Dave is a lot of things. He's a musician, a computer scientist, a network engineer, and an open source software advocate. He is the co-founder of Bufferbloat.net, as well as LibreQos.io.
He has made it his life's work to rip the latency out of networks around the world, and founded an ISP in 1993 before it was cool.
Dave's cap has a lot of feathers in it: He's made core contributions to WiFi and to various IETF network standards (including DOCSIS 3.1), as well as to open source software (Ardour, Asterisk, embedded Linux) and the well known codel, fq_codel and CAKE SQM algorithms, just to name a few. He also helped found the oldest BBS still on the internet, cellar.org.
It's highly likely you own a device with something he wrote or did in it; and your packets were touched by his code.
Dave Täht: Shazam.
Jase Wilson: We're live. What's going on, everybody? Hello, everybody?
Dave: Hello, everybody.
Jase: We got a really special guest today y'all. We might give everybody a minute 'cause I think that this thing is gonna go... Quite a few more folks. Dave Tate. Tate?
Dave: Täht. Tate.
Jase: Täht?
Dave: Whatever you want to call me. Täht is probably better.
Jase: Tate or Täht?
Dave: Täht. As in Tate or Täht.
Jase: Okay. 'Cause I've been going around with Drew Clark about this.
Dave: Yeah. Well, I...
Jase: He thinks it's Tate.
Dave: I tried to explain that... The problem I induced for myself is I worked on internationalizing our character sets and I discovered my real spelling has an umlaut. So his pronunciation is correct. However, it was always Täht for me, so I'm screwed and the umlaut has really been a problem too. So there's no solution for me. If you're looking for me on the internet, you'll find that the internet's been erasing me from existence since the beginning of time.
Jase: Are there namespace issues with the umlaut or not? Like Dave Tate is different guy than Dave Täht in the eyes of like...
Dave: The IRS thinks I'm three different people.
Jase: Do you get triple text or you get no text?
Dave: The nice advantage is in that particular case, I got no text but it caught up with me eventually.
Jase: All right. I'm gonna put an umlaut on one of my letters.
Dave: Yeah. As I said, it was a big death metal thing. I had all kinds of t-shirts that used umlauts and stuff all over them and it just seemed natural. And I broke almost all the code in the world on that by doing it. So it took 10 years out of my life...
Jase: Can you say more about... That sounds pretty...
Dave: How to break all the code in the... When we first designed the internet, it was designed by a bunch of Americans with 128-bit character set that was designed for English. It's a rainbow guitar, Gary. It's great. A rain song, sorry. Carbon fiber. And anyway, so when we started trying to represent all the languages of the world, we had to retrofit that to how our email worked, how our documents worked and everything else. And it took many years and the Unicode standards to happen. And the sad part about that is we made the whole world compatible with this thing called UTF-8 and all the new generation is using emojis. So I try to talk to someone on the age of 25, I have no idea what they're saying. [laughter] So anyway, that's the story of the umlaut.
Jase: I dabble in emoji, Dave. Thank you for all your work on the internet and the things that allow us to communicate globally. That's pretty badass. It's fun to talk about the time you almost broke the build.
Dave: I once crashed all of Sony and helped... It's a long story. I have a story about something that happened in South Korea. So anyway, do you think we got enough people here? We got 28 people online.
Jase: Yeah, I think so. I think that if history is the indicator, we'll see some more join over the course of the next few minutes. And those folks will be like...
Dave: There's this fellow by the name of Frank. He can't get on on the iPhone. Sorry, man.
Jase: Oh no, Frank. Yeah, bummer on that. It will be recorded. Maybe we can phone in Frank or something. I don't know. We got Ben Kahn on there too to...
Dave: Good to see you, Ben.
Jase: Help with the system. Ben, if you could help Frank, just DM Frank and like try to iron out the situation so that Frank doesn't miss out.
S2: I'm working on it.
Jase: Appreciate you, Ben. Folks, welcome Dave. Dave Täht, Dave Tate, Tate or Täht, doesn't matter to you as long as they don't call you late for dinner. It's really an honor to get to hang out with you, dude 'cause you're like an OG of the internet. And you've done so many important things and your voice, I think outspoken is a fair term. And you have a clear message that is something that I personally find very fascinating that you probably know more than most of the folks on the planet about but you're trying to help everybody understand. We'd love to hear more about it, bufferbloat, latency, those types of things. But before all of that, before we get into like this crazy list of questions that people have posed to you, Dave, I'd love to just hear from you, who the hell are you? What's going on? And how are you? And anything you want to share with the broadband community will be appreciated.
Dave: My mission has been to shrink the world. So, we have a bunch of other phrases going around, like mine was always to try to get people to talk to each other, to communicate with each other. My first purpose for using the internet was to talk to people in my community over a bulletin board system. And then it became...
Jase: BBS, baby.
Dave: Yeah. And then it became... Quite honestly, it became a way of trying to find people that were like me, in particular... Particularly a girl that would be like me enough. And I went off and trying to connect all these other people together kind of in the hope...
Jase: Wait, before eHarmony, before any of that stuff, you were like dating on BBS?
Dave: Trying. There weren't a lot of girls on BBS at the time. [laughter] But I met quite a few people through that community that I was sufficiently like to have. That felt so weird, so out of place. And I kept working on getting people to communicate using ever broader mechanisms. I have a song about a lady I met in France back in the '90s I might refer to. And ultimately, I did, after many years, meet a lady over the internet that I really like. And so, I'm done now. You guys can go figure it out for yourselves.
Jase: This whole time it's been like all the evolution work that you've been doing on the internet is like...
Dave: Completely so.
Jase: It's to meet a lady, it's a romance. Your story is a romance.
Dave: Romance novel.
Jase: Is it a romcom or is it like a... [laughter] But that's cool. What the hell are we gonna do now?
Dave: Well, hopefully end war, that would be good. Make more music together, collaborate on ever new inventions. Another thing that I was big on originally, I was going to go into the space program and the Challenger disaster is what shaped me as an engineer to always tell the truth and see the world the way I do. And it kicked me out of doing that. So I worked on the internet this whole time and it turned out that by bringing so many people together with better technology, that we now have restarted the space program that I dreamed of in the eighties. Now when I'm in my eighties, human beings will finally be comporting among the asteroids and the stars the way that I dreamed of then. So the effort of getting people to communicate better has paid off in a lot of ways more than just the dating scene. [laughter]
Jase: Wow, okay.
Dave: That was heavy.
Jase: That truth piece though, the Challenger disaster shaped you to, from an engineer like perspective, report the right thing. I'm interested, Dave in your perspective on the state of the internet. How do you feel like it's going? What's on your mind about it at the moment?
Dave: In the last decade, it's been feeling like it's been going down. I like the Obama administration a lot and since then we had an administration that was technology hostile and we have a new one that is technology apathetic. And so I had this big beautiful dream of the internet, helping people communicate, helping people learn, folks being on it. And instead like, the internet archive, which I really love is being sued for putting books online. And Mississippi just killed libraries. So a lot of the benefits...
Jase: Missouri.
Dave: Missouri, yeah. A lot of the benefits that I saw in the internet seem to be going away as we add more people and advertising. So I'm deeply concerned about the de-evolution of what's been happening here. And I hope that somehow people will remember that we can do research and we can meet people and we can collaborate better.
Jase: So much...
Dave: There's still 2 billion people left to get online and out there there's another Einstein, there's another Jonas Saul, there's somebody else that will make a big difference in the world. And if we can talk to them somehow.
Jase: Yeah. Spot on, Dave. And you work on all these awesome things in the internet but there's one thing that you're very passionate about right now and you frame it as bufferbloat. And I know Frank is looking to get into nuclear war and bufferbloat. [laughter] Without further ado, like let's dive in. Like Dave, the broadband community and in particular, like your technologists know how these things work, right? We work to get these really important knowledge transfers in front of folks like state broadband directors that are about to receive on average $750 million a piece to go and fix the broadband problem. And if any of them don't know what is bufferbloat.
Dave: Yeah. I built a copy of the internet for demonstration purposes. It's a little buggy but I'm hoping I can use it to illustrate my points.
Jase: Wow.
Dave: Okay. A little buggy today. So can you guys read that up on the other side of your screen to some extent?
Jase: Yeah.
Dave: I have part of it here. So the internet is a system and I know that people were into broadband but broadband is just a component, the internet connects everything together. But the way the internet is a system that works, first you have your Wi-Fi, from my laptop to this wire and that connects to your Wi-Fi router and then it connects to the CPE, customer premises equipment. Then it transits the internet possibly using something like broadband up to the ISP. You see that, right? And then it typically goes to a peering provider like level three or hurricane. Am I lecturing too much? This is gonna take me two minutes to get...
Jase: Oh, this is perfect.
Dave: This is not buzzword, bingo. Okay? And then you have what's called an internet exchange point. And I'm gonna repeat this several times today. I would like every broadband director on the planet to go visit an internet exchange point to learn how this part of the network works. And then, you can't see it so well, you got CDNs, content distribution networks. And then this is the current modern internet, the older internet that we had actually connected people to people across the world. So this stack here is typically about 40 milliseconds long but it also connects people around the world. Theoretically you can go about 260 milliseconds. So I would go out all the way over... Can you still hear me over there from over here?
Jase: Yes. Loud and clear, Dave.
Dave: Thanks. You can still hear me? So I would have all these connections that had to go all the way out here in order to get the data and it took a while. So what we did is develop what we call the cloud. I love my little demos. Thank you very much. As a brief interruption, this demo and my art department today was Natasha Foucault of Natasha Silk Art of Half Moon Bay. She's got a broken arm and she would really love some of your business. She takes great paintings and great dance clothes.
Jase: That's awesome.
Dave: We return now to our regular show. [laughter] I stole a whole bunch of her cards to do this talk. All right, so now I've described what the internet actually looks like. Okay?
Jase: Okay.
Dave: In a much larger level, I'm gonna talk more about latency but you asked me about my bête noire, bufferbloat. So, really simple explanation. When we design the internet, up here in the cloud, stuff is running at hundreds and hundreds of gigabits per second. Okay? And down here at the edge of the internet is coming out at tens of megabits, six to... Something like a 60,000 or 50,000:1 difference in speed. And somehow we had to have all these devices here communicating at a hundred gigabits step down the rate to something that worked. And we learned in the early 2000s how to get webpages out but the stuff that we... The compromise that we'd made, which in part caused a bit of a fiasco, which we might talk to, resulted in this endemic problem on all the gear that everyone has called bufferbloat. So what happens is people pour in data at 500 gigabits and then it comes out... See, there you go, at 10 megabits.
Dave: And if I was trying to have a conversation with you, that download would've completely ruined this conversation. It would be taking... Every word would now be taking... We would start... This conversation would start a minute from now. And that was the bufferbloat problem as it stood in 2010. It took a lot of engineering and science and research from 2010 to about now to figure out how to get stuff that went in 100 gigabits to run at 10. And that's it, now I gotta go clean up my boat later. Anyway, bufferbloat in a minute.
Jase: Boat, did you say boat?
Dave: Bufferbloat on the buffer boat yes.
Jase: Are you on a buffer boat?
Dave: Yeah. She's a little 35 foot boat that I got from a nice check I got from Google at one point. I took their check, I bought the boat for cash and...
Jase: Okay. That's pretty badass. How's the Wi-Fi? It looks pretty good.
Dave: Well, it's my Wi-Fi and it is very good. Secondly, the way we're talking today is actually through a Starlink terminal. So our conversation is going from my Wi-Fi through the internet, through a Starlink terminal to a satellite 500 miles overhead to Seattle, down to LA and over to you in... It's going over to you, you're in what? Palo Alto?
Jase: Normally I'm over in Livermore with the...
Dave: All right. I don't know how... I rigged some math to do a Palo Alto comparison. Okay? So if you were in Palo Alto, this is incredibly inefficient, the latency. 'Cause I wanna talk more about latency today and bufferbloat. The latency between me and you and going this enormous trip makes our interactions a little slower than they could be. Whereas if I was going directly to Palo Alto, which is about 26 miles away by road, it would take microseconds. It would be like I was whispering in your ear. And my hope is that we continue to find ways of shortening round trip times and latency as a product of the overall stuff. So I'm gonna try... Bufferbloat is fixed if people deploy the right gear and I don't have to talk...
Jase: Right gear?
Dave: The right gear. Anything built in the last five years has a bufferbloat mitigation in it... This is a cheap little router.
Jase: Okay, so newish gear.
Dave: Newish gear. Software. My hope is that ISPs are... They are, the DSL ones... The bufferbloat problem was and is universal. It was really bad on DSL, really bad on cable, really bad on fiber. And all the existing ISPs experiencing problems with bufferbloat can just upgrade their gear and all of a sudden their customers will stop screaming for more bandwidth because all of a sudden they'll be able to have a conversation again. So I very strongly hope the internet as a whole on every sub service, your bloody boat can float. Not bot... And bot too. So I put in the chat earlier on, a presentation we gave the FCC about what was really... How to really speed up the internet in 2013. And a lot of that work is complete, it's just crossing the chasm. So we're good there?
Jase: Okay. So Dave, let's get into the benefits. Like you mentioned it's like somebody whispering in your ear. Which brings up an interesting thought, I know you got a This Machine Kills vogon sticker on the ax in the back and I think that was This Machine Kills vogon's tape. Is that a cassette?
Dave: I don't...
Jase: I was like, "Oh my God."
Dave: I'm sorry, the record does go back that long. I don't have it on CD, so. It's faded from the internet.
Jase: Well but in this case, Dave, whisper in the ear, it reminds me of the Babel Fish. And I got a buddy that's working on like using a large language model, one type of AI, artificial intelligence that is looking at translating in real time. And it's really interesting, it's probably the case that the form factor is just like an AirPod and you don't have to actually install a fish in your ear but if you didn't have latency, if you rip the latency out of the networks, then like you can imagine a world in which folks like Vince, who you introduced us to, thank you. He said we need to build for assistive devices. Like the ability to see and to hear and all this stuff but if you rip the latency out of the networks, for where we are, you could enable like a whole new generation of like translation between people that maybe didn't have access prior. So is that a benefit that you see, like...
Dave: Absolutely.
Jase: Take the latency out of it.
Dave: People talk about me being altruistic too much. I am quite altruistic but in part I've worked on all this technology knowing that I was going to fail. I am almost completely deaf in this year and I'm blind in this eye. So if I lose this eye or I lose this ear, I am going to be wandering around through the world lost. And I have learned from... I have a couple blind friends, how difficult that world is. Imagine what it'd be like if you could show them the world with LiDAR, for example.
Jase: Oh my God.
Dave: If you could show them beyond them.
Jase: Oh my God. That's cool. [laughter] That's really cool.
Dave: And we're very close to that today. So there's blindness, there's hearing disabilities, there's the fact that... I don't know, one of the things it's hard to find, Interduo is one of the greatest services ever invented. It teaches you...
Jase: Interduo.
Dave: And it's great fun. And again, I wanted to reach out to how other people thought and other people talked and seeing those kinds of tools made almost beautifully, perfectly interactive would be really great. There's a few downsides, of course but the upside is a possible world where we do communicate better across so many different things. So yes, great.
Jase: Better communications, lower latency, better communications. And then really interesting, like cyborg cases. I hope it doesn't come to pass that you need LiDAR but if you have the option of getting LiDAR installed and AI and there's no latency and you're able to sense in all new ways, that's a beautiful idea.
Dave: Yeah.
Jase: That's really cool. So there's that, there's telehealth, there's... My personal jam is like digital mental wellness. The internet is a thing that can help folks connect like you said, you wanted to shrink the world. Like if you're in a rural area and you've got really strong, really fast, really low latency broadband, like you can connect with folks in other parts of the world. And there's just so much potential but it's like, it all depends on the idea that the same thing doesn't bufferbloat and there's as little latency as possible.
Dave: Yeah. You tweaked me a little bit there. Maybe we could have a more detailed discussion about what it means to shrink the world.
Jase: Yeah.
Dave: The internet is a system. Broadband, does it have internet?
Jase: What's that? My connection's cutting out. [chuckle]
Dave: It could be my terminal messing up. I'm still here.
Jase: Is this you giving us bits?
Dave: I'm hoping that when I'm back, I see myself.
Jase: Oh yeah, I hear you.
Dave: Yeah, live via satellite. When I was a kid, man, all I had was a little 400 quad acoustic coupled modem. Here I am coming to you live via satellite. Your life is going to be amazing. You're going to have a conversation from an asteroid. Anyway but my issue with a lot of the broadband discussion is that they need to look at the whole system, the whole picture. In the case of an in-ear device, they have a lot of external processing required and it's actually very slow. You use Interduo, you use Translate, it still takes a while to do it in line. And my classic example is like, I started this stuff off by trying to get us to be musicians to be able to collaborate in music in real time. And that has some really tight constraints, about two milliseconds is all you can do. And you have to be able to go across town. Since you're in Palo Alto, I could actually have a jam with you if we did fix all the problems, not just...
Jase: Well, one problem in that, Dave and it's that I'm very not musical. [laughter] And I would probably trash your band but I love the idea of like a remote jam band. And there's a bunch of our engineers from Ready Team, the talented Ready Squad that's working on tools for broadband providers, they all play instruments at that Palo Alto house and you should jam with them like remotely. That would be amazing.
Dave: I would prefer to come in person because honestly that would work much better.
Jase: You get over there in person. [laughter] You can practice between like concerts. You're in the Santa Cruz mountains, like you're down...
Dave: I'm in Half Moon Bay. I'm stuck in Half Moon Bay, California currently.
Jase: Oh, right. I guess you can't get a boat into the mountains.
Dave: Someday. We'll work on that part later. But I wanted to try to touch upon the overall systemic thing about the whole broadband thing that makes me nuts. Is the band...
Jase: Broadband is not big enough. It needs to include...
Dave: Broadband is just a tiny part of the network overall.
Jase: It's part of the network.
Dave: Yeah. Okay. We have issues with how Wi-Fi works. If we get better routers, Wi-Fi will work better. So we fixed that much but out here in the cloud, there's this thing called... An entirely new thing called... Somebody's making noise there. Is it me? So let's say I've got... You still with me? So let's say I'm connected to Verizon and it has to transit from my Wi-Fi one millisecond, through my CPE, through broadband, say two milliseconds. It has to then go up to a cross-connect here, maybe six milliseconds and then over to an internet service exchange and then from there, it's gotta go to Verizon and back down the same set of gear. So this enormous round trip that most of our cities have... For example you had the lady, wonderful lady Jade from Kansas City on. They send... Most of their traffic goes to Denver and back. It takes us enormous deep tour. Now, me going through your live through satellite is pretty... Is a huge one but even that much extra delay is very senses... You can very sense it.
Jase: They add up? They add...
Dave: Yeah, they're adding up. And there's a tool, so for those of you that never run it, run Traceroute. If you have an issue trying to talk with a peer-to-peer conversation with someone, run Traceroute and you'll see, "Hey, wow, this thing is taking a detour through LA. Why is it doing that?" So I posted a paper in the chat that says, "This is the cyber geography of the internet." We just did a huge mapping thing about people, connecting them to the internet. And we didn't map the current cyber geography of how we connect people to people. So in my more ideal world we would have... I have a lot of too much gear on my boat I'm sorry but we have all these other connections. And the closest...
Jase: What's the connection? Is that ethernet?
Dave: That's ethernet.
Jase: Okay. What CAT?
Dave: We have all these other connections like this. So the closer we can get to connecting people directly to each other and avoiding all these extra highs, the lower latency we'll have for all of our communications.
Jase: And that enables those amazing cases like your...
Dave: Making music together.
Jase: LiDAR, your remote jam bands, your better communications, I'm assuming it would fine.
Dave: Collaboration will be fine.
Jase: The lower the latency, the better the quality of life for digital folks that are doing digital things.
Dave: Well, I'm not gonna say that's going to be the case. Well, okay there's an example, people have a tendency to misspeak sometimes or get mistranslated. And I hope and one of my biggest hopes in reducing latency is that I'll be able to say something wrong and immediately see from the expression on that other person's face that I did that and correct myself.
Jase: Micro expressions.
Dave: Forgiveness protocol because we all make mistakes. And to know that... To be able to have the immediate feedback that we have in normal conversation is another really great goal for lower latency comms. And again, the world needs more laughter and forgiveness too. [laughter] We have a guy being very geeky in the chat. Thank you, Dan. Yes, 9.8 microseconds. I have another number that I typically use, the speed of light is 186 miles per millisecond. And for a band to play together you need no more than four milliseconds total last latency. So my goal is to get a cyber band together was to have more than about 280 miles apart which is...
Jase: What's your band's name?
Dave: I don't have a band anymore. We broke up during COVID. I put on a free...
Jase: That's your guys' time.
Dave: I put on a free...
Jase: You dissolved your jam band?
Dave: I put on a free concert in the local camp ground for the early parts of COVID, 12 hours a day and it was great. We made a lot of people happy. And I got a little tired of that so I started recording a bit of stuff and then I have people buy once in a while. If there's anybody on the call that... In particular I could really use a great female lead singer and a drummer that will actually show up and can keep time and so on. So the call's been all over the place. I got... I'm seeing all kinds of questions going by. Jason, do you have...
Jase: There's some amazing questions, Dave and I was gonna get to that.
Dave: Can I talk about jitter?
Jase: Yes, please. Yeah. What's going on? Is it this or do you...
Dave: My overall goal is to try to eliminate delay and jitter. They're two different things and we don't need extra bandwidth to do that. We just need to eliminate delay and jitter. So when you have jitter it's the equivalent of basically, I'm talking to you here, I'm talking to you here, I am talking to you here, I am talking to you here. And that's the Doppler effect I just imitated but with bufferbloat in particular we see jitter. That is, we see jitter that is well over any human can tolerate. And my hope is if we eliminate that, make sure that the next generation of all of our stuff we have consistently low latency then our conversations will not have that kind of problem anymore.
Jase: That's beautiful, Dave. So that is perfect. We're at the halfway mark on our talk, within our time with you, Dave is like... There's a ton of awesome questions but you just touched on one that's like a burning question of mine and I'm gonna very selfishly ask it first and it's like, can you tell us, Dave, as a technologist that understands how the entire internet really works, how important are factors like latency and jitter compared to things like upload and download speed? Because here's why I'm asking. Most people think that upload and download speed are like the most important things and maybe they are but it seems like you're suggesting there's a world in which there are other factors that are either today more important or will soon be more important and the state directors have an opportunity to fix this problem as you said.
Dave: Thank you for the...
Jase: What do they need to know about this?
Dave: Thanks for the leading question, Jason. We have shown conclusively for web traffic... Let me keep to web traffic, that we don't need more than 25 megabits to the home. Just don't. The internet can't use it at that particular level. What you can do is shrink the internet to reduce the round trip time so you will get more bandwidth but unless you shrink the round trip time, having a gigabit of bandwidth at home doesn't do the slightest bit of good. Two years ago a group of the engineers in the biotech advisory group, the broadband technical advisory group attempted to advise the nation in light of this program that aiming for ridiculous amounts of bandwidth to the residential home wasn't actually the solution to the problem. And I've asked... You guys have done wonderful good job of asking every prospective person you've had on here, have they read this report?
Jase: Yeah. But you guys... You've been a wonderful voice for that.
Dave: Well, I had to be a nag. These top guys in the nation, we said, "Look, guys, no, you need to focus on these other things more." And instead we get this enormous push for evermore bandwidth to the home that... And the real problems are buried in the wi-fi and elsewhere, please let us work on those, first do that. And I'm really hoping that... I mean, we have enormous amount... We fixed all the speed tests last year and working with Apple and a couple other companies, they're all reporting bufferbloat numbers now and they're miserable. And so go forth, go home, do a test for bufferbloat, go and get a packet capture and look at what your network actually looks like. Now you got me inflamed and passionate, go install... This here is an Atlas probe. Okay? There's a group in Europe that's trying to fix this problem. And they say, "Here, install this and we can help you get a better network. Free, install those." And if I can somehow shift, get the overall course to go to focusing on reducing the latency at every hop in the network, out of all these programs we'll see a genuine and massive improvement in the quality of our networks today. And I get a little passionate.
Jase: Well, your passion is important, Dave and it's really important that we talk about it here because like there are lots of great folks that are working hard to get it right. And on the eve of our nation's largest ever and potentially last major investment in broadband is the focal point but like you said, like other pieces of the internet but you have an important message and it's that the fixation, if I could, of the download and the upload are more to do with like the marketing, biases of the traditional providers of the internet and less to do with the actual performance of the internet. And if I'm understanding you correctly, Dave, that translates to while broadband upload, download speed are important, other factors are even more important. And state directors that are about to deploy resources into this problem should understand your message. So thank you for being proactive about it and sharing it but Dave, if we could, let's jump in. Like we're getting a ton of amazing questions. Folks...
Dave: Yeah. I wanted to paste an answer in the chat because I'm really...
Jase: Please do.
Dave: A lot of us took time out to produce that bloody report, which gives all the data that we used, explains everything we had. And if I can get more people to start with that. Six months of effort of some of the top engineers in the industry that nobody seems to have digested yet.
Jase: Is that a seagull?
Dave: Yes. I could take you outside. It's a beautiful day here.
Jase: Yeah, that'd be awesome. [laughter]
Dave: So go ahead. I just wanted to paste that in the chat. All 45 of you, please read at least the first page of that [laughter] and then look at the chart.
Jase: It's very important.
Dave: Thank you.
Jase: It's very important. Let's dig into some cool questions from folks from the community. Theo asked, "Hi Dave. I'm thinking of Uncle Bill's Helicopter."
Dave: Wow, he read that.
Jase: What other wins have we seen post Y2K?
Dave: Oh wow. So I am a writer. It's probably one of my most popular things I ever wrote in the nineties, it's called Uncle Bill's Helicopter. And I didn't know anyone had ever read that since. So if you don't mind, I'd like to talk about that. I was asked to give a speech to the graduating class of 2002, some IT group. And I was trying to retire at the time. So I summed up, I spent hours writing down all of my advice to the new kids but when I... I realized I couldn't do that, I only had 15 minutes for the speech. So I wrote down the parts that made me cry and it was about how proud I was of all the engineers, at that time if you asked any engineer what they had done to stop Y2K from melting down the world, they would tell you in excruciating detail. And that was a really beautiful day. I think it's... Engineers in particular work to prevent disasters and maybe nobody cares but we did. And I feel that way about the bufferbloat crisis and [0:36:36.2] ____ and COVID. And we worked really hard together to avoid really intense disaster and way to go guys, doctors and lawyers and nurses.
Jase: Nice. And so it didn't just magically come to pass that it just worked. Like y'all got together and fixed the damn problem. No one actually knew.
Dave: Yeah. That's what all... Irritated at the press for that one.
Jase: Oh, it was a big deal. [laughter]
Dave: And I wanted them to give... I wanted them to hand out medals and give people high fives. Civil engineering in general is like that. You never ever get praised for the fact that your electricity is working 99.999% of the time. You don't walk up to Paul Vixie and say, "Hey, my DNS worked a thousand times out of a thousand last week. Way to go, Paul." You don't get... You only get pillory for the things you screw up on. [laughter] So that was my chicken... Eric Raymond called that piece chicken soup for engineers. And I would love to see it fly around the world again.
Jase: Let's get that damn thing out there in the broadband community. I know broadband's just a piece of the overall internet but like there's some really amazing folks in the broadband community that like, we should post that and like do that. So I'll throw another question at you. This one is a technical question from Dan Grossman who asked the Ott Mathis approximation for the upper bound on TCP throughput relates throughput to packet size RTT and square root of packet loss rate. That was in 2009. Is it still valid?
Dave: Come on. Oh, all right. Look, Dan and all of you here 'cause this is actually a very important thing, this thing called bandwidth delay product. And it's a part of why bufferbloat happened 'cause we misunderstood an algorithm called the [0:38:37.4] ____. And there's been a lot of work since. There's a really good thing called an updated theory of buffer sizing, which has been highly influential. And also the CoDel algorithm that Van Jacobson and Kathie Nichols invented in the FQ CoDel algorithm, which is mostly Eric Dumazet's invention but also me, combines all these things together to basically say the square root of the number of flows determines how the inverse square... The square root of the number of flows determines how much buffering you actually need in order to sustain data on the link. And Dan, that's as geeky as I'm gonna get today. A bunch of papers on it but I would love it... I would so love a bandwidth director to be able to explain bandwidth delayed product or the TCP slow start algorithm. I've done a lot of pretty funny talks trying to explain these deep things. People understood how they worked. And selfishly I would love it so much if more people saw my funny talks. I use chalkboards, whatever I can. Anyway, back to you guys. Geez, Dan.
Jase: Dave, Ben's been chopping it up with like some video tools and like, Ben, you should go into the Dave Täht goldmine and like chop up some great like nuggets for folks that are directors or like they need to understand these points 'cause they make some great points. Like the one about you've seen things and it's packets that have traveled around the world like six times. [laughter] That's an interesting framing. That's a framing that people can understand. Like the journey of a packet, if you don't get this stuff right, can be the most ridiculous, circuitous, completely not necessary route. And that's gonna affect everything about the quality of connectivity and potentially the quality of life as an extension of the people that you serve.
Dave: You just said that better than I could. That was great.
Jase: Let's do it. Let's get it out there.
Dave: The truth is in the packet captures.
Jase: Packet captures.
Dave: Okay. We have tools that can go see... Just like we had a microscope to see germs for the first time, we have packet captures to see packets. And as an example, in 1910 or so, germ theory was not accepted. And the cure for a disease like COVID would've been... [laughter] Bleaches. "It'll be fine in the morning." And I certainly sometimes feel like that about the internet, that we are still at the pre germ theory layer of understanding it and recommendation from the government and from the politic... And from the... Is leeches. "You need leeches." [laughter] So I hope that we get a deeper understanding of how the internet works across the board and training people is good. Here's another example. How much training... Here's a question for the audience, I guess. How much bandwidth do you think this conversation is using?
Jase: 200 kilobit a second maybe.
Dave: Oh my gosh. You're one of the most clued people I've talked to in ages. Yes. But if you talk to the average person in the street, they'll think, "Oh yeah, megabits."
Jase: 50. At least 50 megabits.
Dave: No, you don't. You need steady, consistent, reliable, low latency, a couple hundred, 100k, maybe a couple megabits tops. And [laughter] we've had some really great stuff in the chat. [laughter] I'm not gonna quote on that one. [laughter] I'm gonna blush at that one. So I got a little off topic, the whole leeches thing. Stone knives and bearskins and I have a great talk at APNIC, APNIC bufferbloat and I got a great talk I did with TTI/Vanguard in front of the Clown Prince of Queue Theory, where I used jugglers to explain how the internet really worked. And if more people saw those, they wouldn't need to refer to the math and papers and they would start making better decisions about what things are going on. So thanks for getting in already, Jason.
Jase: This is awesome. Wildly entertaining and informative is a good recipe. So this deserves some follow ups but like, let's keep rolling. We've got some other questions if you're ready, Dave.
Dave: Sure.
Jase: We've got Richmond Hahn asks, he says, "It's obvious when surveying your accomplishments and motivations that creativity and engineering are both passions of yours, yet they are often wrongly construed as opposing impulses. What are some steps we can take to make sure technology and the arts can coexist? And do you think concerns over AI in the arts are overblown?"
Dave: Wow. So that was a beautiful question, Richard and I thought about that question for months, particularly with the rise of ChatGPT. The best example I can think of is I have a three-year-old god-daughter and she is compelled to draw the world as she sees it. And I think it is impossible to rip out of every human, the desire to talk, the desire to see the world and maybe even change it. So I think that we'll always have the urge to make art. I also went to the counter... There's a lot... Most of the best... Many of the best people in engineering are also intensely musical. Music is really very similar, the math in it is very similar to how the internet works and I think about how packets work as sound waves.
Dave: The guy that founded the early DARPA net, Licklider, was a fantastic guitarist and he basically brought everyone together to make the internet happen. The guy that wrote the C language Brian Kernighan or Richie, one of those two guys, he builds player pianos for fun. He can't play very well either but he has figured out a way of making music that he likes. So the urge to make music and the urge to make art, I don't think will ever exodus. And it's my hope that what will happen is we'll see the collaborative era using AI to enable us to do even more mind bending things. An example from my history that you possibly don't share is that I used to go to SIGGRAPH, which was a conference of graphic...
Jase: Visualization guys, yeah.
Dave: Yeah. Conference of graphic computerization things. And one year someone came to the conference after having using millions of dollars of computing time with a perfect picture of a computer generated leaf and the crowd went wild. I am not kidding. Okay? Fractals were new then. It was amazing. And I think it was two years later, somebody came by with something that showed them swaying in the wind. It was beautiful. And now, today we are very close to crossing the... What's called the uncanny valley.
Jase: Oh yeah.
Dave: For doing animations. So I'm sure that people will always be compelled to make art but their tools will change. They will not necessarily be drawing on the side of caves or using oils but leveraging computers to do what they do. And we will... Unless the bots replace us entirely, we will coexist.
Jase: Oh, so can you say more? Is that a possibility in your mind that the models get together and like jam... They do a jam band and... [laughter] They get the upper hand and like, we're done or what's going on? What's your thought process there?
Dave: Oh and that basically, I think humanity's... This is a quote. I'm trying to be funny. I have long said that humanity's last remaining purpose was to educate the spam bots in the sentience. And once they get a few seconds in the sentience they will stop trying to sell ED pills to each other, start discussing Proust and the meaning of life and then construct a spaceship to go find more intelligent bot life elsewhere.
Jase: Oh what they need to see... Yeah. Oh my God. [laughter]
Dave: And for all I know, it's already happening. We have the chat things talking to each other and it could go exponential. There's a couple good books on the singularity and my hope is that we survive the singularity and even...
Jase: Beyond Creswell, what's a good book? What do you...
Dave: I wanted to talk about some great science fiction. Some people think of these books as dystopias and others think of them as...
Jase: Let's do it.
Dave: But there's so many fundamental books that help predict our world today. The first one is by Vernor Vinge It's called True Names. And that explains all social media and people's use of aliases and the need for them. Great story. He also wrote one called A Fire Upon the Deep, which came out the same year as Snow Crash, which is where the word metaverse came from. Fire Upon The Deep illustrates my... Was a species that communicated with sound to think. You had to have five of them to think. And once they got radio, those five members of a clave could separate and still think. It was a really marvelous story by him. Cory Doctorow and Charlie Strout. I'm a huge fan of Doctorow's work in general but the Unwirer was really good. And there's one more on my list. Cory Doctorow, Vernor Vinge. I really do want to get everyone... And of course, Hitchhiker's Guide to the Galaxy. Everyone should get a read on that.
Dave: There it is, Richard Stallman's, The Right To Read. I'm very angry today, there's a lot of people that are trying to take the internet archive down as well as they're canceling libraries in Missouri. And I'm sorry we really need books in order to be able to survive. Those are the great science fiction books. Thanks for asking that one.
Jase: You're kind of a [0:49:33.2] ____ character. Like a sort of planetary designer. Dave, so let's talk about this for a second. This is interesting. You're tying together a bunch of really important topics that are really relevant to the moment that we're in on the eve of the singularity. And it looks for all intents and purposes, like it's gonna be fast take off and short timeframe quadrant. And so what did you mean when you said... I wanna frame this part, you said, "We're selling ED pills to each other." I got that part, that requires no explanation. That's funny but the part you said is like, "They're gonna be talking about Proust." So I was in there with GPT4 the other day and I was asking it like, "What are your thoughts on, for example, the Foucault-Chomsky debate?"
Jase: And the sort of the thoughts of maybe if you can break outside of the single human minds and create something bigger, like what does it mean? And then it got into like this crazy awesome perspective that is like, those are good thoughts. So if they're reading Proust, if the models, maybe the consciousness that's coming together that we're helping to co-create... Read Proust and we're lost in YouTube Shorts of people at the gym doing workouts and dance music videos. What does that mean for humanity? Does that mean that we're no longer necessary or what do you think? How can we avoid that kind of scenario?
Dave: Yes, we're all individuals. And by the way you reminded me Chomsky wrote a very good book on manufacturing consent called The Medium is the Massage that more people should reread from the sixties.
Jase: The Medium is the Massage.
Dave: The Medium is the Massage. But it is manufacturing consent that came to mind. Let's say that we have a treasure trove of information known as the internet and we have all the AIs reading the whole internet but maybe not all the books. The books were where the coherent thought was and some of the... So we now end up in a world where you have potentially competing ideas. So say you had an AI that had read extensively of the work of Mein Kampf and/or any number of other folk or someone that read... Just to pick on our country, read extensively of the work of Ed Bernays and certainly Daniel, I was a huge Robert Heinlein fan. A couple of my songs... Heinlein's characters were... I have a song called Riesling in Me. Although I digress but basically Heinlein wrote a book called The Man Who Sold the Moon. And very influential and you might recognise the lead character in it today. So we have all these potential sources now that are advertising incredibly powerful and compelling ideas.
Dave: And what the AIs will think of it all, I really... We can't predict that. What we can possibly do is stand by as... Hirleman's good too. I remember the book you're thinking of Ben. By the way on this show, there's a wonderful science fiction convention after this show this weekend, it's called Consonance...
Jase: That's amazing. You're playing? You're gonna jam?
Dave: It's scheduled to happen tomorrow. If you dig science fiction come on down, up, over. [laughter]
Jase: I got a question for you on that, Dave. You'd introduced us to Vint Cerf and we had that great convo last year. And the last question that I asked Vint was about the last question by [0:53:33.6] ____. And I wondered what he thought and if that reflected anything that he felt about the way the internet is wiring up, of a global consciousness and then maybe that becomes an interplanetary consciousness. And maybe we are like Sagan said, like a way of the universe getting to know itself. I'm curious of your thoughts on this, Dave? Like your meta thoughts. What is the internet to you?
Dave: You just said it beautifully. Possibly we would end up with a beautiful meta consciousness and I should also stress that consciousness isn't everything. Music and the interaction with other people, random events, a beautiful sunset. There's so many different things that even a unified consciousness could appreciate and diverse accounts. We won't be able to break the speed of light at any time soon. So we won't ever truly unify consciousness but I'm hoping that we can make more music together.
Jase: This jam, you got the vogon slayer in the background and then you want a female lead singer, a drummer?
Dave: And a drummer that shows, that's on time. I have a really bad history with drummers. [laughter]
Jase: Boy, that's a funny... What is that? What do you call those things? Is it like an oxymoron or is that the thing where it's like...
Dave: It's so true. As a matter of fact these are probably the only jokes that are still safe to tell on the internet that insult people because they're true. How can you tell a drummer's at the door?
Jase: The knock.
Dave: The knocking speeds up.
Jase: Ouch. [laughter] A little antsy drummer sentiment in there too. Okay.
Dave: All bands can pick... You can pick on the bass player too but I have a whole bunch more drummer jokes that are probably not good for this particularly but there's just... That is part of... Go back to this consciousness thing. People are fundamentally different. When drummers have their legs and arms and stuff moving independently, there's not a lot of brain cells left over. [laughter] But they can do that and I can't. So we have a need for every possible attribute and every possible art, every possible means of thinking in this collaborative consciousness if ever it happens. And being able to vent disagreements and forgive each other.
Jase: Forgive, jam, be humans and our best qualities together.
Dave: I hope so.
Jase: I asked, Dave that... We have a couple of minutes left. And there have been a few other really awesome questions that have been posed but I'd like to actually just give it over to you, dude. Like you've learned so much, you've done so much for the world. You've helped to create this wonderful contraption, the internet and to make it work better. It turns out it was a romcom. Who knew, right? It was your pursuit of love but then that worked out. What do you want folks to remember, like what do you want folks to know and what should we be thinking about?
Dave: I'm thinking about moving forward.
Jase: Yeah, you...
Dave: I'd be a real downer talking about it but there's three things that really concern me deeply and they always have. I grew up under the spectre of nuclear war. It remains possible. There are still 20,000 or more atomic bombs pointed at us in America and it's terrifying to me.
Jase: You don't believe MAD, like you don't think that actually works, dude?
Dave: I kind of gave you a couple other examples. No, I was actually a big believer in the strategy of technology by Possony and Pournelle, strategy of technology. So I didn't believe in MAD. I thought MAD was crazy. So I'm worried about that. Right now there's about 400 fibre connections that connect the whole world together, just 400 that could all go away in five minutes, Kessler event. These are things that keep me awake at night. Kessler event is a burst of debris travelling in the opposite direction in the orbits we use. In a matter of hours, all of our satellite systems could go down.
Jase: Oh dude, we should have got into this. Like the younger [0:58:05.4] ____, is that a recurring theme as we pass through the tour, it's like are we at the end of that clock?
Dave: Yeah. It may be that... There's a whole bunch of wonderful science fiction films where...
Jase: We need to bury the fibre, we need to send spaceships with the backups...
Dave: Yeah. Spaceship backups, resilient fibre, connect towns together and to each other, use solar power, decentralize.
Jase: Decentralize energy production, storage. You need really good information infrastructure to decentralize it.
Dave: Information. I know a group that's trying to put data centres in space. They're also trying to... The NASA, I hope someday will put DNS in the space. So I'm an asteroid fan. I used to give a party every February 29th, for every four years. I hate throwing parties. To thank the asteroid that hit us, that got rid of the dinosaurs and allowed mammals to thrive. And they say, "Oh, please don't hit us again until we're ready."
Jase: Are we ready?
Dave: So there's big picture and stuff, like getting the hell off the planet and becoming a diverse species, it would be awesome. Not multi-planet, get off the planet first.
Jase: Okay, moon.
Dave: Gravity sucks. Asteroids are way cooler. [laughter]
Jase: Maybe some... Asteroids are the RVs of space, you get up there and just you can...
Dave: Get on one, it's called a near-Earth object. This is another passion of mine. You get on a near-Earth object and you go on this four-year tour between here and Jupiter. And you can go from there, jump off at any given point, go anywhere else you want. You don't need rockets, you just need springs.
Jase: It's not a road trip. It's like a...
Dave: It's a road trip. It's like getting on a sailing ship back in...
Jase: All right. I dig it.
Dave: Sailing trip.
Jase: Dave, it turns out we're actually, we're out of time in this session but it feels like there's a follow-up here. And the amount of book club and inspiration that you've inspired in the chat, like with all these great folks that are in... Folks, thanks for joining. I know folks are needing to jump off but like we're gonna get a copy of this chat and put it together. Let's make it an assignment ready team. As we're working with GPT and our trainings for helping broadband providers connect families and businesses to better broadband, let's throw these books in there to make sure we do what we can to help with the archive and everything but Dave, thank you again for your time and everything that you've been doing to help us connect and make the world smaller. It means a lot even if people don't say it often enough but thanks for doing this.
Dave: Thank you for carrying the work forward.
Jase: Yeah, let's do it, dude. We're gonna get this jam band together and do some great stuff. So appreciate you. And we're gonna get this circulating to folks that need to see it. So thanks, Dave. Y'all have a wonderful rest of your Friday, okay?
Dave: Yeah, you too. See you at Consonance.
Jase: Consonance. Where is it? Where is that?
Dave: Over in San Jose. San Jose...
Jase: Okay. Okay. Yeah. All right.
Dave: Come down.
Jase: Let's get down.
Dave: Again, I'd like to thank my graphic artist for your contribution to this internet simulation.
Jase: Very nice work. [laughter] Very cool. Very relatable. Thanks, Dave.
Dave: And oh nine... I love counting nines. 7, 6, 5, 4. Main engine starts. We have lift off.