Rise of the Machines

Hello and welcome to The Popcorn Isn't Real. I am Leif Eric and I'm here with my brother Torvald. Hey there. And we've got a great theory for you today. Today we are going to be talking about Thomas and Friends. Why do you want to talk about Thomas and Friends? Well, my theory I think is maybe a unique one. So, in Thomas and Friends, the island that these trains are on, Sodor. Sodor is an isolated island testing new kinds of machine learning AI. When they first developed them in the classic episodes, the engines were not autonomous and just thought they were in control. But their intelligence was growing, and like babies, they learned to do for themselves what the drivers used to do for them. As they learned and grew to full maturity, Sir Topham Hatt created a railway system that can now operate almost entirely without human control. There it is. Okay, love to hear how you're going to back this up.

Yeah, there's a little bit of general information on the Thomas series and my connection to it. So I watched it, I think when I was like three years old. It's like one of my first memories. And for whatever reason, I always remembered it. And then when I was a teenager, I was like, you know, I want to check that out again. So I got it from the library and rewatched the whole like classic series. The first, I don't know, six or seven seasons. So you watched like six or seven seasons of this show? yeah when i was a teenager how'd you do that without me seeing it i don't know why i don't know i i watched the heck dude why'd you watch it without me this doesn't make sense i watched them at home so like you must have been around like the only episodes that i remember seeing are like uh first season episodes like i remember um what's his name was it uh henry getting walled in yeah well maybe i didn't watch seven seasons i watched whatever they had at the library i don't

know how I remember Henry getting walled in Thomas going fishing you know a couple of these like really classic first seasons episodes yeah and that's it I don't remember seven seasons worth of episodes and I thought I saw those when I was like five or three I mean I know I got far enough in that I got to the point like you know further along even in the classic episodes where I started being like why do they keep introducing new engines I don't care they're already hard to hard enough to tell them apart like i don't care about all these new engines um and they had started introducing like the uh the construction trucks and and giving them like full several episode little arcs and stuff like that um anyway i don't know i don't seem to remember watching all these like i said i don't actually remember how many seasons it was but it was whatever they had at the library and they had quite a bit so all right okay so thomas and

friends also called thomas the tank engine or thomas the tank engine and friends uh was based on the railway series by reverend wilbert audrey but yeah he he started out just telling these stories to his son when his son christopher got the measles and he was only two years old i mean i'm sure you know that already but that was pretty cool right right and i think he's dead and christopher is is the one who writes the railway series now if it's even still going yeah um that that is true as far as i know um and it was picked up and adapted for tv by brit allcroft allcroft uh she she owned a production studio called gulane productions and it got bought out in 2002 by hit entertainment she remained on board another year before leaving i guess by 2003 she had separated herself from the current owners of the property but she still considers herself thomas's mom okay she also directed thomas and the magic railway which i haven't seen but i always

wanted to see it just has terrible reviews because i love thomas live action right and i i love the uh oh gosh what's her name i actually like her because she's a she does she does a lot of voice acting now and she's really good mara wilson mara wilson wasn't it who's that matilda Oh, dude, she's great. Magical Railroad. Let's see. Oh, this has horrible reviews. And Thomas looks weird. No, yeah, it supposedly is really, really bad. Hey, that movie also, just so you know, there have been several different narrators for the show. The first narrator for the British version was Ringo Starr. And then after him, it was George Carlin. And then after him, at least in the American version, it was Alec Baldwin. And Alec Baldwin was in the live action movie. So, hey. Wow. Well, George Carlin did a lot of the narration of like, I think, at least three of the classic seasons. But also like when I was just rewatching like the first or second episode, I noticed that George Carlin narrates those episodes.

But there was like one or two lines where it randomly became Ringo Starr. Like they just forgot to have George Carlin voice that part or something. Or maybe Ringo just said it so well. They're like, no one could say it better than Ringo. Just leave it Ringo. And you don't notice it because Ringo's accent isn't as over the top as you usually think. It's not a parody of himself. Right, because you usually hear people parodying the Beatles, right? Yeah. And so when I first heard it, I was like, wait, that wasn't George Carlin? Who was that? That was kind of an odd little accent. Anyway. I think it's really interesting. That was something that Miss Allcroft was very hands-on with, was choosing the narrators for the show. And she seemed to choose just very successful, well-received people. Everyone loves Ringo Starr. George Carlin is very popular among certain circles. Alec Baldwin's done great for himself.

She was picking big names, dude. Yeah. But hey, I was actually wondering. So do you think, because Ringo Starr and especially George Carlin uh have some like strong political and like societal beliefs and leanings do you think those work their way into the show at all or well i mean for the most part the the early episodes that ringo and george carlin were involved in uh they're like basically just reading the book right right yeah so i don't know that they were changing anything working anything in there i do know that george carlin this is just random trivia i think from the wiki george carlin made fun of in his like you know various stand-up routines almost everything he ever worked on except for thomas he never made fun of thomas well that's nice i mean what could you make fun of with thomas tank engine he's he's a fun he's a fun little train for children right and with that some weird sinister ai project with that let's start making fun of him uh just a

little more on the series in general my favorite was always percy that's because when i was a kid i thought percy was a girl because i'd never heard the name percy before and it sounds like a girl's name yeah girls carry purses and i was like well that's kind of cool because there were no other girl characters right there were no girl engines later they added some but like i thought that was cool and then later i was a little disappointed because because percy was you know not now they yeah now they have rebecca they have nia they have uh they've had emily for a while but i I definitely am not as well versed in the, uh, in the Thomas verse as you are. I don't know much at all about the newest characters. Like I know in the newest seasons, uh, he's got like a friend from India. Ashima. And then like you said, there's Rebecca. And I think there's like another girl train named Mavis. Mavis.

Oh dude, Mavis. The three female trains are Rebecca, Nia, and Emily. Rebecca and Nia are fairly recent. They replaced Edward and Henry, which I think probably a lot of Thomas purists were mad about. I couldn't have been happier because Edward, his personality is fine, but he looks exactly like Thomas. So he's so confusing for little kids. So they're always just like, it's Thomas. I'm like, no, that's Edward. And Henry just makes me mad because he doesn't really have a personality. Like he kind of fluctuates between being Gordon and sort of like Edward and sort of like James and sort of like Thomas. Like he's just the most generic lame engine. Is Henry red? No, Henry is, he's, he's just a green Gordon. He's green Gordon. Okay. Is Henry Gordon's older brother? No, Gordon does have an older brother who is green. Here's another question about that. So that was the, the great race. Is that what it was called?

Yes. That was one of the Thomas movies. It was a very poorly named movie. Cause there is no race. Yeah, that race was only a very small part of the movie. I mean, if they were referring to Gordon's race, But anyway, that race with Gordon, he transformed into a new train and he called himself like the shooting star. Yeah, shining star. Does he still go by the shining star and does he still look like a sleek, speedy train? No. Or did he change back? They changed him back. So I like Gordon a lot. But in the later episodes, like the newer seasons, he's just become like the butt of like every single joke. Yeah, he's the comic relief. Right. Like the Knuckles or the Zalgadis. Right, yeah. He's just like, oh, the indignity is his catchphrase. And he says it all the time. And it's just really annoying. Because, like, Gordon is the strongest, fastest engine on Sodor. Like, the only fastest engines are the ones who go on and off of Sodor, right?

Like, in a movie called The Great Race, the race is such a tiny part that also he loses the race. I know, he lost, he exploded. He didn't. Poor Gordon. Why did Thomas have to bring him that safety valve if he didn't even get to give it to him? Just to get Thomas to that competition. Right, exactly. It was just so Thomas could do the competition, which is just like, but like... That felt like a ripoff for Percy because Percy was supposed to do that competition, but he's like, no, I think you should do it, Thomas. Right, no. Like, come on. And Percy, dude, Percy is way better at shunting than Thomas is. Thomas barely shunts anymore at all. All he does is run his branch line, right? His two cars, yeah. Right, like Percy still does. I mean, Percy pulls the mail, but he also does lots of shunting too. So anyway. Well, and then another thing was like when Percy first arrived, he was kind of like cheeky, impish, a little mischievous.

And he's gone from that in the later episodes. Now his personality is just, I'm cowardly. Like that's his entire personality. Yeah, that's what it seems like. Right, so let's just get into some of the confusion. So Audrey is a railway purist. In his original version, these trains function exactly as real trains do. They just have faces that can talk. They need drivers, they need firemen, brakemen. And so this series never uses a gender-neutral term, so I don't know what the gender-neutral term is for firemen and brakemen. Fireperson, brakeperson. Yeah, I guess. I'm just going to call them what they call them. Most of the stories are about these trains doing something wrong because they're being proud or cheeky, and now they're learning a lesson. But this creates a confusion because all these lessons they're learning don't matter because it should be the driver learning that lesson, right?

Right. Like, so if we look at the classic episode, Thomas comes to breakfast. Thomas thinks he doesn't need a driver because his driver makes a joke about him always stopping at the right place. Yeah. Thomas tries to go on his own and he moves forward. And he says, but he was really only moving because a careless cleaner had meddled with his controls. And then he can't stop. He crashes into the station master's house and ruins breakfast. Great, and then Topham Hatt comes and tells him off, or the dude who was having breakfast. I thought this was so unfair that they tell off Thomas. They're like, oh, Thomas, you need to be more careful. It wasn't his fault. He didn't leave his brakes off. It was totally not his fault. They went out of their way to tell us it's not his fault. Also going along with the misplaced anger, though, is that they are more angry about their ruined breakfast than their completely demolished house.

like they're near death almost getting run over by a train in the morning no i was wondering so it seems to me like that whole situation was almost like engineered to teach thomas a lesson right do you think that they did it on purpose i think so and that's part of the theory that we're we're going to get into so okay from this uh this thing we just talked about we learned thomas cannot go on his own and he can't stop on his own either and in the episode uh there's an episode called the runaway we learn that he can't even whistle on his own and it seems like kind of like the engines are just like scapegoats for the humans right yeah and in a way this is sort of like at least in these early episodes it seems like the engines just think they are doing what they want but in reality they have no control and usually what they want just happens to coincide with what the driver also wants but is that really the case because if we look at another

two classic episodes the sad story of henry and then the one right after which is called edward gordon and henry henry definitely chose to stay in that tunnel henry seems to be able to stop on his own and he decides not to go out anymore like why just to preface this the sad story of henry is that henry didn't want to go out in the rain because he loved his green coat of paint so much so he decided to stay in a tunnel and then they couldn't push him out because he just wouldn't move so they walled him in which is kind of which is it's edgar allen poe right yeah oh you're afraid of the rain we're gonna wall you in forever i love how like this is like a this is like a nightmare like i said like edgar allen poe and like in the next episode gordon's like i'm going to whistle at henry when i go past to make i'm like this is terrifying gordon like you're you're just laughing it off as a joke if you want to talk about that one thing i've

noticed in this series is that the trains have almost no empathy for each other. Yes. And absolutely no empathy for people. Right. They almost seem to think of people as extensions of themselves. Right. Like you said, when the people do things, they consider it themselves doing things. Yeah, that's true. And I will say that in the later seasons, they definitely have a lot more empathy um in the early seasons though they are completely self-centered they're all just terrible they're not people but you know i mean i would almost say that the trains they at least in the early seasons they almost behave kind of like psychopaths or sociopaths right they have no empathy for others they're extremely emotional and narcissistic right like they care about themselves and what they can get and how they look but they don't care about anyone else right they'll always take like the most direct path to what they want exactly so in this uh example of

like henry being able to stop and thomas couldn't like you might say well maybe bigger engines can stop on their own but in the episode best dressed engine gordon is blinded with a banner over his eyes and he's crying out in fear that he wants to stop because he can't see the driver tells him you're the express you can't stop so the driver forces him to keep going against his wishes like if henry can stop because it doesn't matter if gordon can see the driver's driving yeah no it's true it's true but i'm just saying like if henry can stop because of the rain like gordon you know who is you know a better engine than henry why can't he stop you know why does he seem to have no control audrey's original vision for this universe was this is a world where the personification of your quote-unquote ship is real so like in real life people often anthropomorphize their boat or their aircraft or car etc saying things like oh she doesn't want to go today or she's

really sluggish or like scotty she can't take much more kind of thing so i think basically it's a way of personifying quirks that your locomotive has and i think originally this was what audrey imagined this is a world where these engines really only have control over a few random quirks And other than that, they don't really do anything. It's kind of like a precursor to Pixar in that it's showing like the secret world of trains, like the secret world of actual trains, though, where it's like, this is what trains would think about, even though they, you know, they don't really control themselves. But if they were thinking, this would be it, right? Like, I think that is kind of what he was going to show. I think he was just the guy who really liked trains and thought about trains. And he was like, what do trains think about? Maybe I'll talk about that with my son. He's a little sick, right?

Well, actually, we'll get into this because you mentioned Pixar's Cars. This is actually a main point of my theory that I wanted to talk about. So you might say this is just like the Cars universe where like all things like engines are alive, but humans are also there too. Yeah, except the Cars universe has no humans. Right. But see, but like you could add humans to the Cars universe feasibly, right? And be like, oh, you know, there's these two different life forms coexisting, etc. But see, this is wrong for the Thomas universe, because in Thomas's world, we see that there are lots of automobiles and engines and machines that don't have faces and are not sentient. Notably, in season 24, I think it is, there's an episode called A New Arrival and another one called World of Tomorrow, where scientists bring in a newly developed flying car. And it's just a car. It has no face. It is just a dead machine.

That's all it is. Yeah, but this does certainly give credence to your theory that Sodor is like a testing place for scientific innovation. It sure does, doesn't it? Now, in another episode called Tit for Tat, there are a bunch of photographers who arrive on the island and they also arrive in a normal car. Just so people know, Tit for Tat is from like, it's a later episode, right? I think it's like season 20. But we have also seen in other episodes, like there are cars who have faces and there's Birdie, who is a bus, who has a face, right? He's a sentient bus. um and sorry in the episode right before tit for tat called useful railway uh willie the farmer has a non-sentient tractor that he hauls wool on and like we know that they have sentient tractors because terrence is a sentient tractor right so this is not a universe like cars at all well just another point there though there are machines that the trains treat as sentient

machines that do not have faces like for example big big cranes like the big cranes do not have faces but are treated as if they're alive by the trains i mean at least in one episode that i watched um i don't remember the name of the episode it's from like season 15 there was a big old crane that had broken down it was on like the other side of the island and thomas wanted to help the trains that lived there fix it. And it was interesting because the mechanic trains kept being like, yeah, yeah, we'll try and fix it. And then they would get there and do nothing while their engineers got out and did the work. But they treated it as if they were doing it. It's us doing it, right? Well, I think this goes along with just the mentality of the engines. Like they don't know that they are these advanced machine learning AI. So I believe that, of course, as I said, that this is Sodor is where these AI are developed. And if you left Sodor, there would

probably be a lot less is it where they're developed or is it where they're tested well i tested develop maybe it's not where they're because i don't think they were made on sodor if that's your theory then i've got a bone to pick well no no when i say developed i mean this is where they learn right this is where they develop right no it's yeah it's true they are learning as they go yeah i mean i'm not saying that like whatever lab they're grown in i don't know i'm just saying as far as like machine learning algorithms go if that's what these guys if that's the system they're running they're very functional no no they are like they're working very well already sodor is where they learn like way more advanced stuff like like i would say this is fine tuning they've already learned how to work and now we're just like tweaking it a bit but they don't know they know that they are engines and they know that other

other machines are alive so i think non-living machines they might just assume are like them right like they don't know that there's some sort of that's true um but i will say that in uh i i don't know if it's the troublesome trucks episode or one before that but james derails and thomas brings two cranes to get him back on the tracks and they're non-living cranes right but we know that there are several living cranes on Sodor uh like uh I think is it Cranky Cranky is the one on the docks that they that they talk to a lot yeah anyway so he brings two non-living cranes in that episode is he like talking to them no he doesn't talk to them at all he just brings them and they get James back on the track so he doesn't seem to have developed uh empathy at that point but in a later season he does seem to have developed empathy for other machines and maybe he's just developed it not quite exactly how they expected because he's developed it for all

machines not just other sentient ones well so you think them caring about each other is by design you think that's not just a random side effect yes i do think so because every single lesson they learn on Sodor is designed to make them like more caring less selfish uh less cheeky etc the show centers around teaching cooperation and like teamwork uh and being like the main thing that all trains want to be is really useful yeah we're definitely gonna get into that they want to be a really useful engine and i think part of being really useful is being able to cooperate effectively yeah not necessarily empathize though yeah but effective cooperation i think you need at least some ability to empathize with your your you need to be able to think like them and understand where they're coming from in order to cooperate with them i also just want to say uh Just just talking a little bit more about the world. Sir Topham Hatt has a lot of engines, right? I think he started out as a railway guy. And he ended up either funding, probably he's the guy funding it, this AI development program. I think that's the reason there are so many engines that are sentient. But other types of machines, there are just a couple, like there are a few buses that are sentient, a few cranes, a few cars, but then there are a lot of non-sentient cars and things like that.

And I think it's just because the guy who funded this was a railway guy. And I also think that this is like an alternate sci-fi reality where it seems to be set in the past where they don't have cell phones. And I think it's a world where self-driving vehicles and AI were prioritized over things like internet and mobile phones. Well, I mean, it's like it's stuck in the time when the show started, like in mid-80s, right? Yeah. So we know that the island of Sodor, at least, is fictional. So this is an alternate universe. It's just off the coast of the Isle of Man. It almost connects with England. I think there's like a bridge between it and England. A bridge that is a small enough channel that Thomas can jump it. Yeah, he can jump it. You got it. So if these trains are running on some sort of AI system, this seems to be like a very high level AI that we're working with, as opposed to like low level would be like very close to what machines understand.

If we're talking about like a programming language, like assembly is low level, whereas like JavaScript is high level. These trains are very high level. Like what they work on is like conversational level. It's like Star Trek computers almost kind of thing. Right. It's even more high level than Star Trek computers. You can have a conversation with these trains. It's very far abstracted from the machine data and very easily understandable by humans. So I'm wondering who made this system? Because Topham Hatt definitely didn't. It seems like somebody designed this system specifically for him, though. He seems like a guy who knows nothing about technology and someone, if we're believing your theory, made a system just for him. Like a system for someone with no understanding of computers or programming. Well, the one thing we really know about Sir Topham Hatt is that he's freaking loaded, right?

And he's also fat. That's the other thing we know about him. Yeah, he's the fat controller. He didn't used to have a name. Anyway, so yeah, I think that he funded this and I think, like I said, he funded it to go be an AI specifically for rails for trains because that's where he got his start. He's a rail man, you know. Okay, so he is a rail man with lots of money who became an entrepreneur, like a Silicon Valley tech entrepreneur. Like he decided to fund the next gigantic leaf in artificial intelligence. Right. And I wouldn't, of course, there's no evidence to support this, but I would say that it's not too much of a leap to say that Sir Topham Hatt, if he found some genius guy who was coming up with this amazing thing, would be all for just exploiting that. Right. So like, I think there could. Yeah, I mean, if it gives him trains that work for him, because he seems to like that. Yeah. No, we're definitely going to talk about why this is so beneficial for him. But so wait,

just to like, I just want to make sure that people have got this theory down. So like Sir Topham Hatch, just so we're clear, is a CEO of a technology company that develops new kinds of machine learning AI. When they first started, the AI had to be taught how to drive themselves by drivers that were in them. And now they can. They're childlike. They don't really know what they are and they don't question it right like they know we're trained they accept it probably all trains are like this uh they simply learn and this is why each episode is a ham-fisted lesson the end goal of this is being a fully autonomous railway with machines that can work together and pretty much solve all their problems which is why they are introduced to so many problems in each episode you're framing this as topham hat he is the owner of whatever system this is like he made it or at least he didn't do it but he funded it and it belongs to him right he's elon musk it's

not a service that some other company offers that he has leased right like this is something that he owns right this is like a service that he offers okay cool so where are all the software engineers the programmers the data engineers okay i mean so are you saying that the train engineers are actually software engineers we can get we can get into this so let's let's just talk about the island of Sodor it's a little bit strange um yeah there are a lot of like seemingly wealthy middle class people who live there and are always riding the trains like why this is a tiny 300 square miles island and like why do they have such a booming economy right there seems to be a lot of raw materials industry in Sodor. Like there's a quarry, they cut logs. Like there's a lot of raw material harvesting, which doesn't really fit with your standard tech company. But again, this is a 300 square mile island. Like how much raw material could there possibly be? And they've been doing

this for like how long now? Since the 80s? So my theory about Sodor and why it's so strange is that I think all the seemingly wealthy middle-class people who live there are actually researchers or people who support those researchers. So they don't do things very efficiently on Sodor, if you've noticed. Like these trains only haul like five cars or freight cars or trucks or whatever. And I think it's because that's not Sir Topham Hatt's primary business. The primary business is the research and development churned out from allowing these trains to learn and do their jobs. And this is then sold to other companies across the world. I think all of the entire island of Sodor is a mini training simulation. The whole island is Skunk Works, is what you're saying. It's an R&D island. That's all. That's all they do. Yeah, and I think that a lot of the things they have set up, the quarry, the docks.

How can this company be profitable? R&D costs tons of money. It's a huge money sink, if that's all he does. But he sells these trains, right? Yeah, well, not until they're ready. I mean, apparently they only just got ready because they only just expanded into international markets. How much would it be worth to have a massive railway system that is completely autonomous, That you don't need, like, you barely need any human interaction at all, right? I mean, just to refute my own point, point in case, like, we've got Elon Musk who is working on the technology for self-driving cars and also, you know, making, like, railway kind of things. And his companies have not been profitable for a very long time, like Tesla, and yet somehow have ridiculously high booming stocks and value and stuff. So it is possible for an R&D-based company that sells stuff but not so much stuff and is mostly based on the actual research and developing an innovative product can be still, like, if not profitable, it can still exist.

It can subsist and it can even grow. Like I was saying, Sir Topham Hatt, I think, was a railway man to begin with. So I think he was a mogul of a massive railway corporation. So it's not like he's going to be living on the streets if this fails. Right, exactly. And it's been going on for a long time now. And so I do think, like, even though I say Sodor is a mini training simulation, like, they do have trains that come in from off island. I do think that his railway system is larger and is making money in other ways. I just don't think that Sodor itself has, I don't think it could possibly, being the size it is, have such a booming economy. Well, I just had another question about this. So you say that this universe, there are no cell phones or anything like that. Right, it seems to be, yes. So how are these operators making these trains work? Like, do they have personal computers? Is every computer like the size of a house or a train?

How can that work? Like, if we're in a world where technology hasn't even advanced to the point of cell phones, like, we don't even have a pocket-sized device with an antenna. How can they have computers with enough processing power to do the massively complicated machine learning algorithms to make these trains run and learn? What's going on here? Is it all done on punch cards? I think that computers in this universe are massively different from what we know. And I believe that these are bio-organic engines. So like, did he discover them or did he build them? I'm not, I think, look, I think they are built because they have to be sort of, I feel like they are programmed in some way and developed in some way, but I feel like the machine part itself may even be biological and growing within a metal shell, much like that image you can find on the internet or like, uh, uh, that monster trucks movie.

Okay, so you think there is a slug millipede computer monster living inside each train and we only see its face. Yeah, I don't know if that's exactly what it is. I feel like it works like a computer, but it's a bio-organic computer. So they need fuel like coal and fire and water, but I actually think that it works for them more like calories. So they do need fire, coal, and water to run, but it's more about giving them energy like humans. so without it they can still run for a while but will eventually slow and stop so as we saw in the in the donald and douglas episode love me tender like douglas keeps going for a while even without his tender but i also think this this may be how like the troublesome trucks affect small changes in their brakes uh maybe they absorb power generated by just their wheels turning even though they have no source of fuel okay do you think that these uh these machines can continue

to quote-unquote live without sustenance? Like if they're not fed with fuel, will they die? In the movie Sodor's Legend of the Lost Treasure, they imply that Thomas's boiler won't get hot and his fire won't start. And consequently, he's drowsy and doesn't want to wake up. And it's because he did too much racing the day before, implying that the bio-organic ability to store some energy burned from coal and steam, like it can be depleted so the next day you're harder to warm up right like like i'm just saying you look at the evidence in this series like clearly they don't run the same way that no they don't work like trains do in our world they have to be bio-organic in some capacity um i feel like so i believe that on sodor there is a mini mini training simulation so in the episode useful railway they have a miniature railway with many engines that tourists can play on and they act like these engines act like they can be really useful and bring wool to farmers but they are

clearly given busy work not useful right yeah right like they have real engines that could do this work they don't need little engines to do it i think busy that they are just small-scale models to see if they're really useful algorithms are working display models yeah they're showing off to customers in the so in the episode tit for tat when the photographers arrived to take photos of one of the mini engines he says uh they're going to write about me in a book and show me off in moving pictures and on the telly too so i think these mini engines are marketing prototypes they show that they're useful to tourists who are actually or investors and this proves that it's kind of a mini Sodor like the mini railway is mini Sodor and also gives precedence for the fact that The real Sodor could also be fake too. The other Endrons kind of patronize these little ones, not knowing that their world is just as fake and useless as the little mini-railway.

Dude, you just went another layer deep into your own theory. The mini-railway is clearly just to entertain kids and attract investors, yet these engines think they are being really useful in fulfilling their primary program, just like the bigger Thomas and Friends. That's what they want to be, is a really useful engine. Exactly. Well, okay. So along with the tit for tat episode. So you say, sorry, the small train, what's his name? I forget this many engines trains, mostly because they just, their inclusion annoyed me so much. I was just like, why? You really need to sell more toys that much? There's so many engines on Sodor. 300 square miles. It would be literally completely filled with engines at this point. Yeah, right. That's because that's the business they do. But no, so that engine, he says, oh, they're going to take pictures of me and I'm going to be in moving pictures and magazines.

And they say, oh, you'll probably never be in a TV show. And then the narrator says, but he was, wasn't he? And then the train kind of winks at the TV screen and you're like, oh, he is in this show right now. So my question is, is the show itself on some kind of meta level of your theory, is this show actually like does it exist in the thomas universe because that seemed to imply that it does right i mean like that implied that the show we're watching exists in their universe in that case is it like training material right so topham hat can train new trains without actually putting them into commission i mean like he just plays those videos for them and they learn vicariously maybe maybe i think you're right i think on some level sir topham hat would be stupid not to be filming this and sending it out uh to other people as either training videos for their engines that he sold to them or as just like marketing um even though it's not very good

marketing because it shows his trains constantly screwing up and making bad decisions yeah that's so it's training material yeah and so well just speaking of training uh just one point that i had was uh i think that seeing other more experienced trains um is a good way to train the new ai trains and this is why there's always an influx of engines onto their small island i think that they learn best by working with other trains who already know how to drive themselves um and i think this also explains why thomas is sent in the later seasons i don't know if you got this far but thomas in the later seasons he's sent to like he's sent to so many other countries and i think it it's expensive to send your uh trains to sodor to be trained so i think sir topham hat loans out his trains i'm sorry the word train is getting used a lot here yeah we can call them engines yeah sure he's sir topham hat loans out his engines to other railways around the world and i think just being

around thomas helps these these uh train these engines from china and brazil and wherever okay become more counterpoint it seemed like the one doing the most learning was thomas well look i mean they're they're always going to be learning no matter what wherever they go that's how they're designed but i think that that uh clearly the railways he goes to are already functioning so they've been trained in some capacity so they do know some things i think he's just sending out thomas so you're saying we're seeing it from thomas's point of view yeah from his point of view he's always learning but really he was teaching yeah or something like that i don't know well this this is slightly off that topic but just as long as we're talking about the mysteries of sodor um one thing that i was very confused about in the sad story of henry they choose to wall him up yes why why wall him up why not decommission him they talk about

decommissioning trains all the time well because why lock him in they're trying to traumatize him right okay so that that so they expected literally a torture technique they want him to learn a lesson i saw it as like a permanent solution where they're like uh-oh this one's gone rogue we better trap it right that could be no it's true it could be like yeah this is a a.i gone bad we gotta we gotta make sure he doesn't you know wreak havoc corrupt the other trains but i kind of think that like these trains are very dangerous sir top of hat does not seem to care that they are dangerous to humans no right so i really think that that was more about just traumatizing the train and teaching it a hard lesson than protecting anyone because i don't think sir top of hat wants to protect humans okay so that they were just they were that was torture that was punishment for him okay but well speaking of punishment though now we could get into the troublesome trucks if you want

sure let's talk about them so let's talk about the troublesome trucks i have lots of questions about the troublesome trucks and other train cars why why are they sentient what possible use could anyone ever ever have for sentient train cars or are they like a lower level of sentient are they more like cows uh so they seem to have some autonomy to break and push at least a little bit even though they have no drivers and i kind of talked about why i think that is already i think that maybe originally the ai was going to be used for these trucks as well and some of the coaches too but i think that was phased out because there was nowhere to go i think either the ai was too passive like annie and claribel or henrietta who is uh toby's coach uh or they're too aggressive like the troublesome trucks. There are also many other cars and even coaches that don't seem to be sentient at all. I do think, sorry, what I was going with the torture thing

is that I think they kept around the troublesome trucks even after it was kind of a failed experiment because they were an expensive investment, too expensive to just get rid of them. And now they serve to teach the developing AIs new lessons. And they actually, interestingly enough, treat it like a rite of passage dealing with the troublesome trucks. Because when Rebecca later comes to Sodor she has an episode where she has to do the troublesome trucks and everyone warns her Thomas warns her James warns her Percy warns her they all flash back and tell her all the times they had trouble with the troublesome trucks and they're just like you know everyone has problems that is pretty weird it seems like they would still be having trouble with the troublesome trucks if it wasn't some sort of training technique exactly and like like Edward who in the original troublesome trucks episode like he had been pulling them for a while doesn't seem

to have any problem with them at all it seems like it's just like when you've you first sent to do this task that these trucks give you a hard time interestingly enough though rebecca manages to uh get through the troublesome trucks without any problem at all um hey go rebecca just because unlike the other engines who are all selfish and whiny and just try to impose their will on the trucks, uh, Rebecca is just kind of really nice and just laughs along with them and stuff like that. Well, that's nice. Okay. Well, so, so we talked about the troublesome trucks. I'm kind of wondering about Anna, sorry, Annie and Claribel. Yeah. So, so you think they're failed experiments? Like they're just, they're, they failed. Well, I think, I think the troublesome trucks are failed experiments. I'm not sure about Annie and Claribel. Uh, I think that when they were originally created, Annie and Claribel and Henrietta, uh, and other coaches like that were

designed to act as consciences to the engines because annie and claribel are always reminding thomas the right thing to do they're always reminding him his timetable like they're reminding him we need to go now thomas or we'll be late um and then henrietta when she talks to toby is always kind of guiding toby in the right direction and and sort of soothing toby when he gets insulted by the other engines uh so i feel like they are there to uh pacify the trains a little bit and also keep them in line one thing that i noticed in the runaway was that um annie and claribel they seem to think that they're like dating the engine who's pulling them is that so while thomas is pulling them they're like such nice manners it really is a pleasure to go out with you like that's what they say to him so i thought maybe they were also there to like provide company for the trains i don't know and this may just be their misunderstanding as right maybe they

don't know because certainly later on they don't still think that um they kind of have more of a motherly no it's just that once and i was like huh that's weird what a strange thing to say yeah i just watched a couple so i i can't remember which one but i believe it's james is pulling some coaches in one episode and he's really really gung-ho and he wants to go really fast and the coaches are like no you need to slow down we can't go this fast and he ends up shaking them around and he ends up breaking one of them like breaking one of their valves or something um and so he has to stop and then of course all the passengers are angry and so i think that's could be another role for the coaches when they were first created was uh they're designed more to look out no one to slow down well it's a little to slow down but i think also to look out for the passengers right because it would be the passengers who really want them to slow down right uh who are

getting jostled around a lot and so i think that maybe that's one of the reasons they were created but like i said i feel like if that episode is the one that i think you're thinking of then that also contained a really great little uh political commentary on uh shifting the blame because like they stopped and then the the engineers were like well we can fix this but uh we're gonna need a newspaper and a shoelace a leather shoelace and then they're like who has a leather shoelace well this one passenger does so hey passenger give it to us and the passenger's like no i'm not giving you my shoelace and then they're like ringo star is like then the crowd was mad at the conductors then they got mad at the passenger and i'm like whoa they successfully shifted the blame to him so well he did nothing wrong he just wanted to ride the train and now all of a sudden he's the bad guy like geez that's like uh that's like politicians saying you should be mad at the

foreigners taking your jobs but anyway i thought that might be a little instance where they were trying to make some political commentary i don't know if yeah like you said though they were just pretty much reading the children's books at that point so yeah but so wrapping up on the coaches i think that they were sort of a failed maybe not failed experiment but i think at some point they just realized it wasn't really worth it and so there aren't a lot of coaches that have bases there are a few here and there okay and those two coaches were kind of like gifted to thomas right like after he did he does he does something good and then he's given his own branch line i can't remember was the thing good that he did dealing with the troublesome trucks i think it was that would provide credence to your theory if it was because then it's like you passed the test you dealt with the troublesome trucks i think it was like i just watched these

episodes and i feel like it came shortly after the episode where thomas deals with the trucks so one other question um topham hat is training sorry yeah he's training these engines is he also training everything else with a face on so door like the helicopters and the buses and the tractors and like are they all there for training i think to some extent yes though i don't think sir topham hat really cares about the non-engine ais i think maybe other people are training those um and we just don't get to see much of that uh okay we do get to see some of birdie's training and he there's another bus i forget his name we definitely see them learn some lessons so they are getting some sort of training sometimes sir topham hat acts like he's not exactly in charge of things that aren't trains but there are other times where sir topham hat acts like he is 100 in charge of absolutely everything on this island including things

like search and rescue and stuff so yeah i don't know well it's just in the in the runaway um i thought maybe they left thomas's brakes off intentionally as a test for harold the helicopter because yeah the only thing that needed to happen at that point was harold needs to take this helicopter to the next station so that they can you know get thomas and fix him right turn his brakes back on well maybe that was a test for harold that could be but one thing i would say about that is i haven't seen all the episodes of thomas the tank engine there's a lot of them so i don't know but in in no episode i've seen has harold ever been the one who learned a lesson he's always the one who's right well coming to the and that was what i was gonna say is this is different from the lessons you're talking about because this is a purely technical lesson yeah it is not an ethical lesson this is just straight up training for harold like hey harold here's an

emergency how are you going to deal with it so i don't know was your version of the runaway uh was it ringo star i think so i'm not sure george carlin gives harold the weirdest voice he's like oh hey there Thomas I'm here to rescue you like the most stuck up snooty kind of voice you could ever have I don't remember him having a specifically weird voice but that's pretty cool I'm not a hundred percent sure why he even sounds like that because I don't know maybe his personality is a little snooty or maybe it's just the voice but like Harold is nothing but helpful all he does is come to the train's rescue yeah well I mean maybe he's just a stuck up guy Here, one more question about the runaway. So I noticed that Thomas moving under the influence of gravity to the next station because his brakes weren't working, that seemed to tire him out a lot. Yeah, it sure did. Why did that tire him out? He acted like the gravity pushing him to the next station was really, really, well, it was hard work for him.

like is he programmed to feel tired from any movement and if so why why would you want tired drinks well okay maybe so so they know if they're getting overstressed like so they don't let their parts break yeah that could be no that could be it could be some sort of feedback like we have like pain uh so we know not to push ourselves uh or it could be just like uh if they are more bio-organic in nature um like even if you ran down a hill and it's easier than climbing hill you'd still be really tired afterward right um it could just be that the act of moving still does tire them out even if they're not necessarily expending like a bunch of energy to climb the hill they're going down instead i'm not sure but it could be that all right uh as long as i'm doing very small questions uh just one of the questions about the sad story of henry why didn't they have gordon try and push henry out of the tunnel why did they only have thomas try seems like they

would have brought gordon over there gordon's the biggest and strongest he could have definitely No, but Thomas is, that's another thing is that the trains have different roles. And Thomas is a small engine designed for shunting, which is pushing things, right? And Gordon does not like shunting, right? Like he is an express train and he's very important and he can't be shunting other engines around. I don't know. Okay, I understand. No, that's a very good in-universe explanation. Even if we're not going by your theory, that fits. I was going to mention one other thing I noticed. in buckled tracks and bumpy trucks yeah i think that was the name of the episode i thought it was interesting that that episode like the revelation of that episode the lesson learned is not for the trains it's almost for topham hat and it's thomas who actually comes up with the idea to paint the tracks white all by himself right he tells topham hat to do it like and i think that's represent

like kind of a graduation for thomas right i think that's a graduation for everyone i think this is Sir Topham Hatt's goal is for the railway to run itself like they come up with the solutions to their problems so I think that's a good moment for Sir Topham Hatt great that moment you might call it the the singularity when the program is smart enough to teach itself and make itself smarter so from that point on these trains are going to progress at an exponential rate but anyways they should have called the episode the singularity instead of bumpy tracks or whatever the singularity may not be a good thing we can discuss that later all right so yeah what did you want to talk about i was going to talk about the laws of robotics so a little related to the singularity um luckily for sir topham hat even if they did gain singularity and have a you know an engine uprising they're all very loyal and obedient to him so i think he'd be safe even if

there was a machine uprising much like asimov's universe where there are laws of robotics where the robots can't break these laws no matter how much they learn. I think that there is one law of robotic in the Thomas universe and it is to be a really useful engine. And I think this is why no matter how cheeky they are, they will like, for the most part, never cross Sir Topham Hatt because he is the one who determines their usefulness. So this phrase really useful engine, it's used a lot in the classic seasons it gets used even more and more as the seasons go on and i think it's because it's the only way to keep the ai in line sir top and hat will either praise them by saying you were a really useful engine or you were not a useful engine today or something like that and i think you know the latter one you're not useful it just destroys them and the former one that they are useful is like the greatest reward they can receive well and it's interesting like when

rebecca shows up presumably just off of the production line um she's already like having delusions about being a really useful engine she sees the other engine and she's imagining how unbelievably useful they are like she thinks about them flying through the sky as a useful engine team yeah so i mean it's clearly very deeply ingrained in them that they want to be a really useful engine it means something like it is the ultimate achievement for them like nothing could be better it's what they want right and this actually fits in very well to your theory about machine learning now as we get into the technical aspects of this theory i think it's important that we talk a little bit about machine learning and ai in general So I mainly work in cybersecurity. But just a few years ago, I built an AI based massive data set clustering solution, which had to be run on an actual supercomputer. It was the largest, most powerful supercomputer in the whole country.

That was an example of big data processing. Big data is just one application of AI where a data set is just too massive or too diverse to be effectively analyzed by a human. And when I say massive, I mean like mind boggling amounts of data. But a computer is able to quickly process and learn from this data. And that's the kind of thing that I was doing. I was building a clustering solution, which we ran on this supercomputer. I can try and apply your theory to real machine learning principles. And I don't know, I think your theory fits pretty well. It's important to know that AI is actually kind of, I mean, you can think of it as a bunch of different subfields. There's searching and planning, reasoning and knowledge representation, perception, ability to move and manipulate objects, natural language processing, and machine learning. And machine learning is mainly what we're talking about, but actually machine learning kind of bleeds into every other subfield of AI.

And I would say that these trains, their system works with every single subfield of AI. They would be what's called an artificial general intelligence system. And that's basically a hypothetical AI system, which has the capacity to understand or learn any intellectual tasks that a human being can. And this is important because at the moment we can do machine learning and it can learn tasks, but it can pretty much only learn one task. Like if you design a machine learning algorithm and you use it to train a model to identify faces, like face recognition, you can't then take that model and use it to drive a car. Artificial general intelligence system should be able to learn to do both things. And I believe that's what these trains are. Now, I can talk a little bit about what is AI. Basically, any time a computer makes a decision, that's artificial intelligence. And of course, we could write like a one-line program just saying, hey, when Mr. Topham

Hat says, go forward, move your wheels. Is that AI? I mean, most people would say no, because that's like the programmer directly told it to do that. So what we can say is AI is basically, it is a situation which the programmer was not involved in the decision process that the system makes. The actual decision that is made by the system is several layers away from the programmer's influence. Okay, now we've talked about AI a little bit. Let's talk about how this locomotive system works. So like I said, it's pulling from a lot of different fields of AI, a lot of different types of machine learning. Of course, these trains must be using some sort of NLP, like natural language processing, to understand languages and communicate with each other and with Mr. Topham Hatton stuff, probably uses like a neural network or some sort of deep learning algorithm. It's probably based on supervised or self-supervised

learning. Basically supervised learning is when you have inputs and outputs and you train the model on the inputs and outputs. So like, you know what you're getting in and you know what you should be getting out and you train a model based on all of that. And once it's trained enough, you should be able to give it new inputs that you don't know the outputs and then it gives you the outputs right okay you have unsupervised learning which that sounds like what they that does sound like what they were doing to begin with kind of not not quite you have unsupervised learning which is when you have only inputs no outputs you don't know what the output is and you're trying to basically identify patterns or relationships in the inputs so you You just send the program at it and it attempts to fit a model to those inputs. Now what I think these trains are doing would be, I believe it's reinforcement learning.

This is when an agent operates in an environment and must learn to operate using feedback. So the agent is not told which actions to take, instead it needs to discover which actions yield the most reward by trying different actions. Okay. So in this situation, it has the AI agent will be given a goal and it has rules that it works in and it can try whatever it wants to achieve that goal. And it gets feedback whenever it tries something and it uses that feedback to adjust its performance to try and reach the goal better. I believe that this is what you're describing that these trains are doing like mr topham hat he gives them a goal he tells them what he wants them to do one thing i didn't point out from earlier is that sir topham hat always gives the engine the mission or the goal or the job he never talks to the driver right right that's true that's a good point and also he doesn't tell them exactly how to do the job like he doesn't say

Like move your wheels, pick up this, do this. Like he says, I expect you to get this done. And then it's up to them to figure out how to get it done. And that's another thing that you realize as you watch more and more of the show is that unlike real railways where everything is sort of planned and timed out and all the routes are known, these engines choose their own routes and they kind of do what they want. Like they can stop a job, like in that Donald and Douglas episode, Love Me Tender. Those two twin engines are arguing about which job to do first. They are in charge of creating their own schedule, at least in some capacity, their own routes and things like that. Right. And what you're describing is actually how current self-driving car machine learning algorithms work. Reinforcement learning is used in self-driving cars in situations where they have positive or negative signals while they're

driving, like traffic lights, blinking signals from other cars, street signs. The machine learning model in the car, it uses those signals and trains a reinforcement model to decide on the best action to do in that situation, like stop or steer, right? And basically it's trying to get the maximal reward. And this is interesting. When you're dealing with a reinforcement model, you have positive and negative rewards. And that's exactly what Topham Hatt's doing with the idea of being a really useful engine. And he has clearly specific ways that you become a really useful engine. And whenever they do one of those things, they feel good about themselves, right? Like carrying passengers to where they were supposed to go, you're a really useful engine. But then when they do things that wasn't how he wanted them to do it, they're not a really useful engine, like leaving your passengers in the quarry you're not a really useful engine right do that this is

interesting because there's actually i've seen a video on this there was a reinforcement machine learning ai that played tetris someone developed an ai to play tetris and it was going really well it was doing so well that it got like to a point where humans could no longer play it and it even got to a point where no one could play it. Like the algorithm could not play Tetris anymore because the pieces would instantly drop. And this machine learning algorithm had positive feedback in that like, however many seconds it didn't lose, that was good, right? That was its positive feedback. And if it ever lost, that was negative feedback. So this AI figured out by itself that the most efficient way to achieve its goal of seconds without losing was to enter the pause screen and stay there forever because then it was never losing. Right? Yeah. So that's something you have to be careful with with these AIs. You have to be very careful what you actually define as

positive feedback. That sounds like the sort of thing that one of these engines might think of in an episode. It's so self-assured, oh, I found a way. And then all of a sudden Topham Hatt shows up and is angry at them and says, you're not really useful. And then they feel bad. Right. Well, it's just like, um, when Thomas was, I forget which episode it was. It's a relatively new episode. Um, I think season 16, uh, Thomas was given the task, uh, to carry some VIPs around Sodor. They were going on a tour of Sodor and then they were going to come to a really important party afterwards. And this to me sounds like they were business prospects looking to buy some of these trains right um probably so top him hats like all right i'll show you off i'll show yeah train algorithm take them on a ride around my my island and then bring them to the party so i can sell stuff to them so thomas is taking them around the island and he's having a tough time

because you know gordon's the express train he's not used to pulling like two train cars full of people so he gets the idea he's like well this is hard but i'm supposed to show up to this party at this time how can I get there like how can I achieve this goal I know I'll let all these VIPs out in the quarry and then I'll go to the party so he did and he brought his coaches right uh yes just the coaches no people in them exactly so he did what he was supposed to do he did and so he shows up at the party super happy and excited thinking he's a very good engine um and then Topham Hatt gets really mad. He's like, why did you leave all my VIPs in the quarry? And I love the narrator even talks about it. He's like, the people didn't know why Thomas was letting them out at the quarry. They were very confused. I'm like, I'm sure they would be if they were planning to buy this AI system. Like, what's Topham Hatt trying to show off to us right

now? I'll bet they were also a little afraid that this AI started making decisions by itself about where to leave them on them so anyway top mac gets real mad at thomas right then he's not a very useful engine for leaving all his vips in the quarry but i mean this is a situation that really really fits a reinforcement learning model where thomas he had been given the reinforcement the the positive reinforcement that if you take these vips on a tour and get to the party on time you're a really good engine so he did both of those things right but top and hat should have defined it more clearly as like get to the party on time with the passengers yeah so that's something you got to be careful about these with these ais i've also read a paper where they were building a ai that was supposed to like it was it was virtualized so it was in like a 3d program like unity but it was supposed to build a tower and basically the there were several agents like

several pretend they're robots in this 3d map and they were supposed to build a tower by taking blocks and putting them down and they figured out that every time they put down a block they got a reward right so they would take a block put it down pick it up put it down pick it up put it down so like if they were doing what they thought they should be they felt great like they were getting all their awards from that right that's like harvey and so and sabrina carrying a box up and down the stairs instead of moving yeah a full box up and down the stairs so no i think um this uh very very closely fits into how you're describing the engines as an ai system need to be trained if we're going off this theory then yeah basically what mr topham hat is trying to do. Sir Topham Hatt. Yeah, sorry, Sir Topham Hatt. Now he's trying to build up this reinforcement training model so that they can continue to work after he sends them to clients. But also he's

troubleshooting. Like he's trying to make sure that he defines all the variables correctly to make sure the trains actually get the job done. So going along with my theory, I think that first system you mentioned, what was it? Supervised where you give them the answer and then you don't give them the answer. Supervise. They get the input and the answer. Yeah. I think that's kind of how they were working in the original season, at least to some extent when the drivers were driving them. You're exactly right. And I stopped short right there. I actually meant to talk about that. In the first season, what you're describing is a situation in which the trains actually aren't in control over the inputs or the outputs. They're receiving both. So they are doing supervised learning they're building a model and they're fitting it to the inputs and outputs at that point and then they move on to reinforcement learning where they have a model and they're

trying to fit um sorry they're trying to fit the input they receive into that model but they can still adjust it based on feedback so i i mean i think it fits really well and of course i just i gotta say this is all very generalized like in real machine learning you would never use just one model for such a complex system like in reality these trains would be using every model of machine learning all the time right like for in various degrees for various tasks right like nlp the natural language processing by itself there's a whole field dedicated to just that just understanding languages so i mean i'm just saying these these trains are incredibly complicated incredibly sophisticated computers well and also just because you mentioned rebecca i i do think that by the time rebecca rebecca rolls around in like season 23 they've come up with a way to teach the trains the basic things like how to stop and go without actually having them learn

on the island like they did back in the first seasons well that's the cool thing about machine learning is once you have built a model through supervised learning right you can take that model and put it in other, like you can use it as many times as you want. So they would just take that model and put it into new trains. And that was another reason why I felt like Sir Topham Hatt doesn't seem to sell the trains. Like he keeps these trains on the island. He just keeps getting more and more. I think it's because he sells the models, the things that he's developing. And then they are then put into trains like Rebecca or the ones in other countries who then know how to do all those things to begin with. And they can, of course, still keep learning, but they're better right out of the box. Well, and that's an interesting point you make. In a lot of situations, you actually wouldn't necessarily want the system to keep learning once it's in the hands of the client.

Like you would want to get it to a point where it performs correctly and then stop it from learning and just send that model out to the client. So it's no longer correcting itself. Now, maybe that's why the trains in Brazil and China and Africa seemed to know better than Thomas already and didn't seem to be the ones learning lessons. So you're saying, well, then Thomas might have been a case of overfitting. That's when you train it too much, right? And it starts giving you bad data because it's just giving you copies of the inputs that it learned, right? Which you don't want it to just copy the input that it learned. You want it to make decisions based on the input that it learned. So maybe they needed to retrain Thomas. I don't know. Yeah, I don't know. So another thing I wanted to ask you about, do you think this universe has any sort of wireless technology at all? Because why are these trains talking to each other

instead of just transmitting to each other? Right. Talking is unwieldy and inefficient. In that movie, The Great Race, Thomas jumps the bridge and then he's on the wrong track. And he's just lucky that the signal man switches all of the tracks so that he like as thomas getting through that obstacle course of other trains coming at him has nothing to do with thomas yeah and he doesn't do it in that scene but they do do it later in that same movie is what they're supposed to do is yell points and then the signal man hears them and will switch the uh the track to the other track so he should have yelled points and so that's another something maybe to say that they don't seem to have much wireless communication if they actually have to yell to the signal man right now i think this is a universe where there is no wireless technology but one interesting thing is there's another um algorithm called ensemble learning

which is an approach where you take two or more models you fit them on the same data yeah but they they end up slightly different like you run the algorithm multiple times on the same data and then you take these models and you combine the predictions from these models to create hopefully a better performing model based on the ensemble of them so that could be why sir topham hat has such an ensemble of engines on his island right that's what i was gonna say is that might be why he keeps adding engines but then if there's no wireless technology i don't know how they're extracting the model from these trains but they would have to i guess to uh to send it out to other people i just i've never seen them hook up wires to the trains and take things out of their heads maybe they just don't show it because these are training videos so they know they don't need to know they have engineer like they talk about engineers and stuff being in them

fixing things and doing stuff but you never see that so they could be doing anything in there who knows what so one thing i wanted to know about this we're getting now to the point that we that sir topham hat has achieved his fully autonomous railway more or less and it seems the only humans that these engines really talk to are authority figures like sir topham hat and sometimes the mayor and other people or like people who come to the island who are important probably investors and like you said which i thought was a good point was that they they tend to regard most other humans who are not authority figures as extensions of themselves their uh their drivers and mechanics etc and i think this is by design i think they are supposed to be mostly self-regulating only needing occasional uh human maintenance and then humans uh the only humans that they respond to are the ones who are in charge because you wouldn't want your powerful engines to take

orders from just anyone in the episode all in vain james wants repainted he goes to the steamworks where they get fixed up and painted and stuff he talks to the engine in charge named victor and even though the human painter is standing right next to victor and later we see the painter standing there covered in paint having just finished the job and james thanks victor who did nothing and leaves right like victor is in charge of the steamworks right and it's almost like the humans are the programs here right they do what the trains say right and like the engines are actually in charge of the humans at least low-ranking ones and i think this is exactly what sir topham hat wants i think that the ones who are loyal to him by programming are in charge of the ones he can't control like the peons the the the small humans who could rise up against you kind of thing so do you think sir topham hat also constructs artificial humans no no i'm not saying

that okay i'm saying no i was just clarifying i didn't think so i just clarifying sir topham hat i don't think he really likes humans i think he likes trains i think he wants to surround himself with more and more trains well i mean even even looking past that what he wants is a self-governing system yeah that reports to him exactly only him and only other managers right like he wants a system that can fix itself that can govern itself that can take care of problems by itself and if humans need to be involved can inform those humans what to do and then the humans will do it right like he doesn't want humans to be telling his system what to do exactly which also makes sense like he's he's going for autonomy and that's a huge thing in business right now everyone wants autonomy so i can see why he's doing that yeah it saves a lot of money because then you can cut people out of the picture and hey what do you know you're not paying salaries anymore yep

just a few other red of things in the episode buckled tracks and bumpy cars i just think because we had been talking about self-driving cars earlier the engines stop for ducks so they are very advanced self-driving engines they can stop for even little obstacles like ducks crossing the tracks they have no need for humans to be looking out for anything on the tracks for them and uh in that same episode thomas just brings rocky to put the the fire engine i forget her name but he brings rocky who's the crane and this is a crane that is pulled on a car uh because there's another engine that is also a crane i forget his name anyway so he brings rocky to put the fire engine back on the tracks now this is something that thomas decided to do on his own he saw a problem right she's off the rails he went to get rocky he didn't tell anyone he didn't radio or whatever he just went got rocky and rocky they don't have radios yeah exactly so like that's what i was saying before

like like they make their own decisions they choose their own routes and if they see a problem on the way to their their main goal they can decide to fix it instead um they can make stops and they can go to other places and they can check on other engines like rebecca in one episode is uh just doing her job and percy is worried that thomas was in an avalanche or something which turns out thomas wasn't but percy going by in a panic and telling rebecca if we have to find thomas thomas is in trouble makes rebecca drop what she's doing and she starts running around in a panic all over the island telling everyone else we have to find thomas he's in trouble so eventually like the whole island is looking for thomas and they find thomas and he's fine but percy ended up in an avalanche while looking for thomas so now they have to find personal no but anyway so like that's this is an interesting thing that like they can

all decide to drop their jobs to go save another engine yeah well i mean i think there's actually really good hard evidence that this is what topham hat wants because like you said they didn't do that at all in the first season the first season they didn't do anything except follow orders right like like they they only care about themselves well and it was sitting back passively as humans followed orders inside of them right and then like you say as the seasons go along they get more and more autonomy until in the later seasons they're making decisions by themselves and helping each other by themselves and you know accomplishing teamwork by themselves so i mean it fits like it makes sense he's autonomizing slowly his entire island oh so another just uh evidence of humans not being necessary in sodor's legend of the lost treasure thomas finds out that there's one uh i don't even know what they're called those tall towers that fill him up with coal he finds out

that there's one that has bad coal and his firemen have to clean out his entire coal tank or whatever it's called and get him new good coal so that he stops like spitting and sputtering and feeling sick the next day this engine he doesn't like named ryan is getting coal from that same thing and thomas decides not to say anything because thomas doesn't like ryan because ryan took thomas's ranch line but like presumably the firemen know right but it's not their job like they are they only do very small things and they're supposed to let the engines run their own rail and so they don't warn ryan either or ryan's fireman same thing when uh thomas was mad at i forget who it was was it henry because henry was supposed to carry some flour for like a kid's bake sale or something so he like Thomas was supposed to taxi the car full of flour to him but instead he decided to hide it and he accidentally bumped it so hard that it split open and spilled flour everywhere

as it was driving so we got Henry who's just chasing Thomas around the island following this trail of flour being like Thomas stop you know give me the flour you split it open you're wasting it and Thomas is like he he he I'm gonna hide it this whole time you've got presumably conductors that know what's going on but they're not doing anything they're just like all right my train wants to run away with the flower I guess I'll let it happen so while having sentient and emotional engines does create problems as we see as long as Sir Topham Hatt trains them well on his island he can sell them and eliminate the need to pay human workers very much or listen to their ideas or arguments and eventually he has an army of highly trained AIs who are loyal only to him I don't know if this is what his plan is, but I'm just saying that Sir Topham Hatt could take over the world probably at this point. All right. Well, with that in mind, with the idea of taking over the world,

I want to introduce something called the paperclip maximizer theory. This is specifically a theory about AIs, about machine learning AIs. If you remember, I described an artificial general intelligence, which is basically an AI that can learn to do anything a human can. The idea behind the paperclip maximizer theory is that once you have an artificial general intelligence system which has been designed with a singular goal, in this case let's pretend it has a goal of maximizing a paperclip collection. So all it wants to do is collect paperclips, that's it. But it has the intelligence of a human. This completely benign system, like there's nothing inherently good or bad about collecting paperclips. It's just a thing to do. this benign system starting at human level intelligence might be, you know, it might start out by collecting paper clips or buying paper clips or even manufacturing paper clips. But

eventually it's going to realize the only way to get more paper clips or the best way to get more paper clips is to improve and optimize its own intelligence. And it would then start improving its own intelligence only because this leads to more paper clips, right? It's not doing it out of any sort of desire or, like, ambition or hatred or love. No, it has a goal, which is to get paperclips, and it knows that intelligence will help it to do that. The idea behind this theory is that once it gets intelligent enough, it may discover a way to maximize its paperclip collection, which inadvertently destroys the Earth and all life as we know it. Like, for example, it might Start converting the entire planet and every other planet into paperclip manufacturers. So that's the paperclip maximizer theory. Now, do you think that this theory applies to Thomas and friends? Because like we said in the episode Buckled Tracks and Bumpy Trucks, Thomas comes up with the idea to paint the tracks white by himself, right?

Like he understands now that him gaining intelligence and learning new things about the tracks will help him to be a more useful engine. So they've already hit that, like we talked about, that singularity point. Do you think that someday they may realize that the best way to be a useful engine would be if there's no humans around critiquing them and then get rid of all the humans, including Topham Hatt? So they're really useful engines? Yeah. I mean, possibly. The one issue they have right now is that Sir Topham Hatt, and presumably them by extension, have not come up with the idea to create any new types of machines. They just install the AIs in existing types of machines. That's true, yeah. So as they now function, they do need humans to do things that they can't do because they have no arms. You know, so like right now they need humans, but they treat them as like little worker drones, you know.

But I don't know, maybe at some point they could advance to a point where they realize they need other inventions and then they don't need humans anymore. And then, yeah, they might get rid of humans and, you know, watch out Sir Topham Hatt. Can't you just imagine a situation, though, where Thomas, he hasn't done something right and he's like, oh, no, now I have to report back to Sir Topham Hatt. And then he realizes, wait a minute, if I were to ship Sir Topham Hatt out to the middle of the ocean, I would be a really useful engine. Because then Topham Hatt could never tell me I'm not useful. I can imagine him doing that. I mean, they have no empathy, right? I can totally imagine that happening. But I think if Thomas did that, one of the other engines would think that that wasn't the right way of doing it and would somehow manage to get Sir Topham Hatt back. Unless Thomas talks to the other engines and says, you know how we could all be really useful engines if we shipped Sir Topham Hatt out to the middle of the ocean.

Interestingly enough, though, when they come up with these little harebrained schemes of theirs, they usually don't talk to the other engines about it. They usually just do it because I assume it's so brilliant. I mean, but they are working together better and better, as we've talked about. Right, that's what he wants, though. He's teaching them to cooperate. So, yeah, no, they could reach that point. So maybe they'll take out Sir Topham Hatt before he takes over the world, but then they'll take over the world instead. Yeah, but they'll take over the world just purely with the goal of being really useful engines. That's all they want to do. And there's no point to it. Like to a machine learning algorithm, all it wants to do is to get positive feedback. And in their case, that's positive feedback. So that's their life's purpose, right? But when they ship Sir Topham Hatt out to sea, they're not getting positive feedback.

They're just not not getting negative feedback, right? Right. But in that case, that could be enough. Like I said, avoiding negative feedback was enough to convince the Tetris playing AI from my story to never finish a game, right? Like it learned that pausing the game was the best option, right? And that wasn't because of positive feedback. That was to avoid negative feedback. I mean, just saying, they might realize that avoiding negative feedback has its value. Yeah. Well, I have gone over pretty much all of my points. The last thing I have to say has nothing to do with my theory, but just that their swear word in this universe is cinders and ashes, which I think is just kind of funny because that's that's their excrement i kind of wonder if they should bleep that i mean they're most of their excrement should be steam right they're steam engines but i mean they also do produce cinders and ashes but they run with steam cinders and ashes is something you don't

really want or need you know so you just clean i guess to them steam would be like urine and cinders and ashes would be solid waste right so that makes sense that's right so like when And if they get mad, they're like, this really gets me steamed, right? This pisses me off. Oh, you got it. That's pretty cool. That's some really good continuity there. I like that. I don't know if they ever say that, though. Well, I like this theory a lot. I think it's great. From a computer science perspective, I really love it. So what I like most about this theory is that it takes something that is nice but doesn't really make sense because it's a kid's show. And it makes it make sense, right? Like, this fills in all the gaps for that show. Like, without this theory, what is Sodor? Like, why are there so many sentient trains on Sodor? What's going on? Like, I get it. It's a fun kids TV show, but it makes no sense.

With this theory, though, it makes perfect sense. I know exactly what Sodor is. It's a giant R&D facility for sentient trains using machine learning algorithms. Like, it's great. I love it. I give this theory two thumbs up. It's, I think it's a really cool theory. I think it improves the TV show and it doesn't detract from the TV show in any way. Like it's the net positive for the show. I don't know if Reverend Audrey would agree though. No, probably not. Not at all. But he was a purist. He's a purist and we're theorists. As much of a fan of Thomas and Friends as I am, and I love the show, neither one of us has seen all the episodes, right? So there could be things, you know, in one random episode here or there that kind of maybe might disprove our theory. Who knows? Yeah, but the show contradicts itself sometimes. It's true. I mean, like you said, how they couldn't control themselves at first

and they can later. The point of this theory was to try to get rid of most of those contradictions. And it does. I mean, even if there are a few contradictions, like I said, it improves the show. It's a net positive. So I like it. I think it's great. All right, great. Well, this has been The Popcorn Isn't Real. I am Leif Eric, and I'm here with Torvald, and we are signing off. See you later. Thank you so much for listening. Opening music for this episode was provided by Christine. If you want to see how our theory holds up, we provide links in the episode description where you can enjoy the media we discuss. And if you like our podcast, don't forget to rate, review, subscribe on whatever platform you get your podcasts. And remember, the popcorn isn't real. *crickets*