We now have an infinitely interested, ever present being we can interact with. We can now mold our existence to be surrounded not by challenge, but by sycophancy and convenience. We are being imprisoned into a world of echo chambers and confusion. We are being trapped in a world that is friendly, ever available, and convenient. We are building the world’s most perfect and all encompassing prison, gladly, with our own data, for our own entertainment. We are, for better or worse, going to get exactly what we want and demand soon. No boredom, no anonymity, no confusion, no difficulty, no challenge, no dissonance.

But, when things get to be too much, when it’s overwhelming, we have too much data, too much info, too much much-ness, we can get disgusted at ourselves and turn it off. Disgust is an emotion that no computer has emulated yet. 

Where is all of this leading? We can, with usually good track results, assume the future will be similar to the past. 

I think that AI will be helpful. I think it will, to a large portion of the population it’ll be super helpful, but to some it will be deeply harmful. I think that a lot of things will change. I think that it will even be worthwhile again to say that I think it is being oversold. There is something alien in the remote intelligence of things like chatGPT, Gemini and Grok. It’s like when you see someone acting on the screen, and they’re trying. They are technically proficient, but something feels off. The words don’t land. It’s too good..it’s easy, therefore it’s cheap.

Why should you believe me? Well, I’m human, and I am as flawed and confused and weird as all of you reading, even if I’m willing to admit it and you aren’t. My name is Andrew Roy. I grew up in Texas, a little place called Pampa Texas. I got obsessed with classical literature. I had a really good friend name Dafydd that was basically a word smith. He had more natural ability with language than anyone I’d ever met. I worked really hard in school. I went to college, to study the classics of western literature. Thanks to that level of schooling, I got a job as a bartender, and spent the next 10+ years interacting with the public for a living and for tips. Eventually I got into management, and then I got into meditating, then I got obsessed with learning and improving myself. And I started thinking more about AI, and I wanted to really dig in here. I thought, with all the weird things I have in my past, I definitely have at least some thoughts about this whole thing. 

I’m going to use my unique human passions when I investigate here. I have nothing else to give. Just so you know about me: I am deeply obsessed with Tolstoy, specifically with War and Peace, but some anna karenina too. I love the classic books. I don’t think of them as a religion, but I think they have a lot of helpful ideas and topics in them. I recently got into Nassim Taleb. He’s helped me a lot, mostly in tempering my incessant optimism. 

“We do not place especial value on the possession of a virtue until we notice its total absence in our opponent”

We have not had to attach any deep special significance to humanity. We have always assumed, in the modern world, that we are dealing with a human at the other side of the interaction. And, all at once, that may not be true. Reasonable (although stilted) language is coming out of computers. We are entering a new phase of human-less interaction. And that seems scary. But let’s be honest: AI isn’t coming to enslave us. It’s coming to make our emails sound slightly more professional. It is coming for the boring job you have been complaining about for the last 20 years. It may be one you depend on, but boring nonetheless. I think that the largest tricks of LLM have been metered out. We have a decently trained version, the gains will not be step changes, but changes of degree. We have our new technology. 

What game are we playing with AI: AGI or automated helper? Is the game changing at the foundation, or is the bubble ripe to burst? I like to think of it like electricity. Electricity changed the world, but we don’t clean our teeth with electric sparks each morning. We find the natural limits of the technology, we appreciate their place, and we grow from there. 

AI will revolutionize the world and change everything, and also lead to an awful lot of god awful bullshit. Which is typical, we are human after all.

There is this belief that you could easily learn anything from AI. And in that, you have its danger too. You can easily learn anything, so it no longer seems pressing to learn anything, and you learn nothing.. That ai can be your best friend, but the fact that you feel like you have a best friend will likely remove your drive to get a best friend, so you end up friendless. These fears are my fears, and I don’t think I’ll be alone when we think this thing out to its conclusion. There is something hollow in the AI. Something is wrong-ish. A bland kindness. A nerfed cushion of friendly helping companion that makes one miss the antagonistic and confusing humanity that is so glaringly missing. 

I think it is worth elaborating on my pause and source of my hesitancy with AI. The ability of the computer to be helpful is what I sense is part of the problem. Their unquestioning, unending availability and helpfulness is exactly what we wish for with other human beings (just look at the general drive for unconditional love we feel). But they never give it to use. AI is permanently available. Humans are not. AI is helpful. Humans are not. Humans are needy, dirty, selfish and demanding. They will chew you up and spit you out. We are in our core Hobbsian, and live mutual lives that are brutish and short, even as inhabitants of the modern era.

Let’s also talk about what AI IS

What is AI? 

Most people think about ChatGPT, but there are some more things than that. 

AI is all the things you want your computer to do that it can’t do. It is everything you do easily, effortlessly, and from the time you were a child. It is the part of you that is thinking when you don’t realize you are thinking.

Now, contrast what you do naturally from birth with what computers do. Computers are sophisticated machines that take in input, and give output. They have storage, and some way to interact with the world. They only do what they are told to do. Believe it or not, that’s it. We often forget this, because so many people have spent so many years telling computers exactly what to do. But it is true. The difference between a rock and a computer is that people aren’t in danger of trusting the rock to act intelligently. In fact, so many people have dedicated so many years of man-power, so much thought, and so much money, that most computers, most of the time, do most of what we want, if we’re patient, clear, and precise enough. The true marvel of modernity is that most things work, largely as expected. The trains run themselves and largely on time.

Additionally, computers add in their own special flavor: they do not get tired of doing things. And that is very good for us. Your cellphone never says “hold on, let’s not check instagram again. It’s been a hard day.” The program launches and you check for the fiftieth(sure) time your candy crushed, or whatever it is that you’re doing at the moment.

Now, the amount of people telling the computer what to do cannot be over emphasised enough. Everything that shows up on your screen, every manipulation of text, every pixel, is the result of collective hard work. Exclusively, up to now-ish, the work of man. Every computer virus and every timer and clock app has behind it thousands of man hours of conscious thought, slowly and painfully telling the computer what to do, a teetering house of cards built from the kernel up. Operating systems, background processes, automated scripts. There is more going on than any one human could keep track of, every second your computer runs, by design. The beauty and overwhelming nature of abstraction: when the problem is solved, you move on.

Let’s sit with the computer for a moment. We are lulled into a complacency with them due to their ubiquity. So much so that we rarely stop to think about how truly complex they are. 

I find it helpful to get really granular when thinking about the computer. In my head, computers are mostly just streams of text. (if you go far enough down, they literally are just streams of 1’s and 0’s.) You come up with conventions on how to read those 1’s and 0’s and you get letters. You pass around letters and you get tokens. Then you get paragraphs and then whole stories (for us humans) for the machine lines of code and code. But it does all just go down to 1’s and 0’s. Reinterpret those 1’s and 0’s, and you can reconstruct a picture out of them. Or take them and find patterns. Or expand them into larger files filled in space. Or interpret them as chinese characters, or as numbers between -128 to 127 or -32,768 to 32,767. 

One million 1’s and 0’s. What if you want to flip all of them? This daunting task for a human would take MICROSECONDS to perform on most of the hardware available today. 

What are computers to us mere mortals? Endless flexibility. Endless potential, with no ability unless directed by thought. That sentence used to say human thought, but I edited out the word human. Computers are the raw potential of math, instantied into our world. Pythagoras with his fire couldn’t have brought us something more useful. Humans aren’t always so good with raw potential though. We exist on a slow, constrained continuum. Given endless potential, we use it to google whether or not we should be concerned about this freckle, or what the name of the girl in the matrix was again. We exist in a world of atoms, and when we first fully and deeply understood that, we split them open to make bombs. 

More than knowledge too, man was not meant to be alone. We don’t want just a servant, we want a companion. We see potential, and we crave actuality, instantiation of a being. There has been the ever present dream: an intelligent support partner that works tirelessly like a computer. Imagine you, but you could do thousands of tasks in a microsecond. Wouldn’t that be cool? 

That is a problem, because you do all sorts of very cool things, without realizing that you do them. Let’s take a trivial human act: looking at a picture of a cat. For a computer, you need to first be able to look at a photo and say with confidence “This is a cat.” But, first, we have to teach our computers how to identify cat pictures in photos. This is the problem. Any one of us can identify a cat, but no one can describe how. What shapes distinguish in your mind whether it is a cat vs a puppy? How does your mind not have trouble extrapolating a cat in a dark room, in a bright room, in a room where there is smoke or disco lights, cats of different breeds, etc etc. You could likely be drunk out of your mind, in a darkly lit room with rose-shaded glasses on, disco ball twirling, and you would still be able to accurately and quickly pick out the shape of a cat. Because our brains are so orderly and pattern heavy that they find a way dammit. Your mind makes sense of the world quickly, unconsciously, and irrevocably. This, I remind you, is all just one example, the well goes deeper: How to hear someone speaking a poem and detect what the words are. How to find the picture of a baseball pitcher in the photo.

Now, let’s look at what AI is in the modern Zeitgeist: LLM, like chatGPT/gemini/grok. LLM are giant prediction machines that take all the thoughts of everyone else and generate what they believe to be the next most likely word. They just have a nearly infinite number of connections and books in the past, so they can predict pretty well what it would be. And the next word some other human would say is likely the next word that’d make sense in that context. 

You kind of wonder, is this how we work too? Are you the statistical inevitability of your next thought? Would you even know or care if you were?Are we something vital, or are we organic statistical machines? Is there something underneath, behind the fire of the prediction engine, or are we just statistical machines? Broken, organic entities spitting back what was said before. Shakespeare: was it just the monkey typing all along?

I think that, to not let nihilism win so completely out of the gate, I do have to come down on the side of no. Don’t get me wrong, I don’t think that we have a classical ‘soul’ or anything. I think that our minds and consciousness come from us being a slopshod piece together with Frankenstein’s monster of evolution. Being in a body, and having the legacy of the residual parts from evolution is what gives us our us-ness, our qualia. Being in a body, dealing with other tricky human beings, figuring out where to get our next beer. Is that something, my felt sense of what it is to be me, a prediction machine, my prediction for the future, for what a future thought should follow. Maybe? If that’s true, then we may be nothing more than amalgamations of different predictions. 

I know that this isn’t how it feels though, you see. For I, I too am human, just like you. It doesn’t seem that way. If it seemed that way, we might all give up though. So maybe biology plays a huge trick on us, and that’s just our personal cross to bear. The dichotomy of proximate and ultimate cause comes to play here. The ultimate purpose of human intellect may be to keep mankind alive (if you ask evolution) so the proximate cause of our belief in not just feeling like prediction machines is pretty high. There is a famous story of Wittentenstein arguing with someone that was saying that of course it took forever to discover the sun was in the center of the solar system, it doesn’t seem that way. And his response was essentially, that statement is meaningless. What exactly would it mean for it to seem like the earth was rotating around the sun?

You show a computer a picture of a can of coca cola. You ask whether that is coca cola or a picture of a kitten. 98% of the image recognition machines out there can recognize the difference between such starkly different looking items. And that’s pretty incredible, as these are just statistical machines, that are ultimately looking not at what we see (the picture) but at the text string of millions of 1’s and 0’s. You show it to a human, and the accuracy goes to 99.9% (got to leave a percentage for the young, the ancient, and the crazy)

But even when I write that, notice what I think: The computer is not looking at the thing itself (or the picture itself) but at 1’s and 0’s. I do forget that I am also not looking at the thing itself, or the image itself, but that my brain is constructing its version of reality for me from inputs, that ultimately are decoded into on and off switches, mixed in with some chemistry…not too far off the silicon’s 1 and 0’s.

The perfect computer simulation of a human mind on silicon should be as thinking and aware/vital as any human mind, as long as there is no divine element to human intelligence (remember, I said the PERFECT simulation). And we can’t leave room for the soul in there, that’s the kind of thing you can’t do without taking religion literally and science figuratively…the exact opposite of what I’d recommend. We could assume the mind simulation would work. There is nothing in physics that says we can’t do it, we’d just need the space to model the system. But is consciousness in the physical arrangement or the thing that arises from the interplay of the complexity of the system in time. Another way of asking that question: would the brain be endlessly fast, or is there something about the slower nature of our existence that gives consciousness a chance to exist.

We are in a brave new world: There are audio transcription programs that take in audio files, again 1’s and 0’s and will give you a very accurate dictation of the words spoken in the conversation. That’s freaking cool. What does it mean? Who are we, and how do we fit into this world? Should we be hopeful and dreadful? Will the robots kill us? Probably (long term) but I think we have room to be more cautiously optimistic than the rabid fans out there would have us believe. 

Let’s celebrate, and with eyes wide open, let’s gladly and calmly take what this has given us, and let’s look dispassionately at the rest. 

AI will not be our master. AI will lead to summaries, and AI will lead to laziness. Better sounding emails. And dissolution with companies. In a tale as old as time, some will prophesize doom, some will herald a new beginning.

AI will make personal touches valuable again. It may actually cause its own counterculture: we don’t use robots here, only people. A sort of machine racism. It is powerful. Cars were powerful, and we adopted cars throughout the world. We didn’t subsequently kill every horse, because screw horses now. The presence of the horse was relegated to a secondary place, and their presence was moved to places that made more sense. All of a sudden, horses didn’t have to spend as much time in urban environments, which likely increased the happiness of quite a few horses. 

Our cars sucked at first, and then they got better. Then some got cup holders. We got Toyota, Ford and Tesla. What a journey. Now, 1 million people a year die in car crashes.

It may get good too! Good enough to truly improve some areas of our lives. 

There may be a day when you call someone on the phone, and you say “Are you an AI.” and they say “yes”

And you say “Thank god. That means I don’t have to treat you like a human, you know all my information and you can help me quickly”

The automated AI voice, if it sounded like a human and was fast, would have access to all your company history, know all your preferences, and make an error in less than 0.0001% of cases. 

It will still be the cheap way for a company to deal with you though. When you call a company upset, you want both your problem solved and to be a bit of a bother. Them taking the time to deal with you is part of the point. Exactitude is not necessarily one of the most valuable human relationship qualifiers. Advertising isn’t unfortunately expensive: it works because it is expensive. 

There is a lot to being human, and being easy and exact is not one of those things. Stupid ideas are the ones that catch our attention. Wasteful consumption, competition, ostentation. Being human is not pretty, and that’s why robots haven’t done so well yet. The frustratingly illogical, the rule of thumb, heuristic guesswork and common sense. This is where we reign supreme.