episodes I believe are more polished. But that’s great.
I’ve decided to release several of my academic papers on patreon, and I’ll add more as time goes. Check those out and support the show. Of course, I’m happy to send anyone pdfs of the papers for free for educational purposes, just send me a message if any might be up to standard.
This episode topic was actually requested by a friend of mine roughly 2 years ago. I had no idea how to approach it at the time, but as it turns out artificial intelligence is itself an incredible example of emergence. Cheers, Jon, hope you’re well.
And-start
There’s a whole lotta nothing in the universe. I mean that literally, there appears to be so much more emptiness to the cosmos than there is “stuff”. Our solar system is mostly empty space. The sun is massive but its farther out planets (Not you Pluto, sorry bud) like Uranus and Neptune are very, very distant. And our solar system is itself lonely. The nearest star to our sun is Proxima Centauri. Traveling at the speed of light, it would take 4.25 years to reach it, that’s what’s meant by a light-year when people talk about astronomical distances.
Light always travels at the same speed relative to an observer, which has some truly bizarre consequences, but that’s a story for another time. It takes light from Proxima Centauri 4.25 years to reach us, which also means we see Proxima Centauri with our telescopes as the star was 4.25 years ago. Telescopes are essentially time machines.
Another thing about the speed of light - nothing that we know of can travel faster than it. It’s a universal speed limit - truly, universal.
Currently, we can’t travel remotely close to the speed of light. As an object speeds up, the apparent energy required to increase its speed increases exponentially. Meaning, without godlike power, we are heavily limited in our reach. Even a mission to Mars seems daunting to us in our current age of spaceflight. Godspeed to whoever is chosen for the first manned Mars mission, that’s going to be such an amazing voyage and I hope it comes soon. Now, Proxima Centauri, the closest star, is hundreds of thousands of times farther away than Mars. So imagine you could fit several million suns lined up in a row between our sun and the nearest star. Pretty sure I got the maths right on that. Fact check me please mathematicians, or my dad if he’s listening. Given that with current technology a trip to mars is several months, it would take generations of humans to reach the nearest star.
As if that wasn’t enough, we’re not the only galaxy in town. There are potentially hundreds of millions more. Our nearest galactic neighbor is andromeda at about 2.5 million lightyears away. Very, very, very far. Like, the idea of being halfway between the milky way and andromeda…so much emptiness for what might as well be an eternity away. You can’t help but wonder why. Like what purpose does a universe serve that is mostly empty?
I know that there’s so much we don’t know yet about dark energy and dark matter and perhaps there’s a question of whether our universe truly is as empty as it seems, and I also know that the formation of the universe as far as we know was governed by physics, and those physics played out according to the starting positions of whatever there was in the big bang. Or whatever the big bang was, a point of infinite density doesn’t make much sense to a human brain. The laws of physics governed it.
When you dive into the world of the very small, you find the same fact: matter is mostly empty space. It’s elementary physics and chemistry but recall if you will that an atom has a central nucleus made up of protons and neutrons where most of its mass is, and orbiting around that are tiny, tiny electrons. An electron is almost two thousand times smaller than a proton - not the largest size differential - our sun for example has over 300,000 times more mass than earth - but, even so, an electron is really, really small, like beyond the human capacity to imagine smallness.
But it’s absurd because even the nucleus of an atom is tiny in comparison to the overall orbital space of electrons around it. I don’t have the knowledge or skill to make the exact computation, so I found someone who did the math on Jlab - they calculated that a hydrogen atom is composed of about 99.9999999999996% empty space. All the objects we see - everything we can touch and manipulate in reality - is basically empty space containing fields of probability of where an electron might be.
What are the laws of nature? Why do interactions between fields of probability in a specific way create matter and human beings?
To bridge the gap between the facts of our existence and the laws of nature - let me introduce:
Hard Logicality. I cite a term coined by the YouTuber Exurb1a-
In his video “The Mystery at the Bottom of Physics”, Exurb1a puts forth the idea that, quote, “There is something special about our constants and our physics - that they’re baked necessarily into reality”.
Let me offer what I hope is a decent explanation - our existence is predicated on certain events happening in exactly the way that they did, therefore making them a necessary condition of our existence. In other words, why did the big bang create the universe in this way? Because if it had created it in any other way, we wouldn’t exist to observe it. I don’t really know where to take that thought. It seems both profound and mundane. Yeah dude, the universe is a necessary condition of existence and it’s all connected and important. But what happens now? Does our universe repeat? Will it be like Nietzsche’s eternal recurrence, ourselves reliving our own lives, because existence is necessary?
I’ve thrown weird but verifiable facts at you for the past several minutes. We’re now exiting the sciency realm and entering the abstract.
So. Emergence. It can be summed up in the phrase “the whole is greater than the sum of its parts”, derived from the works of Aristotle but not exactly said by him. I know this because I googled that quote and found a really good article explaining its origins. Link in show notes of course. Aristotle’s metaphysics are…complicated, to say the least. For my purpose, what I mean by “the whole is greater than the sum of its parts” is that the combination of parts can bring new things into being outside of what was already there.
For example: a bicycle. What is a bicycle made of? A seat, two wheels, a frame, handlebars.
Now, if I handed you a seat, two wheels, a frame, and handlebars, would you have a bicycle?
I think intuitively, we know, no, you need to put the pieces together first. And what do you get when the pieces are together? A bicycle. We agreed I did not hand you a bicycle. Therefore, something emerged from the sum of the parts that was not there before.
That is emergence.
And it’s a common occurrence in nature. In undergrad I took a psych class about the structure of human thought under Professor Gary Lupyan (might be wrong on pronunciation) anyway in this class we discussed emergence. I remember Professor Lupyan brought up the example of honeycombs.
Why is a honeycomb, so actually perfect? Ever thought about that? How is it that little honey-producing insects create almost perfect hexagons in thousands of repetitions when their brains are little more than pheromone detectors?
Speaking of bees, also, remember the scare with the murder hornets entering North America? Well, considering they can massacre bee populations, I hope it’s going well with killing the murder hornets. Anyway. Bees in Asia developed a natural ability to defend themselves from the giant monsters. In size comparison, the hornet wins no question. Asian giant hornets tower over honeybees like attack on titan - shingeki no kyojin - literal titans biting heads off and stuff. But dozens of bees will swarm around the hornet and vibrate their body heat high enough to cook the hornet alive. One bee alone of course couldn’t accomplish the task - a few dozen, you might say, emerge the property of roasting a hornet alive.
In the honeybee we find emergence of complex behavior from simple inputs. Why does a honeycomb exist in a hexagonally tesselated structure? Our best science indicates that bees don’t build hexagons: they build circles. And when many circles press up against each other, they form roughly 120 degree angles with each other. A hexagon emerges from circles in a specific configuration.
Emergence is inescapable. It’s a fact of nature. No matter the configuration of the universe, no matter if no matter exists at all - in the no matter universe, if matter did exist it would have the possibility of creating an emergent property.
On top of that, if you’re a mind idealist like Berkeley and take seriously the idea that esse est percipi and what exists in the mind is all that exists, you still can’t escape from emergence. Take a brief gander at a human brain. Do you know what you find? Neurons, synapses, myelinated axons. Now, let’s assume we have a mind. I know I have a mind because I’m thinking right now - insert Descartes reference here.
Can you point me to the mind? Where is the mind located in a brain?
We don’t know. Shocker. For all our science it seems the mind is an emergent property of the brain that we don’t have a good explanation for. This is disputed in philosophy of mind but that’s a rabbit hole. Maybe turn these ideas over in your head a bit while we cut to a break.
Okay, now:
The Algorithm. It’s the secret sauce of artificial intelligence. Algorithms control every search engine on the internet, every social media site, every mindless browsing experience where content is thrown at you and you gobble it up for hours because you have no self-control. Or that might just be me. How does the algorithm work?
Excellent question, and one that every youtuber ever would like to know the answer to.
Algorithms are like walking - they take things step by step, instruction by instruction, to perform some specified function.
To take things one step further, we have neural networks. CrashCourse as per usual has an excellent video on the subject, you can find the link in the description.
Essentially, a neural network is a collection of algorithms that take inputs, put those inputs through layers of equations called nodes, and produce some output as a result. Neural networks got their name from the fact that charting inputs, nodes, and outputs of the algorithm looks something like a human brain with its neural connections.
Another thing to know is the concept of machine learning. We don’t get to superhuman AI without machine learning, which is basically the ability for machines to train themselves to perform better at something.
Something that becomes more apparent with the study of AI is the tendency towards embodied cognition. Now, embodied cognition is such an incredibly vast subject of inquiry… To sum it up, just as the Whorf hypothesis in linguistics tied human thought to language, the embodied cognition theories tie the functions of the human body to thought.
One very poignant point from within embodied cognition is the fact that we don’t get anywhere with artificial intelligence unless we have a value system embedded within. Deep Blue and AlphaGo were created given a specific set of parameters to achieve. Of course, it’s more complicated than that, but that is one fundamental axiom of their AI. There is a goal, and the success of the AI depends on that goal.
Machine learning automates the process of achieving the specified goal. Instead of a human reviewer telling the AI whether or not it did a good job, another system of AI is set up to optimize for correct actions or answers.
There is much speculation about the implications of goals for AI systems. Right now, we’re aiming AI at board or video game dominance. And it’s doing great. But what about Wargames? The actual movie as well as just in general the game of war.
Well, that’s the gist of it. There’s a lot of advanced work building on the principles of neural networks and machine learning to create -
Artificial intelligence. Algorithms that can predict things no human would be able to. Computers that can reason and think. At what point can we say that the algorithm has a mind?
Where do minds even come from in the first place? We use them to think, but it’s not obvious at all what they are or where specifically they emerge from. We’re pretty sure they come from the brain. No known location could account for our mental states other than the brain.
But if the brain creates the mind, then an artificial brain would create an artificial mind, right?
That moment where mere patterns of information inside electronic chips become something more, something conscious…well, that’ll be a day. Emergence. Finally, maybe, we will have created artificial life, capable of thinking, feeling, of having an internal world, something it’s like to be that intelligence.
No one has a clue what to do after that. If we create life, true life, from what was before nothing, not-life, do we have to acknowledge this new thing as having personhood and rights?
It’s a question of ethics, and one that probably deserves a lengthy analysis.
More on-topic right now, this thing that emerged-what is it actually?
Software creates the algorithms, the numbers, and how they interact with each other. Each calculated node in a neural network is dictated by the algorithm and the numbers that preceded it. Would life exist in the nodes and their connections? Or could life be found in the software, the lines of code that create the numerical model? Or is life perhaps in the numbers themselves? Probably the most intuitive solution is that consciousness would equate in some way with the physical hardware of the computer, the hard disk where the program is kept. That makes sense to us who have consciousness due to our physical brains.
It’s not a very satisfying solution though. A program does not need to limit itself to any particular physical matter, so long as its parts exist somewhere, in the cloud for example.
It’s not clear whether artificial intelligence will ever be conscious. It may never be clear. How could numbers possibly have feelings? Even the idea of software actually experiencing sadness is absolute craziness on par with conscious toasters. Doesn’t change the fact that we are hurtling toward a reality where we will have to answer these questions.
But that’s all I got for now. Good luck everyone, stay safe, here’s hoping for a brighter future.
References:
Aristotle:
https://se-scholar.com/se-blog/2017/6/23/who-said-the-whole-is-greater-than-the-sum-of-the-parts
Space and the distances between:
https://en.wikipedia.org/wiki/Interstellar_travel
Emptiness of atoms:
https://education.jlab.org/qa/how-much-of-an-atom-is-empty-space.html
Exurb1a and hard logicality:
https://youtu.be/EH-z9gE2uGY
Machine learning:
https://www.sas.com/en_us/insights/analytics/machine-learning.html
Crashcourse neural networks:
https://www.youtube.com/watch?v=oV3ZY6tJiA0
Honeybee honeycombs:
https://www.nature.com/articles/srep28341