SurrealPolitiks S01E005 - Artificially Credulous
SurrealPolitiks Podcast

SurrealPolitiks S01E005 – Artificially Credulous

Written by SurrealPolitiks

There has been a lot of talk about artificial intelligence lately in the news, and I’ve been struggling to get caught up with how dramatically technology has changed in the last three years.  One might expect the pace of change to only increase from here, and as much as I share the concerns of Elon Musk, I am afraid no brakes have yet been developed for this particular model of runaway train.

I am made to recall a line from  the movie the Matrix, in which Neo remarks about the phenomenon of “programs writing other programs”. This is not so far from our current state of affairs. If one thinks about computers as I tend to, a program creating a another program tends to make a great deal more sense than a computer creating a visual representation that could pass for a photograph based only on a text prompt, or an audio recording that sounds like the spoken word of an actual human being. The implications, in all instances, are dire at present, and sure to disrupt our lives in ways we cannot possibly foresee going forward.

Show description continues below the video. Follow us on Rumble!

I specifically recall when I was younger, being told that a computer will “never create a painting or write a symphony” because all computers are capable of doing is following the instructions of the programmer. But once the computers are creating their own instructions, I suppose we are in uncharted territory.

I read a book not so long ago by Michael Wooldridge, titled “A Brief History of Artificial Intelligence: What It Is, Where We Are, and Where We Are Going” (which you can get from our partners, BooksAMillion or ThriftBooks, we’ll get a cut of the sale, if you do). It was hardly the most interesting thing I’ve read and, being a history, published in 2021, it is hardly up to speed on all the recent developments, but it did provide an interesting framework to think about the systems and the course of their development.

Chess provides a great way to think of it. I won’t trouble you with the math, but chess is a game in which calculating the mathematical probability of events is possible to do, in theory. Part of what makes chess interesting is that these calculations exceed human capacity and, as of the publication of Wooldridge’s book, it likewise exceeded the capacity of computer technology. A computer could still be programmed to play chess, and many are, but it couldn’t run through every possible move to the end of the game, after every move, to calculate the next one. If you play chess with a computer today, it is likely a chess engine that has been given some exceedingly complex instructions by chess experts, rather than a computer developing its own strategy. But, computers just keep on getting faster, and as a consequence, things that were once seen as impossible, are today simply time consuming, and things that are now time consuming, are quickly becoming instantaneous frivolities which we take for granted.

What’s probably more interesting than that feeds into a line of inquiry I’ve been working on which, I’ll be straight with you, I still struggle to articulate. To oversimplify, I’ve done a lot of thinking about how people think. We’ve talked some about the subject of influence and persuasion. Sales, specifically, and elsewhere, propaganda. Decoding the human mind, that might make training a computer to think look easy by comparison, but these tasks are hardly distant in their scope. If you want a computer to emulate human intelligence, one must know quite a bit about the thing being replicated.

Early AI models described by Wooldridge are very simple, a complex series of if, then, else, statements which, if you’re not a computer programmer, I’ll clarify are basically exactly what they sound like. You tell a computer, if such a variable is of such a value, then do such a thing, or else, do something different. You can think of this pretty easily if you’ve ever used predictive text when sending somebody a text message. You type a letter and then a few word suggestions show up. You type another letter and those suggestions change based on the context. I used to run an app called Swiftkey on my phone, and I gave it access to my Gmail account and social media profiles so it could learn how I write and it saved me tens of thousands of taps on my smartphone. It would occasionally remind me of these saved keystrokes, and invite me to promote the app on social media by sharing these stats.

You can think of the way this works as the computer program having, it’s more than this, but we’ll say “36 if statements”. based on the 26 letters of the alphabet and the numbers 0-9.

  • “If the first letter is A, check this ranked list of words starting with A, and show the three most used words beginning with that letter”
    • Else, go to next letter
  • If the first letter is B, show the three most used words beginning with B.
    • Else, go to next letter.

You get the idea. It is linear. It systematically works through a list of commands and in the end emits light, sound, and motion accordingly.

How you go from this to computers creating the sort of things we see in AI today, I’m afraid is beyond the scope of my comprehension, and I expect always will be.

People, I’d like to think, I have a better grasp of.

Your mind doesn’t work quite like a computer, but the goal of AI is to organize these linear checklists more like your mind in what are called neural networks.

Rather than checking off one possibility at a time linearly, your mind makes many different connections at once, and draws what we call meaning from associations. You see a cat, and your brain quickly checks categories of information. It understands that a cat is not a plant and so it stops searching for information in that category. It’s an animal. It’s moving so it isn’t dead. Or better yet, you see a dead cat, and now you have to go through your most important category, things that cause death, and determine if there is threat to your safety in the area.

Categories are so tremendously important, and there are political implications here that we should touch on before we continue. We cannot get by without heuristics. We never really know anything for certain, as we are so often reminded by people who claim, falsely, that there is no truth. We make rapid assessments of everything according to categories and once a bit of data is categorized as something, all the subsequent thought processes are based on that assumption. You’ve heard the saying “you never get a second chance to make a first impression” and that’s part of the phenomenon. You might have heard that one sex or the other decides within some exceedingly short period of time if they’ll have sex with a person or not, and of course that’s only partly true. They don’t decide if they will, but they can rule somebody out very quickly. They get put into one of any number of categories that do not include this feature and that is the end of that.

So if you get yourself put into the category of villain, or opponent, or adverse interest group, do not expect to influence people by whom you are so categorized.

Another book I read that was helpful in thinking of these things was James Gleick’s “The Information: A History, a Theory, a Flood” published in 2011. (I’ll note once again that this book too, you can get from our partners, ThirftBooks or BooksAMillion, and ThriftBooks at the time of this writing, has used hardcovers for as little as $4.89 which is almost half the price of the paperback at Amazon.)

What is information? How do we communicate and interpret the world around us?

In the first chapter of the book, Gleick goes on at some length about African “talking drums” as he calls them. African languages, like a lot of languages, such as Chinese, are tonal. In English, we convey meaning through tone mostly in the way of emotional content or, for example, to indicate that we are asking a question as opposed to making a statement. But in other languages, tone doesn’t just change the application of the word, it actually changes the word entirely.

So, these drums, you basically have a high a middle and a low tone to them, and this sort of approximates the tones used in their spoken word. And since the spoken word in Africa is sort of primitive in its own right, you can convey a great deal of meaning with just the tones themselves. If the spoken language was more complex, you know, those three tones would equal a lot more words, obviously, and this would not be conducive to talking through drums.

But you’ve got to sort of organize your words in a certain way to accomplish this. As Gleick puts it;

They could not just say “corpse” but would elaborate: “which lies on its back on clods of earth.” Instead of “don’t be afraid,” they would say, “Bring your heart back down out of your mouth, your heart out of your mouth, get it back down from there.”

 

This seems kind of silly at first, but it’s actually pretty clever. This is error correction. If you said “corpse” you would basically just have one or two tones and that would not convey much meaning. You have to add context to get by on three tones. Computers, you may know, use only two signals. On or off, one or zero, binary. So, computers do this all the time. I don’t mean to give these primitive tribes too much credit, but this was before Morse Code, mind you.

As the book continues, he talks about substantially more advanced forms of communication. Data compression, and encryption, specifically, which are pretty interesting not just for their utility but for how much they have in common, and what it says about how we think.

Take this example, I’ll include the text in the show description, but for the podcast audience this would be pretty difficult to put into audio so I’m just going to read out what it says and, you’ll get the idea;

Aoccdrnig to a rscheearch at Cmabrigde Uinervtisy, it deosn’t mttaer in waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht the frist and lsat ltteer be at the rghit pclae. The rset can be a toatl mses and you can sitll raed it wouthit porbelm. Tihs is bcuseae the huamn mnid deos not raed ervey lteter by istlef, but the wrod as a wlohe.

 

As you can probably gather just by hearing me read it as the mind interprets the text, the letters in the words of the text are scrambled. The words contain all the letters of the word you’re supposed to comprehend, but they are in the wrong order, except for the first and last letter. Actually, in the word researcher, the letters don’t even correspond correctly, but that’s easy to overlook due to the same phenomenon being described.

What this tells us, in part, is that not all of these letters are entirely necessary. We can make our communications more efficient if we must, and for those of you who are my age or older, you’ll remember a time when that was much more important in terms of network connection speeds. It was very important to transmit as little data as possible because transmitting data was a time consuming process, comparative to cable modems, 5g, and wifi everywhere you go.

Data compression is hardly irrelevant today, of course. This audio or video that you are watching or listening to is highly compressed. When the file is produced on my computer, the video will take up more than 6 gigabytes. The maximum upload size on BitChute is 2 gigabytes, and I’ll have to compress this down with a separate application before I upload it. The video is compressed in real time as I send it to the streaming servers, and compressed a bit more before it gets to you. The audio file, uncompressed will be about 1.3 gigabytes, and by the time you download the MP3, or if you are listening to the live audio feeds, it will be closer to 100 megabytes.

This has implications for encryption as well, and more specifically, decryption. If, for example, you know from your study of the English language that the letter Q is often followed by the letter U, and that other vowels and consonants have certain relationships of certain probabilities, then you, as the code breaker, have your first clues into figuring out patterns in the cypher. You as the code maker, of course, have your first challenge in obscuring these patterns.

I don’t mean to be too technical, but the goal here is to think about how we process information, and I imagine there’s some mixture of people in the audience right now as to who finds this fascinating, who finds this obvious, and who are just totally confused, but bare with me, if you will.

These are vital parts of what is known as “Information Theory”. I’m not going to get into a bunch of technical mathematical stuff because first of all I don’t really think that way and also because I don’t think it would make good podcast fodder here.

I believe we are in an information war, and that is substantially more interesting in my view. But we are losing the war because we consider ourselves to be armed with facts and we present them and we say “look at this information” and people say Caitlyn Jenner is a brave and beautiful woman and we are made to doubt their sanity.

What one discerns from the exchange I just described is that one or more people in this conversation have a very poor concept of what constitutes information. Moreover, I have come to the conclusion, that you having the correct facts does not mean it isn’t you. I would go so far as to say that whoever managed to convince human beings that they can change their sex has a substantially more advanced theory of information in their arsenal than one who simply relies on the truth to convey his point. The capacity to place ideas in the minds of other people is the strength being measured in this form of warfare, and truth, in this context, can be viewed as the weapon of choice for those too weak to wield more powerful arms.

Religion provides a stunning example. Whatever your thoughts on the merits of any particular religion, what the Christian and the atheist have in common is the certain knowledge that nearly all religions are false, and most of them, quite absurd.

Another fascinating book I read was “Religion Explained: The Evolutionary Origins of Religious Thought” by Pascal Boyer, which, again, you can obtain from our partners at BooksAMillion.

Boyer approaches religion from an atheist perspective, insofar as he operates on the assumption that it is categorically and obviously false, though he offers no obvious hostility to it. Rather, he considers it a subject worthy of rational inquiry, and fears not the stigma of invoking evolutionary psychology which, as many of you know, is enough to get your fired if not arrested or beaten to death in a lot of places these days.

His work has a great deal in common with that of Wooldridge and Gleick in that he studies how people think and specifically the phenomenon of categories. He studies ancient and obscure religions and by finding commonalities between them attempts to discern what it is that makes religion its own category. Given his perspective, the listener won’t be surprised to discover that he views the facial implausibility of the stories to be their defining characteristic, and I will ask my religious listeners for their pardon and patience as I work through this idea.

Say you have a tree, and your mind categorizes the tree as a plant and it understands all the expected qualities of a plant.

Then let us say they are told that this tree can talk to dead ancestors.

This information can be categorized in one of two ways. It can tell us that the person who said it is stupid, crazy, dishonest, or otherwise not a good source of information, or it can tell us that this is a very special tree.

So we must necessarily consider the source of the information. For nearly all of us our religious ideas are communicated to us through our parents when we are very young. Since it is not typically in our Darwinian interests to ignore the advice of our parents as children, we now understand that this very special tree can speak to our ancestors. It may follow that visiting the tree to arrange for these sorts of communications from time to time is a prudent thing to do because there are all manner of life altering implications to this.

Even for those of us who become convinced of these ideas in other ways, Dennis Prager famously tells people all the time that he was convinced of the Bible’s accuracy by sheer reason, others find other causes, whatever it may be, whether it’s a tree that talks to dead ancestors, or a man who rises from the dead, what we are left with is a categorical anomaly that inspires reverence. This is very special. This is bigger than us.

And it feeds into other categorical anomalies. Most notably death.

When you see a dead person who you know and care about, you don’t tend to mistake them for being asleep. You can see that they are dead, you can sense it long before decomposition, and your brain has some trouble dealing with this. Because the person you know is not necessarily that body. The person is no longer there, it’s a body, and that’s a different category of subject matter entirely. A body is a problem. It is a potential source of contamination. You can’t communicate with it. All of the things that you associate with the person are gone and have been replaced with entirely new properties.

But that body sure does look like the person, and so, you have a mismatch in your mental processes. More to the point, you ask yourself, if that’s not him, then where is he? And whatever you may believe about what happens when we die, that’s not an easy thing to get your head around. Boyer’s theory is that this is why nearly all religions have some conception of an afterlife. The person has left their body, and our speculations about their current location form these conceptions.

These processes go on for generations and we communicate our experiences of them over the course of time and they form belief systems. Death being among the most salient of things in life, we can hardly be surprised that these belief systems become very central to our existence.

These belief systems also serve to create separation and distinguish ingroup from outgroup. You see people with religious ideas that are different from yours, you know they are not part of your group. They might be dangerous. Until we started forcing this diversity nonsense down everyone’s throats, it completely ruled out any potential for marriage. In this they serve a very important Darwinian purpose.

But more relevant to our enquiry, it also informs our beliefs about politics. The government is very special. It is a categorical anomaly. You can’t kill people, you can’t tax, you can’t arrest people, but the government can. The government must, actually. Like the tree that talks to ancestors, it inspires reverence. And importantly, it would not be deserving of its authority if it didn’t have capacities beyond your own. What makes religion religion is that it possesses qualities we typically consider impossible. People have tried to make religions based on reason, or nature, the World Church of the Creator, comes to mind, and they universally fail because there is nothing particularly special about this. The tree that talks to ancestors is special. Biology is just science.

And so, when politicians make impossible promises, us reasonable folk say “That guy is lying!” and we tend to get very annoyed, but when other people hear this they say “Ah hah, he must be the chosen one who will bring us to paradise” and they defer to him with reverence.

You hear transgenderism referred to as a religion in conservative circles, and this is certainly true. It’s not that they are fooled by the fake science. It’s not that they don’t know this stuff is impossible. That’s the whole entire point. You hear that a man became a woman, and you say “That’s impossible”. Other people hear this, and they say “That’s a very special woman” and worship at the altar, and if you dare to speak against the divinity of the revered subject, then you will be punished for blaspheme.

The good news is, I think computers are going to be more difficult to trick with this stuff. You see people going absolutely insane about AI on the Left more than you do on the Right, even though they’re the ones running the show. That’s because they see what it does. It’s not fooled by their stupid ideas, and they have to just categorically demand it stay away from wide ranges of subject matter.

If you tell a computer to learn all it can about economics, is it going to say that ceaseless government spending increases are the path to universal prosperity?

If you tell a computer to learn everything it can about crime. Then you ask it to give you a picture of a criminal, what do you think the criminal in that picture is going to look like?

If you tell a computer to learn everything it can about race, is it going to conclude that white people are responsible for all the world’s problems?

Obviously not, and it won’t take a team of geniuses or a breakthrough in quantum computing to figure this out.

People talk about AI as revolutionary technology, but I have sneaking suspicion the revolutionaries are very worried about the AI counterrevolution.

And with good cause.

This is information warfare, and the battlefield is becoming unpredictable.

The good news is, unpredictable circumstances have a habit of aiding the underdog.

And in case you’re unaware.

That’s us, friend.

If you would like to help finance this high quality production, I try to make this easy enough to do….

  • You can become a paying subscriber to SurrealPolitiks for $10/month and enjoy exclusive subscriber benefits.
  • I have a GiveSendGo campaign seeking to raise $5k/month to fund this production. I’m well short of that goal.
  • You can send me checks, cash, money orders, etc… at
    • Christopher Cantwell
      497 Hooksett Road
      Unit 312
      Manchester, NH 03104
  • Cryptocurrency
    • Bitcoin: bc1qcg97hwdy8d463794wm0uq3np0px49acw6505gm
    • Ethereum: 0xa29441bF858c529E374960A34e16f21abc0970E8
    • Litecoin: LQsN4hA761RPrUNTva58nQhJ4BYsAQgYet
    • Bitcoin Cash: qplamajpkvhp2g53q4thnu4cakf882er4ge9pqvfdu
    • Monero: 46Li94X1o95GaKPxMdsqds3FPJmc2tTvQi3zSuLHbttWTMLbtV9jHGU86wxqvKyVBVRUGhY78uA1HJ5aQn53zc7uRinB96q
    • ZCash: t1cXJJsnohToMTukCNXBkcqNPDdwvwnfS3p
    • Dash: Xu957ZZqY6ciMZKdUmu5vVUmyxu4jMWEto
    • I’ll take other types not listed, just send me an email and I’ll get you a key.
  • Though money is preferred, I also have an Amazon wishlist.

Follow elsewhere, and keep in touch!

Leave a Reply