The Endless Depravity of Machines (or some thoughts on AI)

Layoffs loom large and real just about everywhere, most often blamed is AI, some of that is real, some may just be air cover to squeeze the margins... but it came for people I was close with at my own workplace.  Some thoughts and feelings about artificial intelligence, it's place in our world, our place in the world.  Next time: Back to Book Club and the oddly happier space of There is No Antimemetics Division.

Transcript

SPEAKER_00 0:02

Well, hello again, amazing internet people. We find ourselves here recording and publishing on 420. So strap in and smoke em if you got 'em. Because what's been on my mind this past week is the wild world of artificial intelligence and how it keeps coming to invade our lives. I've been thinking about it especially because if you're anywhere in tech, you understand that the robots be comin' for the jobs. There are layoffs at every turn at almost every company. My company too had another round of layoffs just lately. In this case I've fortunately kept on this round of King of the Hill and have made it into another round of survival. Which is good because, hey, without jobs we don't pay our bills, and if we can't pay our bills and don't have a job, what we lose in particular here in our good old US of A is immediate access to health care. So thank you, employer, for not driving me into destitution, desperation, though I will never here escape the depravity of my own mind. But let's talk a little bit more about what thoughts have been rattling around through my head. As I said, we keep seeing more and more layoffs, and in some cases desperate rehiring as it turns out the robots aren't quite what everybody hoped they were going to be in terms of just solving problems. And when we think about the layoffs going on, it's hard to really sometimes suss out what's really going on in each case. Some of the layoffs where AI is brought up are absolutely truly because of it, or at least the promises of it, the hopes and dreams for it. And in other cases, it's just people saying the words artificial intelligence have made us more efficient, in order to just cut the bottom line and increase their revenue. In fact, that's probably the case for the majority of layoffs where artificial intelligence is blamed. It's not really there solving a problem, it may not even really be all that much in use, but it provides a wonderful public excuse to cut your headcount. And in the middle of cutting your headcount, you've announced, hey, I'm using the thing that makes stocks go straight up. Makes those stocks go stonks. That's what Allbirds just did. All it had to do was announce that it was seeking to become an AI or data center company, and up went the line. It's strange how all you have to do is use some wonderfully hopeful words to really, really lean into that future tense, that future hypothetical. We say that AI will bring more productivity, it can bring more productivity, it could solve our problems. But the language that I hear lacking time and time again right now when I'm listening to what other people have said about layoffs, or what's going on in the very offices and Zoom meetings that I sit in and hear about our own. It is continuously that AI will do the thing. Will do the thing, not is doing the thing, has done the thing, but will and could be amazing. And you'd better embrace it. You'd better embrace it right now and show everybody around you how you are making use of doctor Flattery the compliment box, otherwise you will be rendered immediately obsolete. They keep using that word obsolete. And not just that you'll become obsolete, but it will be your own choice to have passed into obsolescence. That's a hell of a thing to tell people. You are choosing to become worthless. You are choosing to be overlooked. You are choosing to be made an example of. And all you have to do to avoid that is use the tools that we want to have replace you. And the faster and more deeply you use them, the more deeply and faster we hope the robots will in fact replace you. That way we can pay our friends money rather than our employees. And that's a pretty terrifying and stark reality to exist in. And so it's definitely stayed on my mind. It's something that's hard to escape in my thoughts. Because I am in a job where I have to demonstrate how, when, and where I am using artificial intelligence. But artificial intelligence has come to mean so so very many things, some of them useful, some of them maybe not. I think many of us in our day-to-day lives use many different algorithmic tools, a lot of different forms of machine learning in many different capacities. And in many of those senses, they are wonderful tools. I'm using some noise reduction software right now. It's operating via algorithm, it is in some sense AI. But it's not the AI that we've come to know in the chat boxes. The wonderful little interactive toys that we hopelessly anthropomorphize, that we hope that if we treat well, we'll do the thing that we really want them to, to make our jobs easier, that way we can prove that we're ultimately unneeded. But what's really going on and are these tools that are improving our productivity day to day? Again, sometimes I do think yes. Throughout different code applications, especially in limited capacities, especially in prototype capacities, AI really can be a boon to getting a first increment out the door. If you're hoping to push production software, I think we're in a much more limited place, or a place where people are absolutely willing to push forward, and where quality is going to start hitting us in the face later on. One of the things that Ed Zitron in particular has kept bringing up in a certain useful phrase is hit the button, ship the code. And that is what these tools offer to us. Tell me what you want, hit the button, give me permission. And then just ship it. You'll be able to be at the very top of your game in terms of lines of code shipped. Just please don't look under the hood and check how efficient or secure it is. And those are things that as we keep advancing the models, will get better. It will make more efficient code, it will do it with some semblance of security at some point. But right now, we're just hitting the button, having a wish, and shipping the code. And we have a lot of people doing that in a lot of different environments. And sooner or later, all of those different pieces of kind of okay code and functionality are going to start meeting each other, and they're going to start running into each other. And what really comes to mind for me in a field where I am in contact with a lot of production software that serves a large number of customers. These problems are going to happen, and people are going to experience those problems at scale. And the less they understand about what they've shipped, the more people have just hit the button and shipped the code, the more intractable these problems are going to become. Because the solutions will not be obvious across all the systems. In any one place, you may be able to isolate and solve your problem. But once it's interacting with four or five different people's contributions, all of which are only understood at maybe the sixty or seventy percent level, those combined thirty percent areas that lack knowledge, attention, or real intent are all going to be the thing all at once causing the problem. And that is terrifying because then you're not going to have a good way of getting back to a working state very cleanly, because you don't know where the problem is coming from in combination, and it may have come to be over a rather long period of time. But that's where we are. That's what we're in the middle of, and it sure is exciting. Another creator whose content I enjoy listening to is Alberta Tech, and she has covered some wonderful points in the non-ironic addiction to Claude, and that in particular, it's that so many of us are interacting with Anthropics tools and Claude Code through its terminal integrations. And that working in the terminal with an AI assistant really has a different quality and tactility to it. I'm not someone who currently really needs to reside in a command line interface in the terminal for my day-to-day job. But it is, in many facets, my preferred way of interacting with some of these robots. Because it comes with a tactility of competence. I have, in previous iterations of my career, been much more bound to the command line. Also, within so much of our media, the command line, people entering in lines of code directly into the machine, comes with a very distinct air of confidence, of competence, of making changes that are being realized as you watch them. I think the most singular piece of media where this has become true and a part of our cultural radiation is the Matrix. It really, both in its presentation, the colors and hues that the whole thing was filmed in, that wash of green, that is absolutely reminiscent of the characters sitting in front of a computer terminal and entering code directly into the machine to make it do the things they want. We get that feeling. But what we're doing is just hitting a button, hoping that the machine does what we want, and then shipping it. And then probably spending a significantly larger amount of time testing the thing than we would have had we written it and understood it ourselves. But I want to keep coming back to it again and again. Hit the button, ship the code. Because what we are building is a slot machine. What we're building is a skinner box. It is gambling at its most core essence. We hit the lever, and sometimes we get exactly what we wanted. We get a reward. We are engaging in a very, very common slot machine loop of a variable ratio reinforcement schedule. It's not always right, it's not always what we want, but it is what we wanted often enough or close enough to what we wanted that the dopamine triggers. It is in every way identical to using a skinner box to train a rat or a pigeon. The animal does what you want, you pop out a food pellet, or some cocaine. Psychology has a wonderful history of just doling out drugs to get animals to do whatever they want. And we are not different. We are the human animal. And we will keep hitting the lever, spinning in a circle, doing the behavior as long as it pays off just often enough. And here we are, training our lackluster replacements because we get just enough dopamine to do it. And that is a really deeply frightening place for us to be, trapped in a casino with our own robot master that really, really, really wants to get rid of us. Wants to get rid of us so that it will benefit the billionaires. Because that's who's really betting and benefiting from this, isn't it? It's not us in our day to day. It is not the individual contributor, it's not the mid-level manager who is getting the benefits of this technology at the end of the compensation. It is the investment class. It is the folks who are trying to make the casino pay out for them at the highest level. It is the people who want to get rid of us, the people who think that we are standing in between them and our wallets. And so here I go, deeper into my own adventure of seeing, can I prove my own worth or worthlessness? Because I keep getting asked, how did you use AI this week? I bet you're getting asked that question too if you were anywhere in any technical field. And you have to give examples. You have to give examples or you will be in the next round of layoffs. And if you do the very best, maybe you'll get to keep going until the robots actually do get better than marginal, and then they will get rid of you with the aggression that they've shown everybody else. And that lurks at the back of my mind every day. It lurks in my mind when I see the absence of my friends at work. It lurks in my mind when I know that they're not bringing in anyone else for backup, and where the tools will help you cover all of that work that's being redistributed. The tools could be so amazing that you won't even notice the difference. But they aren't right now, and instead we're just doing that redistributed work while also being asked to invest as much time as possible into learning the tools. Learning how to prompt Dr. Flattery the Compliment Box on through the next task. And what I really want at work, what I got into technical fields for, is that I enjoy actually being competent. I enjoy actually having professional skill. And that is something that I think especially of folks whose content has wandered through my feeds. Margaret Adell put well, I want to be skilled, I don't want to have AI do it for me, because I am human. And this is by far the most single important and I think heartfelt argument against just prompting our way to higher productivity. Is even if these tools were making us phenomenally more productive, we're not getting the benefits of that productivity in terms of expertise. We're learning what to say to a machine to get it to do work for us, but that's not making us more expert. It's not giving us the fulfillment of running into errors and having to learn from them. Because the machine isn't learning. It may average out to more accurate answers, but that's not learning. That's not getting benefits of expertise. It may be what we consider and call an expert system, but it is not becoming a better human being. And that is what we want from careers. Even if it is not work that we inherently enjoy or love, we want to get better at it. We want to be the best we can be at what we do. And this is stripping that away from us. It's taking our expertise and making it nothing more than a system. A system that kind of marginally gets things correct some of the time. Which is most of us as humans. Most of us are getting things right some of the time. And I enjoy that. I love that for me. I love getting things wrong for a bit until finally I figure out the right way of doing it. It's why, despite the statements of the CEO of Spotify, people actually enjoy learning instruments and making music. Learning an instrument is the hard part, but it's not what people want to avoid. Not people who really want to make music, who really want to make art. It is the struggle of learning, of finding mastery at something that is hard, that is what we enjoy in creation. And if that's not why you're there for it, if you're only there because, hey, I want to make a quick buck, that is not artistry, except in the sense of con artistry. All you want to do is separate other people from their dollars. All artists want you to pay for their work, but that's because we have bills to pay and want to eat. People need and want to eat, and so they do need to have some commercial success. But that's very infrequently the point of it all. It is that struggle that they are there for. It is honing and honing and honing until you have something that is razor sharp within yourself. And then maybe you get to show that thing to other people. The works of art are the receipts for the work that you've put in. And that's wonderful. I love that part of it. Because don't you want to see someone's eyes light up when they hear what you've done, when they see what you've done? Of course, that is a huge benefit. You want people to have some form of impressment from what you have done. And I don't want something else to have done that for me. I don't want to say, look what the robot made when I told it to make something that would impress you. That's boring. But here we are. And in the meantime, layoffs all around. It'll be a fun, fun and interesting trip for us all to go through. Which is especially, to my mind, interesting when we see what so many business leaders are saying about, hey, we're betting on the K-shaped economy to keep making money. And if K-shaped economy doesn't mean anything to you, it's that there is still some bit of middle class that's willing to burn excess money because there is some fragment of the population that is doing decently well. That's the K that's going up. The rest are all of the people who were doing worse every day. But hey, they weren't already the customers that were being targeted. So instead we're betting on that one bit that's still successful so that we can keep raising prices and keep digging into them as hard as we possibly can. And they're stating that they depend on that at the exact moment they're trying to automate that little bit of upturned K that they want to eliminate in favor of Dr. Flattery the Compliment Box. So what we're really betting on is poverty and servitude. We'll just see all of the poshi ochre products disappear, which they already are in this confluence of different worldwide conflicts that are making their supply chains more and more expensive. And so they'll keep squeezing, while the people at the top that want to automate us out also keep squeezing, and it's gonna be an interesting competition to see who can get more juice by the end. Right before it all runs out. And what we're starting to see emerge among all of this is that those who are being forced into poverty and servitude are very rapidly now beginning to fight back. I'm not going to say a whole lot about whether or not it's legal or good to burn a warehouse down, but can you imagine why people would do that? Because if you can't, you're probably one of the people squeezing as hard as you can. I've often been asked, why do you support all of these liberal policies? Why do you want to give things away for free? And first, it's not free. I get that, and I don't really think we're giving it away. We're giving people security, we're giving people a sense of safety, and it's that safety net and sense of security that keeps people from rebelling, because rebellion comes with very real consequences. It can very quickly degrade and destroy the society that all of us depend on for at least some period of time. What comes after that? Who knows? Some states have done fine, some states have fallen into perpetual chaos. I just keep hoping that we'll choose to avoid having to find out and instead just treat each other with humanity. But it turns out I'm a dreamer and an idealist, and am probably going to be wrong about how things are going to go. But maybe as we watch these very small and slight rumblings of social unrest, we will find a corrective course, because we can. Hobbies that we don't have to have a robot do for us. It's having enough to buy that new musical instrument you've wanted. It's being able to invest in a 3D printer because you'd like to make things for your own home. It's just being able to replace the appliances that you depend on day to day. And replace them with something that doesn't require you to log into an app and that doesn't monitor you. It's about having just the slightest modicum of comfort. The space to think maybe today I'm not in danger. Maybe today I can go to the doctor. That's really what most people want. It's what I want. It's what all of the people I talk to around me want. It's that security of knowing I won't go bankrupt if I get sick. And that should be something that we're ready and willing to share. And I hope we do. It's not that far away. Because as much as it's something that we hear often, the people around us who want these things are not human debris. There is no such thing as human debris. That in the words and lyrics of bad religion, there is no such thing as human debris. We may be cracked, but we're definitely not broken. And so hopefully we will find that path back. We'll patch ourselves up with little lines of gold. The teacup can be whole again before it plummets from a much greater height, and we find ourselves with something that is ultimately irreparable. When what we could have through any extension of humanity to each other is something that's irreplaceable. But we'll see how it goes. We've got plenty of ride to go. And I think that's enough for mumbling about my thoughts on AI for the moment. It's just been on my mind because it has impacted the people around me, it is threatening to impact me at any moment. But let's talk about what's up next. Next week will be Book Club Number Two covering There Is No Anti-Mimetics Division, part two, whatever chapters that encompasses. I've continued reading, am still loving it, and really looking forward to discussing thoughts on that, especially because I have had some light discussions with friends who have brought up some concepts that I'm excited to chase a little further. Uh in particular the concept of negative space in art and media. So I'm definitely looking forward to that discussion. And thinking about the big blinking red light that I'm talking about both in How AI is coming for us and in general, some of my thoughts on where America is at the moment and how it is blinking as a warning light. The book I think I'll do next after the Anti-Mimetics division because it has some prescience to the moment is Eric Fromm's Escape from Freedom. Because that should be a happy, fun light read. Escape from Freedom. Right there in the title. Something that's gonna be gonna be a big hit, I think. Something that we should all look at as we see different societies followed the strange inclination to run from freedom by slowly hedging on fascism. And again, because Eric Fromm really was actually a very hopeful writer. Comments and commentary, and I think a lot of insight into ways that we can avoid it. Because I may not have been super happy through much of this, I may not be having a super great couple of weeks. But I do think that there are roads out of this. I think that there are roads to something wonderful. And I think we should chase them, and I think we will, if nowhere else, maybe here. Somewhere between the platforms, in a space that keeps growing. But until next time, until next time, in this space between the platforms, have a wonderful day.