How the Internet Became a Glorious Mess | Chaos Lever

[00:00:00.05]
Chris: But no, this is something that's super cool that you would probably actually like, but it's probably too cool for you. So I'm not sure how to handle this dichotomy.

[00:00:11.04]
Ned: No.

[00:00:11.19]
Chris: In any event, there's books that are written about the SCP universe and they're very creepy and end up giving people weird dreams. Ask me how I know.

[00:00:22.00]
Ned: Again, I'm not sure if we're talking about AWS or not. Hello, alleged human, and welcome to the Chaos Lover podcast. My name is Ned and I'm definitely not a robot. I am a real human person who does not restrict what you're able to do within your AWS account. That would be a strange thing for me to do as a human. Feels like something else could take care of that. Some three letter thing with me is Chris, who is also a three letter thing.

[00:01:01.04]
Chris: I usually am referred to as a four letter thing. So I don't know if this is a promotion or a demotion.

[00:01:07.07]
Ned: Think of it as a lateral movement.

[00:01:09.06]
Chris: There we go.

[00:01:10.17]
Ned: I'm so glad we got there. I. I still have no idea what you're talking about and I not sure if I want you to expand on it.

[00:01:20.18]
Chris: It's called the SCP Foundation.

[00:01:23.05]
Ned: Okay.

[00:01:24.09]
Chris: That's all I will give you. If you insist, I will send you a good entry point into the world.

[00:01:32.17]
Ned: I can Google, sorta. I'll ask ChatGPT how to search for. Did you see there was an interview that Ed Zitron gave to Adam Conover and it was two hours of him just ranting about AI, if you need some catharsis.

[00:02:01.20]
Chris: What the.

[00:02:06.02]
Ned: We'll listen to that music. It's very cathartic.

[00:02:08.27]
Chris: Something's ringing. Oh, it's fucking. I know what it is. Hold on.

[00:02:13.25]
Ned: Now you're gonna make me edit and shit.

[00:02:16.21]
Chris: Where were we? You were lost in an ikea.

[00:02:20.11]
Ned: I think, as is tradition.

[00:02:24.26]
Chris: Honestly, is there another way to be in an ikea?

[00:02:28.09]
Ned: I often wonder if you just run straightforward, you know, like a juggernaut style. If you could get out.

[00:02:36.21]
Chris: It'S worth a shot.

[00:02:38.02]
Ned: It might be the most effective way to exit an IKEA ever. Unless you're on the second floor.

[00:02:43.18]
Chris: This might be why they don't sell chainsaws.

[00:02:46.28]
Ned: Don't they? Then where did I get mine?

[00:02:51.25]
Chris: Chainsaws Unlimited, obviously.

[00:02:56.02]
Ned: Oh, maybe Elon Musk gave it to me. That fucking guy. What I was saying is that if you're looking from for something cathartic when it comes to AI, you can listen to Ed rant for almost two hours about It. And he's not always right. He definitely said some things that are factually wrong, but, like, the vibe is right. Like, fuck these guys. And I'm all right. I'm here for it.

[00:03:23.01]
Chris: I know you're a big vibe guy.

[00:03:24.24]
Ned: I am. It's almost like computers were a bad idea from the beginning. Hey, so what's our episode about?

[00:03:32.12]
Chris: It's not about computers.

[00:03:34.02]
Ned: Oh, thank God. That would be awkward.

[00:03:36.14]
Chris: It's about computer networks.

[00:03:38.07]
Ned: Ah, shit. Okay.

[00:03:43.01]
Chris: I can't believe we've never done something like this before, but it ties nicely into an episode you did a bit ago. I was going to look into it and get the actual number for the episode, but I didn't. You did a whole thing about computers. I think it might have been the supercomputers episode. Talking about how we started from nothing and created single purpose computers, which became gigantic general purpose computers, which got us to more or less where we are now.

[00:04:09.17]
Ned: I also think I did one that built logic circuits all the way up from transistors. That one was a lot.

[00:04:16.26]
Chris: Yeah. So I'm actually, I'm reading a book about a topic I'm going to do later on in the year and I'm doing Boolean math logic right now. And I'm really upset.

[00:04:25.16]
Ned: Set with everyone, mostly George Bool. You did this.

[00:04:32.08]
Chris: But no, we're going to talk about the history of the Internet.

[00:04:37.00]
Ned: Yay.

[00:04:39.12]
Chris: And what's funny is I wrote this like three weeks ago. I've got no clue what's in this article. So this is going to be fun for everybody. In the beginning, there were computers and God saw them and said, ah, shit, this is going to be a problem.

[00:04:58.01]
Ned: It is.

[00:04:59.26]
Chris: And the people said, nah, it's fine.

[00:05:02.27]
Ned: They were wrong.

[00:05:04.05]
Chris: This is back in the 1920s, of course, when such conversations happened all the time. Everyone knows that God does enjoy a good gin, Ricky.

[00:05:13.05]
Ned: Who doesn't?

[00:05:14.07]
Chris: You just can't get a good one these days.

[00:05:16.13]
Ned: So true.

[00:05:18.11]
Chris: Anyway, between sips, God said, fine, let's see how it goes then. But don't come complaining to me when pizza places start influencing elections. And the people said, what?

[00:05:33.13]
Ned: What's pizza?

[00:05:35.12]
Chris: Oh, God, what a world to live in.

[00:05:37.16]
Ned: I know, 1920s, I don't think a pizzeria was a thing.

[00:05:42.08]
Chris: Oh, I don't even know if Italy existed in 1920.

[00:05:46.00]
Ned: Seems unlikely. What are we talking about again, as a country? I actually don't know when Italy formed as a country, because it was just a bunch of arguing nation states for a long time, wasn't it? I mean, still is.

[00:05:59.29]
Chris: Right. Anyway, back to computers, which were not invented in Italy.

[00:06:06.12]
Ned: Well, okay, I'll back up. Go ahead.

[00:06:11.24]
Chris: In the beginning, computers we had were big, slow, and very academic in nature. They existed primarily at universities, which is also where the whole idea of data science was coming into form. The government eventually got involved as it became clear that computers were helpful. Computer like machines helped crack codes and won World War II, as we outlined just a few weeks ago in the Enigma episode. Not long after this, ENIAC hit the market. I think that was 46 or 47 businesses started using computers for data science. And at some point in here now we're at the end of the 50s, someone said, wouldn't it be useful if computers could communicate with each other from like a distance? And Bette Midler said, I can write a song about this, which was strange because I don't think she was born yet. Anyway, what happened was we figured out how computers could talk to each other over modems. Now, if anybody remembers the movie War Games, you might remember the old fashioned way of doing that with a modem was literally a cradle that fit the phone. You took it off the hook, you put it on top of the modem.

[00:07:31.23]
Ned: Yep.

[00:07:32.16]
Chris: And this worked because at the time the phone was a standardized item. There was one you could get from Bell Telephone. It was called the Phone.

[00:07:41.27]
Ned: That is wild to remember that there was a time where you got your phone from the phone company and they had like one model.

[00:07:51.05]
Chris: Yep. Same phone, same handle. Most important.

[00:07:55.20]
Ned: So it fit nicely on the two cups.

[00:07:58.04]
Chris: Right. And this allowed the computer to both speak into the talkie part and hear through the ear loudy part. Stop me if I'm blowing you away with the terminology here.

[00:08:12.24]
Ned: Slow down. Okay. The transmitter and the receiver, maybe.

[00:08:17.02]
Chris: There we go. You want to be all fancy about it.

[00:08:20.26]
Ned: I am fancy.

[00:08:22.15]
Chris: So the first one of these modems was invented in 1958 for the sage, which was short for Semi Automatic Ground Environment Air Defense System. And there are fun pictures of this on the Internet. The computer has an ashtray.

[00:08:42.20]
Ned: Okay.

[00:08:43.25]
Chris: But modems, it was communication, which was great, but it was one on one communication. You had a phone. I'm sorry, you had a machine. You would plug it into the modem, punch in the number, hit the button, and assuming you're calling a computer that also has a modem, it would connect. As long as this worked, you had a rudimentary network connection across the phone lines. That last. That ran at a blistering 110 bits per second.

[00:09:16.18]
Ned: Whoa.

[00:09:17.15]
Chris: Screaming down the wires.

[00:09:19.14]
Ned: So much information.

[00:09:21.04]
Chris: My rudimentary math on this puts that at about 13 baud.

[00:09:26.20]
Ned: That's quite bawdy. Out.

[00:09:30.06]
Chris: Out. Now, obviously modems got better over time, but the idea of connecting computers together started to spark people's imaginations. The idea started out in academia. Colleges like the University of California had branches all over the state. Wouldn't it be useful if, say, UCLA had computers that could talk to the computers at UC Santa Cruz and all the other computers in the UC branch? I almost said network. It's not a network yet.

[00:10:02.13]
Ned: Nope. Just point to point communication, right?

[00:10:05.12]
Chris: Exactly. Also, the communication was unique to the machines. So say you had an IBM 360 and your friend had an IBM 360, you know, like you do, those two computers could talk to each other. But if you had a 360 and you wanted to collaborate and your Neighbor had a CDC 6600, which is a precursor to the craze, do you do your homework?

[00:10:31.18]
Ned: We may have done a whole episode on that.

[00:10:34.17]
Chris: They could not natively talk to each other.

[00:10:36.29]
Ned: Course not.

[00:10:37.26]
Chris: Remember, super early, this kind of standardization had not occurred yet. Obviously we're going to get there, so put a pin in that.

[00:10:46.06]
Ned: Okay?

[00:10:47.11]
Chris: The idea here was that the network could span larger scales of geography and connect a multitude of computers. And everyone was like, how do we do that? Now here's part of the story that people probably do know about. In the late 50s and into the early 60s, the government took an interest in creating a network that could survive a nuclear attack. Cold War stuff. Good times were had by all. We were under our desks in an air raid drill. The government couldn't figure this out per se, and asked the RAND corporation what to do. What is the best network to survive that kind of what if scenario? And RAND was already one of those companies that had a ton of defense experience and also computer experience. The team that tackled this question was led by one Paul Baran, a noted scientist and researcher. And I don't remember if I ever say his name again because again, it's been many, many months.

[00:11:53.01]
Ned: All right.

[00:11:55.05]
Chris: After a lot of back and forth, the RAND recommendations came up with two fundamental ideas that still underpin the Internet today. The first is what we will call a distributed network design. We're going to call it that because.

[00:12:09.13]
Ned: That'S what it's called and that's what it is.

[00:12:12.12]
Chris: In order to understand what that actually means, though, we should look at the competing ideas because there were more than one at the time fighting for dominance. 1. And the most simple one to understand is A centralized network model. What that effectively means is you have a home base in the middle or central that has a line out to all the computers you want to connect in the network. So if node A wants to connect to node B, it still has to go through the home base.

[00:12:45.28]
Ned: Right.

[00:12:46.15]
Chris: Even if it's right across the street from the other node.

[00:12:49.05]
Ned: Doesn't matter.

[00:12:50.02]
Chris: Does not matter.

[00:12:51.13]
Ned: That's called a hub and spoke topology.

[00:12:53.28]
Chris: Right. This is a problem that still comes up. This annoying wild maneuver that the traffic has to do is called hairpinning or backhauling, depending on whether you're being nice about it or not. Often not having a home based node is good for one reason, and that is that all the traffic goes to the same place. And that means that if you have traffic analysis or access control or user management questions, it can all be done in one place and everybody has no choice but to abide by those rules or they just won't get a connection.

[00:13:29.18]
Ned: Also, and possibly more importantly to this conversation, you have a single point for routing.

[00:13:35.16]
Chris: Right.

[00:13:36.28]
Ned: You don't have to worry about how the packet or whatever gets from point A to point B, because it's always going to go back to the hub. And the hub is where is what makes the determination of getting it to node B.

[00:13:49.12]
Chris: Exactly. So what you get from a centralized design is a huge amount of control, but you lose a huge amount of flexibility. And also what happens if the home node disappears?

[00:14:01.19]
Ned: Nothing good.

[00:14:02.27]
Chris: Nothing communicates, that's for sure.

[00:14:06.26]
Ned: Yes.

[00:14:08.17]
Chris: So this was the easiest one to envision. And on paper it looks fine, but it's immediately clear that it's just not practical or scalable. Because think about America. If we did this as a central design, where would the center be? Where would the hub go?

[00:14:25.05]
Ned: Nebraska.

[00:14:26.11]
Chris: Right. You literally would probably put it in the middle of the country. Connect San Francisco to New York. This makes perfect sense. Straight shot across the country, connect Boston to New York, all of a sudden that hairpinning thing becomes a bit of a problem.

[00:14:41.19]
Ned: It would be sort of weird to send it all the way out to Nebraska and back. Yeah, right.

[00:14:46.25]
Chris: It doesn't make sense from that perspective. It would kill bandwidth. And again, this is across the street. You could just throw a wire. Boston and New York are like 12 minutes away from each other. Don't look that up.

[00:14:56.29]
Ned: Yeah, sounds right.

[00:14:59.13]
Chris: So looking at that and especially taking into the geographical problem, the next model was called decentralized. And really all it did was create multiple home bases. To be fair, this does solve a couple of problems. We were Talking about. So if in the centralized model, New York traffic goes all the way to Nebraska and then goes all the way back to Boston, and we agree that that's dumb, well, now we can solve that problem by having a much closer hub that is a home base subprime of its own, shall we say? Sure, let's just call that New England and we'll put it in Vermont, because all the cool things are in Vermont. And by that I mean Ben and Jerry's ice cream. So you have the idea of centralized control, centralized analysis, all that stuff we talked about, it's just now much more localized and all of those home bases can report back to the main home base. Right. Homebase is administrative primarily, but it will still pass packets because you still have like, for example, San Francisco to New York that'll go through home base. It'll just pass through two local home nodes on its way.

[00:16:16.16]
Ned: Right.

[00:16:17.18]
Chris: So it makes it functionally a little bit easier. It makes it way more palatable to communicate within your geographical network. But it doesn't solve the main problem that was asked, that is what happens if a bomb hits home base? Well, now New York can talk to Boston. That's great. We're glad. Well, New York's probably not glad, because who wants to talk to Boston?

[00:16:42.10]
Ned: It's true.

[00:16:43.09]
Chris: Yeah, but if you lose that central node, again, you lose connection to the whole country.

[00:16:50.14]
Ned: Right. You still have localized communication, but San Francisco can't, or LA can't talk to New York. Orlando can't talk to anybody. Which is probably for the best.

[00:16:59.24]
Chris: Right?

[00:17:00.20]
Ned: Yeah.

[00:17:01.01]
Chris: I've been saying for years that the best thing we could have done is donate Florida back to Cuba.

[00:17:05.11]
Ned: I mean, it looks like you could just lop it off, like no problem.

[00:17:09.18]
Chris: Get one of those chainsaws from ikea.

[00:17:12.24]
Ned: Exactly.

[00:17:14.24]
Chris: So both of those ideas, even though they do have merits and in certain small scale designs, they are still a point of contention and conversation. It didn't make sense for a countrywide problem. What Rand came up with instead was a fully distributed network. And what that meant, in short, is that wherever you can make a direct connection between nodes, you should make a direct connection between the nodes. Furthermore, the more connections that you have, the better. So now imagine the United States with say, thousand cities. I know is more just hypothetically.

[00:17:52.14]
Ned: Sure.

[00:17:53.07]
Chris: Each local city, for example, New York to Boston, have a direct line to each other. And each one of those cities has a direct line to smaller, more local cities. And each one of those has multiple lines going out in other directions, all intended to pass local Traffic as closely and quickly as possible.

[00:18:09.24]
Ned: Mm.

[00:18:11.18]
Chris: What this does is allow you to pass traffic from New York to San Francisco. But now you have a million jillion billion quadrillion different potential ways to do it.

[00:18:23.24]
Ned: Right. So there's no single point of failure to disconnect all of the country.

[00:18:30.07]
Chris: Right. And there's no concept of centralization at all for node based traffic. All nodes are created equal. Really. The only difference is the distance between the nodes and the amount of bandwidth that you have. That's it. Nobody's in charge.

[00:18:49.05]
Ned: Right. There's no person who's saying all traffic from point A is going to follow this path to point B. The path now has to become dynamic and calculated on the fly, depending on links and bandwidth and a bunch of.

[00:19:06.16]
Chris: Other factors and a bunch of networking gear talking to each other.

[00:19:10.25]
Ned: Yes.

[00:19:11.23]
Chris: Constantly updating. And we're not going to go too deep into this, but I do want to hit on one important point. The way that this works, the only way that it actually works and stays sustainable. Because if you think about it, with no centralized control, node will pop up, will be connected to a different node, node will disappear. Such is life. How do you handle that? The way that you handle that is you instruct all the network equipment to be able to say one simple thing. I don't have the information for node X. The last time I sent it, I sent it to node Y. One path, that's all they care about. When you get to the next node, goes to the next node from that same question and you answer that same question over and over and over again. When it gets to an error, it'll try something else. This is called self healing. Because it screwed up the first time, it won't screw up the second time.

[00:20:04.10]
Ned: Okay.

[00:20:07.11]
Chris: The routing tables will be updated by these devices. And that is the whole point of a fully distributed network like this. Every node has ideally multiple paths to communicate with every other node. And because there are so many connections in a distributed network, the amount of damage that the network can take and still function is astonishing.

[00:20:30.12]
Ned: Yeah.

[00:20:33.08]
Chris: It is likely the only correct answer to the how do we solve for a nuclear bomb going off in America and still being able to communicate?

[00:20:42.12]
Ned: This is one of those things that I found kind of mind blowing when I first started to understand it about the Internet. And it's the fact that routing is, routing across the Internet is non deterministic. There is no centralized authority that is looking at all of the various paths and going, you go this way and you go that way. No, it's just A bunch of different network nodes exchanging routing information. And then magic happens, right?

[00:21:09.13]
Chris: Yeah, that's actually in the design document. Yeah, and then magic happens.

[00:21:14.18]
Ned: RFC 1206. And then magic happens. And I made that up. Don't look it up.

[00:21:23.08]
Chris: So that was one of the biggest things. Aha moments, if you will. And the second thing that was decided upon is so much more boring but just as obvious in retrospect. And we have to remember that it sounds dumb when we talk about it now, but this is all 2020 hindsight stuff. Basically, point 2 from the Rand recommendations. We as an industry, meaning it, have to standardize network communication stack. That's it. The end.

[00:22:01.15]
Ned: Agreed.

[00:22:03.01]
Chris: The biggest thing that RAND came up with was breaking the messages down into blocks that were standardized inside size and were speaking a common language. And that's basically what they went back to the government with distributed self healing network design and standardization around packet sizes and protocols. Hmm, that is not bad for 1962. Starting from zero.

[00:22:26.23]
Ned: No, no, that's. That is actually a ridiculously good start.

[00:22:35.06]
Chris: The advanced. Nope.

[00:22:37.08]
Ned: No, no. Because I had to find out. I looked up RFC 1206 because I wanted to see what it was. And this is amazing because it is answers to commonly asked new Internet user questions.

[00:22:50.00]
Chris: That's phenomenal.

[00:22:51.18]
Ned: Totally. And amazing. All right, keep going with your thing.

[00:22:58.09]
Chris: So we had our recommendations document from Rand in 1962. The Advanced Research Projects Agency, you heard of it?

[00:23:07.13]
Ned: Yep.

[00:23:08.21]
Chris: They created what they called arpanet and connected four colleges together in 1969. The first four nodes on what we can legitimately call the modern Internet, where you see Santa Barbara, Stanford, ucla, and then the University of Utah somehow snuck in there.

[00:23:31.19]
Ned: Bunch of weirdos.

[00:23:34.22]
Chris: Something of interest. The devices that these schools were using were not all the same. Now this is the problem I talked about earlier. There was a PDP 10, two different kinds of IBM computers, and then a Sigma 7, which is a thing that I hadn't even heard of before. It sounds like. I don't know, it's like a. You get in trouble for that in the military. I got to go to the brig, man. I got hit up with a Sigma 7.

[00:24:02.19]
Ned: Oh. I was gonna say it sounds more like an elite death squad. I'm part of Sigma 7. You don't want no one.

[00:24:08.00]
Chris: Either way, it doesn't sound like a supercomputer.

[00:24:11.08]
Ned: That's fair.

[00:24:12.22]
Chris: Anyway, so this was 1969. Even though Rand recommended standardized protocols, at this point they were not a thing. Of course, the difference between theory and reality, what we ended up with was the creation of an intermediary device called an imp. IMP or interface message processor. The imps sat in front of the systems and handled the network communication. An IMP would be fluent in the same communication that all the other imps used. And each IMP was fluent in communicating with the computer it was connected to.

[00:24:49.09]
Ned: Right?

[00:24:50.16]
Chris: So they didn't. The network for this time didn't ask the PDP10 to be able to speak fluently with an IBM940. What they did was set it up so the PDP10 could talk to the IMP. PDP10 talks to IMP. IMP talks to other implementations. That IMP talks to the IBM 940. That's how they connected. And that's how it worked for at least two years. And things escalated from four computers quite fast. This is one of those things. I should have whited it out because I would have been really curious to hear your guesses. When ARPANET started, there were four hosts. Two years later there were 18. TCP and IP. Later happened, modem technology evolved, broadband happened, yada yada yada. By the 1980s there were 100,000 systems online. And eventually we had America Online.

[00:25:49.08]
Ned: Just remember, in the 1980s, the World Wide web was not yet a thing. Web browsers were not yet a thing. So the people who wanted to get online at this time really need to be serious about getting online. This is not a trivial affair where you got a disk in the mail and you plug the phone jack into the back of your computer.

[00:26:11.01]
Chris: Right? You really had to earn it.

[00:26:13.18]
Ned: What's interesting is those imps eventually became the routers of today. Because eventually they designed imps that could interconnect multiple computers at the same site and then also be the connection point from a single site to other sites.

[00:26:32.03]
Chris: Right?

[00:26:33.09]
Ned: So it wasn't that they only did the one to one communication between a PDP 10 and some other server. But they were also would have multiple PDP10s and other devices plugged into the single imp and that would communicate between those local nodes, but also then be the way of transmitting information into what was ARPANET at the time.

[00:26:57.13]
Chris: Right? I'm pretty sure those are the people that founded Cisco. And thus the magic started.

[00:27:03.12]
Ned: I believe you're right. Or at least some of the people involved ended up, you know, founding Cisco and making a ridiculous amount of money.

[00:27:14.23]
Chris: Now obviously I yada yada, a lot of things there because there was certain stuff, you know, if people are interested, we can do a second version of this. Because apparently all we do is multi part episodes anymore. But Things do start to get a little more complicated and a lot more people get involved, including names. People probably have heard of the two principal scientists. The most important people for the ARPANET going from a design theory and a potentiality into a real live environment were Vint Cerf and Bob Kahn. And they are today known as the fathers of the Internet. Bob feels uncomfortable about that. Vint is all about it. I think he's got a tattoo. It's awkward.

[00:28:01.08]
Ned: You're not wrong.

[00:28:02.26]
Chris: They wrote another paper that helped explain a lot of this stuff. They designed protocols that again, they exist now. They were theory back then. The biggest one is tcp, which you might know as the TCP part of tcpip.

[00:28:18.28]
Ned: Sounds familiar.

[00:28:20.04]
Chris: And the IP part actually came later, but there was something there beforehand. Again, I didn't want to get too deep into the protocol wars of the 80s and 90s, but if anybody else here has token ring scars. The support meeting is again at 8pm and everybody gets one chance to talk in a circle.

[00:28:40.26]
Ned: And if you want to hear a little bit about that, check out our episode from last week where we talked to Ivan Peplnyak about nat. But he got into the CLNP and IP wars a little bit too.

[00:28:53.26]
Chris: Yeah. Which honestly again could have been its own episode easily.

[00:28:57.21]
Ned: Yeah.

[00:29:00.16]
Chris: So what we ended up with is a way for disparate devices to talk to each other over a standardized network that was decentralized and basically bomb proof. And everybody is using standardized protocols, meaning everybody agrees how we're going to communicate. The Sigma 7 could only talk to the Sigma 7 back in the day. Now it was a very simple addition of this IMP device which eventually became routers. Like I almost called you Ron. Are you Ron?

[00:29:34.20]
Ned: Maybe. Wouldn't you like to know.

[00:29:37.23]
Chris: It did not take long for more and more and more and more and more and more and more and more and more and more and more computers to be online up until now, where the number of Systems Online in 2025 is so large that we can't come up with an actual number because every time someone tries to count it, the number changes.

[00:30:03.23]
Ned: Yeah, we just, you know, come up with something in the right order of magnitude is the best we could do.

[00:30:12.02]
Chris: Right. And even there we're like, I guess because if you think about it, some of the things that we talked about. So we've got the general purpose of the distributed network for the overall Internet, but all of us realistically are still on some kind of quasi centralized connection. So for example, your home network has one big ass distributor from Verizon or from Comcast or whomever that goes out to all the individual houses. Your house is not directly connected to your neighbor. That's just not practical.

[00:30:45.06]
Ned: Right.

[00:30:48.06]
Chris: We could. I don't want to go down that path either, because then we talk about Tier one, Tier two, Tier three, and then everybody falls asleep.

[00:30:55.19]
Ned: We've talked about some of this in our BGP episode, so I guess you could go listen to that if you want to dig into that side of things.

[00:31:05.19]
Chris: But yeah, I think it's pretty crazy to think about. When you start out in 1969, there are four hosts. By the end of 1971, there are 18 hosts. There are a hundred thousand by the 1980s, and in 2025, there are infinity.

[00:31:23.19]
Ned: Effectively. Yes.

[00:31:27.07]
Chris: Not bad.

[00:31:28.18]
Ned: Not bad. We did well. Good job, everybody. We can stop now. All right. Well, hey, thanks for listening or something. I guess you found it worthwhile enough if you made it all the way to the end. So congratulations to you, friend. You accomplished something today. Now you can go sit on the couch, fire up an imp, and send a message to a PDP10. They're very lonely and you've earned it. You can find more about the show by visiting our LinkedIn page. Just search Chaos Lever or go to our website, chaoslever.com where you'll find show notes, blog posts, and general tomfoolery. We'll be back next week to see what fresh hell is upon us. Ha. Ta. For now.

[00:32:16.16]
Chris: It'S a SCP 3008. That's where you need to start.

[00:32:22.16]
Ned: I'm sorry, what?