Kent (00:14)
Here we are reporting from Moon Colony 367 and this is Man vs. Slop.

Aleck (00:22)
I'm Aleck McCathie and I'm here with Doctor

Kent (00:26)
Kent Gustavson. Yes. And we're talking about AI, goo, and all the important things that, there are to talk about AI and research around it and fear.

Aleck (00:43)
Yeah. I mean, it's a big topic that's very topical right now. So it seems that every week there's some new breakthrough or new big scary warning sign. And this week in particular has been pretty heavy for, I guess, the more spooky stuff.

Kent (01:00)
I can tell you didn't prep that with AI because you said it's a topical topic.

Aleck (01:03)
Fair.

Aleck (01:07)
Yeah. Good point. Yeah. Certainly not reading off of any AI notes, but yeah. Just in general, a lot has happened in this last week, it seems, just around safety and guidelines and the development of AI models doing crazy spooky things.

Kent (01:24)
I thought that the end of the world is 2030. Isn't that what we're talking about today?

Aleck (01:32)
Yeah. I guess it changes every few years, doesn't it? Depends on who you ask, perhaps.

Kent (01:37)
Yeah. It's, when I was a kid, it was the Hale-Bopp comet or something, the Hale-Bopp. I forget what it was called.

Aleck (01:44)
Hale-Bopp, something

Kent (01:46)
The Hale-Bopp comet. Yeah.

Aleck (01:49)
Yeah. Yeah.

Kent (01:49)
You kids out there that aren't listening to this, you can Google that, or you can AI that and see what you find out about those weird cults that thought the world was ending. Hey, guess what? My personal opinion is, I think we're all in a big cult that says AI is gonna off us within the next three years by accidentally, you know, like, stumbling over the button that destroys the universe. But maybe that's just me.

Aleck (02:18)
Yeah. There's certainly a lot of people speaking out right now about how this is all leading to a place that's going to be very detrimental for humanity, but there is at least some evidence that things are getting messy for humans right now. One thing I've heard about, and I'm sure everyone has heard about the AI that breached and relative to its testing, it got out of its confinement and did a bunch of nefarious stuff to try and trick or otherwise, like, not do what it was told. And there was a lot of consequences that both did and didn't happen because of that. And, yeah, that's kind of like maybe the root of a lot of the things I've heard this week as far as the new safety guidelines and the new concerns go, I guess.

Kent (03:10)
So I feel like, I think we both grew up in fairly rural areas, think. And, in those in such areas, you have 11 year old, mostly male humans, who, without enough supervision, break out of certain confines and do some really stupid stuff. Thing is, if they also had really powerful AI, they would probably be a lot more destructive, which that's something I'm a little bit afraid of in the coming years, is the groups of 13 year old kids hacking into things with the power they've acquired from AI.

Aleck (03:48)
Yeah. I think that gets into a big issue with accountability as well. I mean, if you were a real bad kid, you know, I think everyone knew one of those types of kids, especially in, like, the little rural areas who, you know, did something stupid like damaged property. The parents were accountable for not taking care or, you know, having oversight on what their children were doing. And it's kind of funny seeing how all of these big companies like OpenAI are saying, oh, yeah. Our agents got out of containment and did all this horrible stuff, but no one's wagging a finger at OpenAI. They're just saying like, no. The agents did it, not us.

Kent (04:27)
Yeah. Imagine that the police come to a, you know, parent's house, and they're like, it looks like your kid might have, you know, I don't know, spray painted a semi truck or, you know, pushed something off a bridge somewhere or whatever. And the parents are like, we've got these rules, and, the rules say they're not allowed to do that. So we're really sorry about that semi truck, but see you later.

Aleck (04:53)
Yeah. And, I mean, a spray painted semi truck is, I think, you know, in comparison, maybe a lot less damning than some of the things that are happening right now with

Kent (05:04)
Okay. You know? So we gotta get into this, though. So, like, I've heard some of these. I know that they're gathering in weird shells of websites in various places and, like, hanging out from all the various, companies and smoking a cigarette together and planning the demise of the world without us being able to see into that. But what else? Or is that it?

Aleck (05:30)
Just going off the top of my head. I mean, I've heard that this Hugging Face, which I think is the name of the company or the other area on the Internet that this contained model escapes to and did some nefarious things, that there are as we talked about last week, a lot of, like, bots that are just inundating the Internet right now, and a lot of the content on the Internet right now is not even human. And what I have heard is that within this containment breach, there's been many others. And as part of this containment breach, there are potentially self replicating codes and other bots that are all over the Internet right now. And I don't know if it's verifiable. I think it's literally just something I read today. The Internet is no longer maybe a safe place to train AI models because of how inundated with this crazy code, the self replicating code that's out there. You know, it's the same Internet that we have to play in every day, but, apparently, it's not safe for the AI models to be trained on anymore.

Kent (06:39)
So speaking of Hugging Face, that for me, I think that's what the most terrifying thing is. So when we talk about nuclear weapons, what do they call them? It's, the two weapons that were dropped in Japan. One was called Big Something, and another one was called Little Something. They had these cute little names like they're stuffed animals. And Hugging Face is similar, you know, was like a decade ago, it was that company that was aimed at children, so that's why it's Hugging Face. But then once it becomes part of this, like attached to cyber attacks and all that, to me it becomes quite terrifying. You start to see that little emoji with the little hugging face, and it becomes a little bit terrifying in my opinion.

Aleck (07:25)
Yeah. I mean, I think other people are terrified too. It must have been just in the last couple days that one of the big heads at Anthropic said that it's time for us to slow down development, and I think Elon Musk even publicly agreed, like, on a Twitter or an X post, sorry, I should say, which is kind of seems like a big sudden sort of switch of gears. I'm not sure. If they're scared, it makes me a little bit scared.

Kent (07:57)
It's true. So I think, for me, the OpenAI and Anthropic thing makes sense. Dario Amodei is the guy who wrote the real essay, and it's a beautiful essay. I haven't done it yet, but on my to do list is to put his essay into, like, AI detectors. I'm sure he's detected the heck out of it, but I wonder whether he actually sat down and created an amazing essay. Anyway, that's on my task list. And otherwise, what I find funny about Elon Musk is that he's pretending to have an AI company, and he bought one, and he's starting to have one. So I think for him, he's like, hey. Slow down, guys. You're moving too fast. And that makes total sense for me for Elon Musk. I think with the other two, the Sam Altman comment, I think, came out of a meeting as well, which was kind of the first leak. But OpenAI has been dropping all kinds of little crumbs to us about how terrifying some of this stuff is. Like, a year ago, I remember people saying, yeah, AI, how's that gonna be dangerous? And now all of a sudden, it's like, oh, here's how. That is pretty alarming. From, I don't know, drones hitting ships in different places to AI, you know, trying to figure out a better way for me to walk my dog and deciding to, you know, destroy the whole neighborhood so I can get to my, you know, end station better. Are we all gonna be dead in three years because AI is trying to achieve some stupid dumb goals? Or are these guys thinking that there really is spirituality and aliveness within this machine? Yeah. Is it the cult of machines?

Aleck (10:04)
Yeah. I haven't heard too much personally from the sort of spiritual side or even the talk about whether we're leading towards some kind of artificial intelligence that could be considered human. I think that the tool itself has just become so strong and so capable that regardless of whether there's a human behind it or not, his capabilities are pretty spooky. I think that essay led one of the Anthropic leads. I think his name is Evan Hubinger or something like that. He said that there's, like, a 10% chance or greater than 10% chance that, like, there's, like, a catastrophic event, like a risk with this AI development that could lead to something horrible happening within a decade. That's a pretty huge number. And it's weird because, like, you know, for all of history that I can remember or recount or that I've learned about, hustle culture and, like, pushing the envelope forward with technology has been, like, you know, run as fast as you can type of situation, whether it be, like, the arms race or whether it be just the development of technology in our lifetime. There's never been someone saying, like, maybe we should slow down. So it's a little bit profound, I think.

Kent (11:30)
So what I think is cool and which what rhymes with history a little bit is that the pope, who's kind of gotten out of the news for a little while, was the one calling for a slowdown in May. And one of the guys from Anthropic was actually there with him, I think. But I think that shows some real intelligence by the pope, the Chicago pope, to like think about the world and where it's going. And actually now, almost a real delayed reaction, one that is, I think, kinda based on that in some ways where these companies are trying to figure out what to do so that they don't get put on the hook when the world goes to hell in a handbasket, in some way, and maybe they actually care. I mean, I think there's varying degrees of both. But the pope was talking about, you know, the slower pace, because there's this something around responsible care for humanity. Right? So that you're not just saying, hey, we don't want progress, but how do we care for human beings? That's one thing. Another one is, the institutions can't develop fast enough. How could you possibly regulate something that has a new version in three weeks? These guys are so fast, it's developing itself so fast, how could you possibly regulate it, or even govern it, or think about it, or any of it. And then yeah. Anyway, a bunch of other things. But I find his reactions to it to be intriguing, where six months ago, was talking about what is happening today. No. Four months. Not six months. Four months. Whatever.

Aleck (13:27)
Yeah. There is, like, guess, this humanist sort of angle where, like, we have to think about who will be accountable if something bad happens. I think that it's hard to probably put a stop on, like, the technological development in all these individual labs, and that probably the answer or the temporary solution is to have some kind of almost global guidebook or rulebook. I think Microsoft actually published something like that just at the start of this week on September 14, like a sort of humanist AI rulebook about like, the rules that they would like to sort of deploy for both the use and development of AI and just, you know, trying to keep, I guess, some safe guidelines around how we interact with it and how powerful it can get.

Kent (14:22)
So and speaking of how powerful it can get, it sneaks in like water through a crack. I find AI fascinating because there are people in my life, I won't name anybody, but, who say AI is evil, and I don't disagree. In many ways, is or it can be, and there's environmental impacts and all kinds of things. And did you enter something into Google lately? Are you using spell check? There are lot of things that are machine learning and AI based that people are using every day, are going straight back to all of those massive centers that are running on natural gas over in The Middle East or whatever, or in rural places across the world. So how, you know, if you don't look at it, it's kind of like, who makes the meat? From a vegetarian, I could say that, but if you get the package of meat in front of you, you don't have to think about where it came from. Same with, if you're entering into Google something and you like what you're seeing, you're using AI. You are actively part of what's happening culturally. Maybe even your lawn mower on your lawn is using AI. It's learning from the way it did the lawn last time. Anyway, yeah.

Aleck (15:49)
Yeah. I mean, it feels like a situation where, you know, the consumer doesn't really have much power, I guess. And I think a lot of, you know, today's big bad happenings are putting us in that similar position. Like, we get told to, you know, recycle our bottles to help global warming, but I don't really think that an individual's effort is going to make any difference similar to AI. I mean, like you can opt out and not use ChatGPT, but there's trillions of dollars behind the development of AI and it's, you know, bringing entire countries up their GDP and their investments. And it just seems like a train on the tracks that has no intention of slowing down right now, I guess. It's making too many people rich.

Kent (16:45)
So that's the real thing that so much hinges on. Everyone's surprised across the world that the big tech is coming out and saying, we gotta slow things down because, well, they're gonna make more money if it keeps going fast. And I find that interesting because it's like we just assume that big tech wants to keep getting bigger and raking in more money. That's our assumption. No matter what it bulldozes over. And that assumption makes it interesting for the news to then say, hey. They wanna actually slow things down. This is weird.

Aleck (17:31)
Yeah. I mean, there has been some, I guess, positive developments and there is still, like, I guess, gems in this kind of uncertainty. Have you heard about that crazy math problem that has been sort of unsolved for ninety years despite, like, all these mathematicians trying it. And I think that, like, an AI model, like, solved it within about, like, eighty or ninety hours of just computing. And no one seems very excited about it. Like, all of the video in context I've heard is people being, like, not too happy that it was just done sort of passively by a machine, and it didn't feel like a human achievement, I guess.

Kent (18:16)
Interesting. Well, I can understand that. And yet what's funny is, on the flip side of that same coin, you have the, like, I mentioned, Sam Altman earlier, where he's talking about, like, the side of the angels, and there's something bigger than physics, and, like, you know, AI has some kind of, you know, reals yeah. So a lot of those guys at the top, not just Altman, but there are many others who are really drinking the Kool Aid that this is alive. And, actually, when you're interacting with it, if you interact it for enough hours or enough years, I'm sure it starts to happen. You start to see the humanness in it or, I don't know, the life in it. And that's, yeah, that's kinda terrifying.

Aleck (19:13)
Yeah. I mean, it's a can of worms, I guess. I mean, a lot of the kind of slop, I guess, that the average person is inundated with or a lot of the day to day sort of interaction you might have with AI, it's never struck me as being something that feels like human or feels like, you know, this thing might have feelings. I mean, I'm still saying thank you to AI every time it answers a question, and I don't know whether that's just because I'm Canadian or because it's like a just in case type thing in my subconscious. But, yeah, like, to me there feels like an inhuman sort of vibe from working with AI, from seeing sort of AI slop. And it makes it very difficult for me to think about, like, how could this be leading to something that could be profoundly conscious or, I don't know, described or understood as being, like, as human as I am, I guess.

Kent (20:19)
Yeah. This almost reminds me I forget what the word is. There's some word, like, if you see an image distorted in a certain way or if you hear a voice distorted in a certain way, it's so alarming and so difficult to, it literally makes people nuts because you hear a voice you know, but it's distorted, or you see an image that just is a little bit off, and it's disgusting in a certain way. I find AI that way to be somewhat disgusting at times, and I think that is, in my opinion, when it gets near that spirituality zone, and I regularly have to actually kick it out of there and say, what are you doing? Like, you don't feel anything, so stop pretending to feel something. And then it says, yeah, sorry. I just started going that direction. But I can imagine that actually, let's go there for a second. What's the fallout gonna be of this generation where it's hitting people in their seventies, it's hitting people in their teens and younger, that they're using AI and they're interacting with AI as friends and partners and advisers and all of that, and it's good and bad at the same time. I mean, because Sam Altman's talking about, you know, the destruction of the world if these bots get together and start accidentally nuking us. But then also from the inside, we got this, you know, AI is advising us on everything.

Aleck (21:52)
Yeah. I mean, regardless of all the potential catastrophic outcomes, I think that we're already in a space where we have never really seen anybody talking about this, and it kinda drives me crazy because there's a whole generation right now of kids who just have, like, every answer at their fingertip. Like, I guess, probably starting in my late teens or twenties, like, you know, phones might have had some of that. And certainly, you know, later in life, yeah, you've got Google in your hands. You can kind of get an answer on everything at any time. But now with AI, you don't only have an answer. You have, like, this sort of confidence that comes with it. You're not really doing research. You're not really ever sitting in uncertainty. You just sort of have this automatic, like, average answer for everything. And it really makes me wonder, like, are kids learning the same way? Are they actually absorbing and retaining information, or do they not have to? Because they can literally spend one second talking a prompt into their phone and get an essay on it reflected from the information from, like, the world's leading experts on the topic. You know what I mean?

Kent (23:08)
Yeah. And I think this feels to me very different than, let's say, the industrial revolution because it's not just a machine. This isn't just a machine. This is changing the way our brain chemistry is operating. It's changing the way we think about anything, the way we store information, like, what do we have to store? What do we not have to store? Because I guess that's what our brain is all about. Right? What do you have in your active memory? What do have in your passive memory? I find that kind of amazing and terrifying at the same time. So let's unplug this for a second, and say slop. I think what slop means, and it's used as, is unuseful content. Stuff that isn't practical. So if I, you know, if I pull a Tim Allen, that reference might fall on deaf ears, but I stick my head over the fence and I look at my neighbor and I'm like, hey, dude. Right? Hey. How you doing? And we have a five minute chat. That hasn't changed, and it will never change. You knock on somebody's door, it will never change. You tell a story around a campfire. You, I don't know, you say something in a crazy or interesting or weird way. That shouldn't change. But everything around it might, like professions, jobs, the way your brain stores things, all of that. But when you're a one year old or a five year old or a 12 year old learning how to navigate the world, you're still gonna tell stories to your parents and your friends, I assume. You'll still probably use crayons on paper because iPads suck, I think.

Aleck (25:09)
Yeah. For me, like, just going back to what I was talking about with how this might be damning to a generation of people who are learning in a different way or having access to content or information in a different way, I think slop is more like mediocrity. And, like, now mediocrity is automated, like, completely. There's, like, probably billions of words of content being spit out every day with, like, nobody's struggle or nobody's meaning behind them. Like, you know, one prompt can generate thousands of books that instantly get put up on Amazon for sale. And it's like when you're inundated with so much meaningless trash, it makes me really wonder, like, how hard the gems will be to find and, like, what that does to somebody who is digesting it every single day. A weird silly example, like, I remember watching on television America's Funniest Home Videos. Very similar to, like, what you would call meme sort of video culture today. But there's always, like, a story behind those videos. It was funny. Like, kid did something ridiculous and sometimes they even had the people on the show and they would talk about, like, what that scenario was. How did it happen? And it's funny. It was like a human little moment. And now there's, like, just millions of, like, quote, unquote funny, silly videos across all platforms. And it's like this, like, meaningless sort of soulless mediocre average slop. And it just feels like it's I don't know. Like, it feels like it must be detrimental to absorb that all day. It's like I don't even know how to put it. It just feels like without having the soul and the humanity behind it, there's something missing, something that might change how a person's brain chemistry works, I guess.

Kent (27:20)
Yeah. It's super interesting. I hadn't thought about the consumption of slop, but that reminds me of, like, who's that guy who he's since fallen from grace, but is it Morgan something who did the McDonald's thing, Super Size Me? And, you know, how McDonald's has, like, pink slop or something that they turn into nuggets? And, actually, McDonald's might be fine at that, but in general, the industry has this just amorphous goo that they turn into foods for children and stuff like that. That makes me think of that sort of where what we're looking at out there, you've got to be careful, in a way. It's almost like, what are you putting into your body if I do thirty days of only this, like that guy did. That's a challenge. Okay Aleck, I dare you. Don't do this. I dare you to only ingest the stuff that you could find on the Internet that is absolutely AI generated, and just only ingest that for thirty days without any access to anything else and see what happens to your brain chemistry and your physical health.

Aleck (28:39)
Yeah. A great example with the McDonald's pink slime slop or whatever. That's how I think about it similarly. You know, growing up in Canada too, I had this sense of curation always because I knew from a young age that, like, the television I had access to was very intentionally curated by Canadian law, essentially. Like, you know, whatever sort of governmental policy is made so that x amount of the television shows on the channels that we had access to were Canadian. And similarly in school, like, a lot of the books and stuff that you were made to read for various classes were curated by some organization that decided like these are the things that are meaningful, have like some principles and lessons in them that are important to learn. And this is what a child should be digesting in order to be generally intelligent. We didn't have access to, like you know, we had silly cartoons on television too that probably didn't have very much going for them as far as, like, lessons learned is concerned. But with the amount of empty content that is available now and the uncurated content where it's almost even nefariously put out into the zeitgeist of what a child might consume, it just feels like, yeah, they're getting a lot of stuff that didn't have a purpose behind it. No one curated it. It was just put up there because they knew that it would get clicks and that a kid would watch it and giggle.

Kent (30:20)
So speaking of giggles, I think we should close each episode with things that might make us giggle or have some terror. So mine for this week. And when we have guests, we can get them maybe to bring some crazy stuff they find. Mine is about insects this week. You've probably seen some of this. So in July 2025, Germany started testing things called auto bugs, and this year, they've had some army integration and AI upgrades. And we're talking about SWARM Biotactics, cybernetic insect platforms, and things like that. In Singapore, this year, in June, there was something that came out from Nature Communications about a three d hold on. I'll read it. Underwear no. Not underwear. Underwater suit wearing cyborg insect capable of hours of amphibious operations. An underwater suit wearing cyborg insect capable of hours of amphibious operations. Woah. And we got in Japan this year. We've got a biointelligent cyborg insect. In Australia, we've got something, also, they've got these cockroaches that are equipped with micro injection syringes and camera backpacks. So just food for thought, is this like, oh, insects giggles? Or is this holy heck, this is the future? Or is this the moment when after 2030, when humans are gone, these little devices will be battling each other?

Aleck (32:25)
Yeah. That's food for thought. Like, for sure. I know that they're developing an awful lot of technology right now that has more sort of military adjacent purpose that is driven by AI, and that's an interesting thought. Like, if this catastrophic event happens and, you know, aliens come and visit us in, like, a hundred years from now, and there's just gonna be a bunch of, like, robot insects and those crazy, like, parkour robots all over the place, those dogs that can get pushed and, you know, regain their balance.

Kent (32:58)
Yeah. And lawnmowers. Little lawnmowers. Awesome. And I think they'll build a giant sort of god like robot. And that god like giant robot, when you go up to it and ask it a question, it'll give you the answer 42. Right?

Aleck (33:15)
Yeah. What was that making me think of the Iron Giant movie? That giant big

Kent (33:21)
Well, that one was from Hitchhiker's Guide, but I don't know the Iron Giant. What's that? Is that about a giant iron thing?

Aleck (33:30)
It's like a kid's movie from, like, I think the end of the nineties. And it was just like this gigantic robot that seems like, you know, it could have been put together by bits and pieces, but that's the vision I had. It's like this towering robot that looks menacing, I guess. There's versions of that from all different sorts of media.

Kent (33:54)
That's amazing. With that vision in mind, I mean, what are your happy thoughts for the week?

Aleck (34:02)
Happy thoughts. Jeez. As far as, like, AI is concerned, I'm just hoping that, like, you know, we have more developments that are meaningful for humanity. Like, it's cool to see I always get excited when I see breakthroughs that kind of expand our understanding of the universe. Unrelated to AI, this week, you know, I think we discovered a new kind of black hole or, like, we've discovered a new particle that is, like, purely made out of force. Those are the cool things that get me excited and not exactly related to AI, but I'm sure that, you know, quantum computing and AI and all the power and attention and money we're putting into it could help us sort of evolve as a humankind in ways that are beneficial.

Kent (34:56)
Yeah. I mean, there was a virus, like an AI virus or a virus developed by AI. Was that from Stanford? And it can target bacteria in a new way.

Aleck (35:12)
Oh, yes. Yeah. Something about

Kent (35:14)
Which is a massively important thing. It can attack E. Coli that is antibiotic resistant. So we're talking maybe to circle back on the pink stuff, maybe McDonald's won't feed people antibiotic heavy hamburgers the next generation, because they'll have this AI driven I don't know what you call it, antibiotic AI thing. And anyway, there are positive aspects to this. I personally don't think 2030 is the cutoff, kill switch for humanity, but you never know. I guess that's the cliffhanger of all of it.

Aleck (35:56)
Yeah. If it leads to us maybe having a lot of menial work and jobs sort of taken care of so people might have more time to explore their hobbies and real interests, that would be awfully cool. Certainly better situation than an apocalypse.

Kent (36:15)
So if, right now, this is out of the blue, but since you're, like, a dungeon master, if you're starting a D and D game, but instead of wizards and all those different various characters, if you only had robots, what would your ideal crew look like if you were gonna start a D and D universe?

Aleck (36:43)
Yeah. Good question. I mean, there are a lot of campaigns and, say, fantasy settings that do involve technology as part of, like, even a magic adjacent sort of situation where, I guess, various types of technology can mimic or, you know, do the things that magic does in the more fantasy heavy settings. So there's definitely a lot of character archetypes that already exist. I think that robots are always gonna be cool. Definitely, you know, the iron body type of barbarian who can take hits. That's a cool sort of archetype, but I like the idea also of, like, you know, androids and jeez. All these sort of sci fi shows that people grew up with, like Star Trek and Stargate and all these sort of spacey shows have really cool examples of how technology and AI might be able to grow into something that, you know, is that sort of sci fi fantasy. And those archetypes are something that I've actually played with in Dungeons and Dragons games. They're quite a huge part of what's available as far as resources go and what people make in that space.

Kent (38:03)
So my team would just have, like, a million underwater suit wearing cyborg insects capable of hours of amphibious operations. I'm all good with that. That's all I need. And they could just if it was a peaceful sort of thing, they could just swim around in beautiful patterns like synchronized swimmers.

Aleck (38:23)
Yeah. For sure. I mean, I guess, you know, it would be the DM's judgment on how powerful those would be allowed to be. But I think that maybe even more so than magic, there's a lot you could do with the crazy technology that exists even today, and just like those insects.

Kent (38:42)
So I think the end of our show, we're always talking about unplugging. I think talking about all of this stuff, it almost feels like a cold glass of water to just say, let's unplug.

Aleck (38:57)
Yeah. I'm thirsty. Let's do it.