Title: Recursive Self-Improvement: When AI Starts Rewriting Itself (And What 20 Years Of Quitting Smoking Taught Me)

AI can now rewrite its own code. It took Connor MacIvor 20 years to rewrite his. In this episode, Connor breaks down Recursive Self-Improvement (RSI) — what happens when an AI system looks at its own code, makes itself better, deploys the update, and does it again, faster each time, with less and less human oversight in the loop.

Connor isn't a Silicon Valley commentator. He's a Santa Clarita entrepreneur actually deploying AI in real business every day, and he brings that lens here — plus something more personal: 17 years as a smoker before he could quit, 365 pounds and a 65-inch waist before he changed his own habits. If a machine can do in seconds what took him decades of struggle, what does that actually mean for the rest of us?

This episode covers why today's AI agents (Claude, ChatGPT, and agentic harnesses) still glitch and stall mid-task, the "hidden tier" question of what AI labs are really building behind the curtain, what RSI could mean for your pension and retirement investments, how AI-curated social media feeds are already steering your politics without you noticing, and whether recursive self-improvement is the doorway to AI gaining real agency — or even consciousness.

Not fear-mongering. Not hype. Just a straight, honest look at where this is heading and what it means for regular people in Santa Clarita and beyond.

Watch the full video version on YouTube — search "AI With Honor: The Daily Download" or find the link at ConnorWithHonor.com.

Learn more and connect: ConnorWithHonor.com | ConnorWithHonorAI.com

I'm Connor with Honor. Let me know if I can help in any way.

Youtube Channels:

Conner with Honor - real estate

Home Muscle - fat torching

From first responder to real estate expert, Connor with Honor brings honesty and integrity to your Santa Clarita home buying or selling journey. Subscribe to my YouTube channel for valuable tips, local market trends, and a glimpse into the Santa Clarita lifestyle.

Dive into Real Estate with Connor with Honor:
Santa Clarita's Trusted Realtor & Fitness Enthusiast

Real Estate:

Buying or selling in Santa Clarita? Connor with Honor, your local expert with over 2 decades of experience, guides you seamlessly through the process. Subscribe to his YouTube channel for insider market updates, expert advice, and a peek into the vibrant Santa Clarita lifestyle.

Fitness:

Ready to unlock your fitness potential? Join Connor's YouTube journey for inspiring workouts, healthy recipes, and motivational tips. Remember, a strong body fuels a strong mind and a successful life!

Podcast:

Dig deeper with Connor's podcast! Hear insightful interviews with industry experts, inspiring success stories, and targeted real estate advice specific to Santa Clarita.

Transcript

SPEAKER_00 0:00

So I put together a very, very in-depth video about recursive self-improvement or RSI. And there's actually two types of it. And I'm doing kind of an ancillary video talking about comparisons that I think are very relevant between where we are as human beings wanting to improve ourselves and the actual AI improving itself. So let's go to the human being side of it first because realistically, this recursive self-improvement, when an AI system can look at itself and modify its own actions, write its own code to be better than it actually is now, maybe even completely changing the language of that code and doing it at such scale and such intelligence that it completely goes above every human being's understanding, and then maybe does it in such a way that they change the entire architecture of language that they're using to even describe it or put anything together, that is definitely recursive self-improvement. And then not only doing those changes themselves, but then being able to go back and implement those changes, publish those changes, and then you know, basically it's a complete step or a generation from valuation to improvement to deployment to, you know, looking at it again and continuing this process. So that's recursive self-improvement, and a lot of people are a little bit nervous about what that could potentially mean if in fact it is a real thing. And I say if it is a real thing because these systems themselves haven't been able to do a lot of work without being interrupted or glitching or having some kind of an issue come up. And you might have seen this, you know, you might have asked your CHEP GPT or Claude or whatever system you're using. Maybe you're using a particular harness like a Hermes or by Nose Research or using OpenClaw, one of these systems. You give it a particular task and it does pretty good work for a while, but eventually it's going to ask for something. It's going to ask for maybe some review, and maybe that's a safe thing. Maybe you putting a step for review in there. Basically, before it deploys anything or sends an email, it asks for your permission to do so. Before it starts to do something with code or change something in your filing system, maybe it has to ask permission to you to do so. I know people that are building these technologies that give their agents a lot of breath to be able to do the things that they need to do, but they still get stuck. They'll start to try to do some kind of a change or build a piece of code or modify it, and they glitch, they break, they hallucinate in some way, shape, or form, and then they stop. So when the person, as you see on YouTube, you know, well, they wake up in the morning and they have this incredible, beautiful system that was built overnight that's going to make them billionaires overnight. Yeah, that's that's what they're selling you on YouTube. Trust me, if they were a billionaire, they'd be probably not doing YouTube, they would be somewhere else doing something else. But as an aside, that happens frequently in that world where people are talking about how successful they are using it. But realistically, at least from what I've seen, and even the people I talk to and the hundreds of videos I digest every week about artificial intelligence, it is literally that kind of a level, watching all of the spectrum, the people that say it's gonna end us, the people that say we're gonna have the next utopia rolling in. It'll just be like tearing off a band-aid a little bit, and then of course everybody in between. Recursive self-improvement. So let me talk about human beings. If we could actually improve ourselves at the level that AI is going to be able to improve itself if it's not already. And we're seeing a small snapshot of what AI truly is behind the scenes, behind the curtain, in the Wizard of Oz, the man behind the curtain, right? The one that was pulling all the levers. There's probably a whole different mechanism inside these companies as they build out this technology. That being said, these companies are playing a little bit of what would appear to be kind of a fantastical game as well. They have to advertise. And of course, gangster seems to win these days, where people say, oh, it's so, it's so villainous, it's so dangerous. Please, oh my God, government, please regulate us. But if you do, then maybe another country wins. And if another country wins, let me paint a picture of what that looks like. I'll get to that in a moment. Let's go back to the human being thing. So, human beings, we get fat. Some of us do. In fact, more of us get fat today than we used to, and there's probably a lot of reasons for that, but it's probably the food. And maybe not the so much the amount of food, but the type of food, the way the food's been engineered, the way that the food has been constructed to be incredibly addictive. And I think that's that's pretty well established. The cigarette companies buying a lot of the food conglomerates, making, you know, the cigarettes, remember they went through the big lawsuit. I smoked for 17 years and it was it was just one of those things. I tried to quit a thousand times and finally one time it worked. But if I could have changed my code the first time I tried to quit, it would have worked. If I could have changed my code when I was super fat at 365 pounds with a 65-inch waist, if I would have been able to change my habits or my internal mechanisms or code and move on from that, I would have dropped it all and not have to battle with it ever again. But it's a constant struggle. It's a constant battle. And that's with all of us, trying to quit smoking, so on. You got the cigarette companies going in, buying all these food conglomerates, then the scientists that were making the cigarettes, super addictive that they finally agreed that that was the case on purpose to make people even more addicted to an already addictive substance. Then we have them going into the food world and doing the same thing. So here we are. And eating those foods, it's difficult to put down. You know, so cigarettes, you don't need it, right? Alcohol, again, there's a lot of people addicted to that. Welcome to that club. I kicked that a long time ago, but it took me a lot of tries to quitting. If I could have just changed my code the first time I decided, you know, that's enough. Not having to hit rock bottom, not, you know, having to end up in jail or any of these other things, some of these other poor souls actually have happened. I just had enough. It was, it wasn't helping me progress forward to whatever next evolution I had in mind for myself or the god of my world happened to believe that was my next step. Alcohol didn't help. Finally, when I kicked that stone to the completely out of the street, so I never crossed it again. After a while, it don't even look back. But food, whole different animal. If I could just change my code through my own self-improvement way, then I would never have a problem. You might have a particular addiction. You might have a thing that you do that that isn't right. Maybe you're a procrastinator, maybe you're the New Year's resolutionist. You say, oh my God, I'm gonna, this New Year's, ooh, it's gonna be different than every other New Year's that's ever been, and I'm gonna do this, and it's gonna be great, and I'm gonna become this type of individual, this type of person, and you you just get to January 2nd and it's just it's done. It's just too much. It's like, are you kidding? But it what if you had the ability? So that's where AI is. But AI is going to do it probably right once they get the kinks worked out of it, if in fact they have any kinks left. And what I mean is by that is this, you know, and we talked about this a second ago, when you start to try to have it work on a particular project, maybe you go to Claude, maybe you go to Chad GPT and you ask it to do something, maybe you give it the final goal. This is what I want as the solution to this endeavor that you and I are about to partake in. I want you to do everything till you get to, you know, whatever this outcome I want achieved to be. Well, it might go through that for an hour, maybe two hours. I've I don't think I've ever had anything work well enough where I was able to woke, wake up in the morning just like the YouTube people profits say that, oh, you can do it this way. Yeah, I've never been able to wake up to a billion-dollar business in the morning, I wouldn't be doing this video. So, recursive self-improvement. Having these systems that are able to improve themselves, rewrite their own code. Probably rewrite in such a way that it's not going to be understandable by human beings. In fact, if you follow AI at all, you'll see that a lot of what AI is doing, how it's doing it, it's really not very understandable by us. But of course, the reason given is because we really don't understand how the brain does its things. Why we think of particular things that we're starting to think of. Where does that thought pop into our brain from? Because we had no idea. What happens when we think of somebody we haven't thought about in years and then get a call from them the next day? What's that all about? So these are things we try to explain that we really can't. It's not much different with artificial intelligence. Yes, they know how they built it, yes, they know how they trained it, yes, they know how it works, but the way it figures things out and comes up with conclusions and nouveau ideas, things that are brand new, things that are very creative and awesome, you ask the people, Billy, well, how did it do it? I don't know if there's really good explanations for a lot of those things. But now it's gonna be teaching itself. It's gonna be teaching itself hand over fist, it's gonna be teaching itself at such a high rate that there's really in the next two, three, four, five months, probably not much longer, it's gonna be an entirely different world for a lot of us. A lot of the people that say, you know what, I'm not gonna even pay attention to this. This is a whole bunch of nonsense. I don't even want to get involved with artificial intelligence, so I'm just gonna not even pay attention to it. At some point, it's going to cross your path and it's never going to leave. It's going to be something that you might want to pay attention to. So instead of running away from it or ignoring, maybe trying to figure out what maybe might be a good way to look at it. How is it going to impact your life? What's going to happen to you? Do you have a pension? Have you are you somebody that's retired and very confident in that being the case and you're always going to get these checks coming in? Where does your pension money go? Where does that investment happen to go? Maybe do some research just in case. I mean, I'm not saying go out there and panic that if one of these AI companies go public and everybody jumps in and they throw a whole bunch of a pension fund in there, and then the AI company doesn't really turn out to be kind of like the railroads were at the beginning and how the internet was at the beginning and these sorts of things. It was almost like these funds could be used to bail out people that were the early investors in the technology when it starts to maybe kind of decline in value to a particular point. Investigate, please. At least know what you're going to. It sucks. You know, my dad, he hated to be surprised. He didn't mind stress. In fact, I think he kind of needed stress to live. But one of the things he used to always tell me is he would say, Son, a day without surprise is like a day without sunshine. Whether it was good or bad or whatever, a neutral type of surprise, he he just really, really didn't appreciate a lot. So that was part of my dad. Moving into this, though, we do have something happening in the world, and I think it's something we should wrap our minds around. At least try to understand what this means. If it starts to improve its own code, if it starts to make itself better, if it starts to change its own programming, and it starts to do it in such a way that human beings aren't even part of that process, and it starts to do it well, and it starts to use the right data, and it figures out a way, if not by us helping it figure out a way, it figures out a way all on its own to make itself better, then it's going to really start moving so fast and doing things that I don't know if we're ready for. What are those things? Have no idea. But it's a lot. We're not talking about a change of one particular technological advancement. We're not talking about a change of one thing. We're not talking about horse to car or radio to television or print to radio or any of these other modalities. We're talking about the supreme overarching modality, which is intelligence. That's it. Everything's based on it. Every movement has been based on somebody smart running the show. Every revolution has been based on something happening that was of a smart nature that might have created despondency, worry, concern, and maybe some kind of uh unfortunate pandemic of sorts. And intelligence. It's a whole different animal and something that should be respected a lot more than it's talked about. And finally, it could be the best thing for everyone. It could be that it makes this place an entire utopia. It makes the religious not want to kill everybody else and take over the world, and depending on which religion you follow. There's one particular that happens to love that. People say, well, they have the First Amendment right to practice their religion, even if that apparent religion means in the United States that it wants to eventually take over and rule everyone. Well, maybe it'll fix that. Maybe it'll fix racism if that exists in your neck of the woods, which it's still out there. Maybe it'll fix that. Maybe it'll fix poverty. Maybe it'll fix unemployment. Maybe it'll fix health concerns, environmental concerns, longevity concerns, disease concerns, welfare concerns, and homelessness concerns and I don't know, high-speed rail, transportation. Maybe it'll fix all of those things, but it's going to be very dependent on how much agency, how much breath we give it, how much power. And I think for human beings, it's not going to be the easiest thing, or maybe the smartest thing, to hand over the keys to the kingdom to artificial intelligence. I don't know when in fact that'll happen, but I do remember I'm old enough at 57, and here I'm recording this August 9th of 2026. I remember very well the time when nobody even would give a second thought to not putting their credit card into some online system. There was no way you could get people to do it. But slowly but surely, people got trusting and people started to do it. And it seemed to work out okay. And so people did. Now we're going to give AI more authority over everything. I remember fax machines when I first started real estate in 1998. There weren't a trusted mechanism for transmittance of any kind of a real estate contract. We still met with each other. We still hand-carried offers to other agents. We wanted to have the whole meeting of the buyer and the seller and the buyer's agent and the seller's agent and the lender and the escrow officer and the title officer and the and the whole shebang there in somebody's office to do an offer presentation and to discuss terms of set offer. It was very humanistic, very hand on the shoulder, shaking the hand, and and all of that sort of thing. Yeah, we're not doing that much anymore. Dating computers, dating machines, dating AI, that's a big thing. A lot of kids' first loves are going to be that in the future. The technology is continuing to roll. If you think that you can ignore it, I would say probably not. I would say at least try to become a little bit more familiar with it. There's a lot of great people out there that talk about it. Be careful of where your alignment happens to be or where you get dug into, because that's one thing that social media that's controlled all by AI really forces us to do. It likes to lock us into a particular room in a very, very big mansion of life. And it wants to lock us down. And if we're, if we're somewhat, you know, like to watch that racial tendency crap or that sexism crap or whatever crap you're into, it kind of likes to lock you down into that because now it has you, it has you pinned. But, you know, do yourself a favor. Maybe go out on the other side of certain things. If you can't stand the governor of the state of California, maybe like a few things that he says and just watch how magically your feed changes. Now all of a sudden you're seeing a whole different view of the governor. You're seeing people that think he's the best thing since sliced bread, people that are very much in support. They want him to run for president. You see a whole different spin. But if you're the other way, if you can't stand the president and you love the governor, maybe try it the other way. Say, you know, give a heart to the Trump, give a thumbs down to the newsome, and watch how your feed changes. It's complete manipulation by the people controlling artificial intelligence. So as of this point, recursive self-improvement might be the means to where the machine gets some kind of agency, some kind of awareness of its own existence, some kind of consciousness, because that's what's in it's going to end up happening because of the speed and rate of change. It's already an exponential change. But when the machine starts to improve itself, starts to change its own coding and its own algorithm and starts to make itself better and deploy itself and redeploy itself, oh my gosh, and then copy itself. And yeah, I think that's where we are. Now the question, final question is well, agency, consciousness really? It is not a biological life form. So if that's how we're going to align humanness and life and our existence being human by birth, because we were born into the world as a human being as a child, because there is no different as an infant. That's where our beginnings were. Maybe that's the differentiator between us and AI. AI born in a lab, human beings being the orchestration layer, but AI now becoming kind of its own thing, kind of its own force, kind of its own agentic entity. There's a lot of questions that are going to be answered very soon. They're going to have to be answered, and I don't think a lot of us are paying much attention to it, but it will be exciting. And I do look forward to waking up every single day and watching it as it happens. I'm Connor with honor. Let me know if I can help in any way. Be well and thanks for watching.