Taxonomy of Hype
I have no scheduled spaces still. So okay. Don't know what I need to do to get this to get in the the what I do have is in the topics. The so now in we used to say, like, technology would be a topic. But now you have to pick which subset of technology.
Speaker 1:Do you have do you have a bad feeling about where this is going?
Speaker 2:Kind of. I mean, I feel like this is gonna be we're gonna be immediately into crypto.
Speaker 1:It it not only are we immediately into crypto, it is like of the 25 topics that you can have a space on, like, 18 of them are different specific currency.
Speaker 2:That's exciting.
Speaker 1:It's like, oh, fun.
Speaker 2:I mean, the the the the spaces I've seen that have, like, the like, you know, 2, 5, 10000 people of them have all been Oh, no. Crypto spaces. And and have you been to the profile tab in Twitter recently?
Speaker 1:No.
Speaker 2:I should let it be a surprise. But there there's some prompts that I swiftly clicked through that I think lets me use my NFTs as my Yes. Profile or something? Yes.
Speaker 1:You could become you can become a hexagon.
Speaker 3:Good. Good.
Speaker 1:Yeah. And if you this is good. You've basically been, like, off Twitter for the last, like, I don't know, 72 hours.
Speaker 2:No. I I'm just very selective and I block a lot of keywords.
Speaker 1:You're you're very good. That's I I'm I'm impressed. You oh, is this you know what? I actually so okay. Point of principle, but maybe a bad one.
Speaker 1:I don't block anybody. Maybe I should. I don't. I do including people who hate my cuts and love to tweet about it. It's like you think I would've but I don't know.
Speaker 1:For whatever reason, I just don't do it. I what I do, I have muted a very small number of people where it's like I literally just can't take it anymore. And I went into my mute I have muted mister Peanut.
Speaker 2:Well, he does have a lot of ads as I recall.
Speaker 1:I literally wouldn't know. And I it's like, oh, I do remember just being, like, begging for mercy from mister Peanut. Like, I mister Peanut, I cannot like, we are past the point of actionability with respect to my Peanut and something, and I we need to stop bombarding you with ads. I I don't even know. Mister do you get mister Peanut ads?
Speaker 1:Are those targeting? I don't know.
Speaker 2:I mean, I I gotta go look and see if I've I've blocked or muted because unlike you, I don't have a principle around not blocking or your mind.
Speaker 1:I don't know if it's a principle. I do feel that it is very lame when you have people that are celebrities that are using block list to indicate the people they disagree with. I think that that that's normal. But I say that as someone who's, like, on those block list. So, like,
Speaker 2:me too. I, you know, I had this moment where I'm, like, well, I guess Mark Andreessen Yes. Isn't on Twitter. It's, like, no.
Speaker 4:Well, no. He's on Twitter. He he he just blocked you
Speaker 2:for reasons I have no idea.
Speaker 1:So I he's blocked me as well, obviously. And I do not know why he he he's blocked me for so long. I don't know that it's even discoverable why he's blocked me. I don't know if we can even figure it out.
Speaker 2:That's right.
Speaker 1:And I'm such a chicken, you know, because we were I I ended up pitching Andreessen on oxide, like, in a 1 on 1 conversation. And I should've been like, hey. Can I just ask you, like, a question? Like, Why do you block me? But I click, you know, and I I don't know.
Speaker 2:Do do that after after you raise the money. Like, hey, funny story. Thanks for the money.
Speaker 1:Thanks for the money. Well, yeah. That that story didn't work out that way. So, you know Exactly. Thanks for cheering me on from the sidelines.
Speaker 1:Like, can I ask you a question?
Speaker 2:So how did you come across this this blog post?
Speaker 1:The hype thing, I was just kinda going around. And I do feel and I, like, obviously, I don't wanna spend any time on the web 3 brain fungus, but clearly, the fact that I'm saying that is just
Speaker 2:Right.
Speaker 1:I feel like we are all collectively, as technologists, transfix. And I felt like Kelsey you obviously you went you go back and you listen to our predictions. But you obviously must have done that.
Speaker 2:I I I read through a bunch of them, but I did not listen. I I listened back to it, like, the night after I recorded it, but not, like, not since.
Speaker 1:I yeah. Kelsey was so that was so good. That was Oh, he's terrific. Yeah. Like, everyone is great.
Speaker 1:Like, Tom was great. Laura Laura is great.
Speaker 3:All the
Speaker 1:pictures are great. But but Kelsey on I just thought it was really, really interesting going back and relistening to it. And the other thing that I just cannot get out of my brain that I thought was a really interesting comment, but one that we didn't dwell on, is where Stephen O'Grady, likened the the movement of web 3 to the, like, the commune movement of the seventies.
Speaker 2:That is great.
Speaker 1:Yeah. And I just felt like, you know, where you had all this turmoil of the sixties. And, you know, the people kind of taking out those angst in you know, becoming very agrarian and getting kind of on getting I mean, it's like, in many ways, like, it's the exact opposite, but it kind of represents some of this anyway. So I think we are all collectively trying to reason about this thing because it's so different than anything that we have seen. And it just be different in that the level of hype is so next level.
Speaker 1:And yet the people that dig have a lot of questions that aren't with the answer. I thought Kelsey did a great job of of answering those. I did follow Kelsey into a space talking about crypto. And have you been in any of these, like, super pro crypto spaces?
Speaker 2:Yes. But, it it like, not for long and, often I find them sort of almost unintelligible. Like, there's a there's a shorthand and a language in some of these that I find indecipherable.
Speaker 1:And it is I agree with that. Indecipherable, and the indecipherable, and that's, like, not praise. Just to be clear. Indecipherable in the in the pejorative sense. But Kelsey was really trying to, like, ask some really good questions that are tough questions, but he's asking in a really, like, friendly way.
Speaker 1:But
Speaker 3:they were, not being answered
Speaker 1:very well. And I'm, like, DMing me being like, that is the last time I follow you at your space. I'm like, look. I'm sorry. I I it it be yeah.
Speaker 1:That was that was a mess. And then several like, maybe a couple of hours later, Nahum sent me a link to a space that Kelsey had started on, like, I've changed my mind on cryptocurrency. I'm like, oh my god. What happened to Kelsey?
Speaker 2:Oh, no. Yeah.
Speaker 1:And I was like, when I laugh last left Kelsey, he was, like, trying to make rational arguments. It was just being, like, it just making no progress. And he, like, something, like, collapsed in, like, the the the in the 3 hours. It's like or whatever. And so, did you see the space item?
Speaker 1:And I would No. No. I didn't. It was just a rec roll. It was really great.
Speaker 1:Oh, thank goodness. Yeah. It It it it it was it was yeah. Exactly. Right.
Speaker 1:It's a call. Sanity is restored. And yeah. So,
Speaker 2:yeah, that this is gonna be, like, we're gonna need to stage an intervention or something. Yeah.
Speaker 3:We're
Speaker 1:gonna see and I think that so I feel this so this blog entry is like a a kind of taxonomizing hype. Something that I saw yesterday that been kind of kicked around earlier. And I I feel that this is again what a lot of us are trying to do of just trying to make sense of this thing and trying to systematize it. So I thought this was a really interesting I I mean, I thought this was an interesting way to think about it. And
Speaker 2:Yeah. Yeah. I I I really liked it when you kicked it over. And and obviously this, you know, it sort of unfortunately, it starts the graphically starts at at, the most severe, which it, so it didn't it didn't kinda slow roll me into it. But as soon as I saw the term othering, I I I could not help but associate that with Web 3 and crypto.
Speaker 2:Right? The this notion that that not only will this lead us to utopia, not only is this imbued with magical properties that that make no sense, But to think otherwise is illogical.
Speaker 1:I'm gonna regret do you are you familiar with Roko's Basilisk?
Speaker 2:No. Oh, no.
Speaker 1:I I was I was afraid that was gonna be the answer because I know. So Roko's Basilisk is the idea that the artificial general intelligence is coming. And that the first thing that it will do when it gets post singularity is it will simulate all of past humanity to determine who accelerated its arrival and who impeded it. And it will reward those that aided the the the AGI and its arrival, and it will punish those who did not.
Speaker 2:Makes sense.
Speaker 1:Exactly. Makes sense. So the the, talk about, like, your your other ring. And we do get to the point where, like, okay, we are definitely talking about something. We we are no longer in the realm technology here.
Speaker 1:We're in we're we're theological.
Speaker 2:Yes. Yes. Absolutely.
Speaker 1:We're talking about the rapture. But I think it is worth talking about, like, past so I I I'm hoping that we can be done with web 3rd moment. I know. I'm sorry that I
Speaker 3:No. No.
Speaker 2:I think I think we got it out of our system at least for the At least
Speaker 1:for the moment. I know. Brain fungus. But I wanted to go back through because, obviously, we we do see this a lot, and I like this way of it kind of tac taconomizing it. Starting with certainly marketing claims, which an exaggerated returns at level 1 and level 2, which we've seen a lot of.
Speaker 1:I feel Ethiopian Futures, we've seen more than 1. And then magical thinking, I can think of at least one example, but I'm gonna be very curious. If you've got other examples that that that you're thinking of. Yeah. I would
Speaker 2:say as I've been, you know, as you text DM me this earlier today, and I still think about it, I I am a little worried that this is, like, a massive troll of yourself. Like, we're going to be be pulling up these things that you have ranted against for years.
Speaker 1:Like, some sort of, like, super rant.
Speaker 2:That's right.
Speaker 1:Well, no. Okay. So, actually, then maybe we should start with this. Like, let's start with not the things that did live up to the hype, but the things that were hyped and if anything under hyped. You know, what were the things that, like, people were enthusiastic about that have over delivered on that enthusiasm.
Speaker 1:And again, I've got a couple of those in mind. Maybe you got a couple of those in mind too.
Speaker 2:No. I'd love to hear it.
Speaker 1:Oh, I well, I'm for sure, because I think cloud computing is a problem.
Speaker 2:Oh, yeah. Yeah. Yeah. That's a great point. That's a great point.
Speaker 1:And there were certainly people who believed at the time that cloud computing was being overhyped. And I don't think it really was. I think it's been it's if anything, it is a much more profound impact on computing than we thought it would be.
Speaker 3:Yeah. I I
Speaker 2:think that's a great one. Matt, I don't know if you were trying to get in there.
Speaker 5:I mean, you know, just to start off with, like, are are things okay out in San Francisco? Like, are they
Speaker 3:putting the No.
Speaker 1:No. No. No. No. Things are very bad.
Speaker 1:This is required for help. Like Yeah. No.
Speaker 5:It Like like, there there are Bitcoins falling from the sky.
Speaker 2:It's funny it's funny you say that in
Speaker 5:the army of smart contracts. You know?
Speaker 2:Matt, I I literally was in San Francisco, yesterday and drove by in a poster that promised a million in Bitcoin. I don't have the details, but I took a picture. But, I'll and I'll I'll send it around.
Speaker 1:It it it is Right. And if you drive right now through downtown San Francisco, it is dis disconcertingly on brand. The billboards in with these are billboards that used to be on the 101 back in the day at Silicon Valley. Have all moved north, and they're all the app center of of San Francisco. And there are I don't know if half are crypto related, but a lot of crypto related.
Speaker 1:Yeah. So on the underwrite, I think, I think, arguably, mobile was underwrite.
Speaker 2:Yeah. I was gonna I I I was wondering what you thought about kinda we talked about iPhone and our and our and your perhaps prescient
Speaker 1:iPhone prediction
Speaker 2:from years ago. But, yeah, I think I think that kinda snuck up on us to a degree.
Speaker 1:I mean, I mean, certainly,
Speaker 2:the the I mean, the the fact that the fact that it is what it is now, I think nobody predicted or those who did predict it were dismissed as as, as overhanging.
Speaker 1:Another one that maybe was, like, not necessarily overhyped, maybe it's, like, adequately hyped, Wi Fi. Do you remember you were the first person I knew who used Wi Fi.
Speaker 2:Well, that's what I was gonna say. I I feel like I I worshiped at the WiFi altar almost instantaneously.
Speaker 1:Yes. You definitely did. And I remember just being like, this is, like, this is crack cocaine and stuff. This is amazing. Like, look, Anoop bite my hands.
Speaker 1:They're all around my laptop. Top. There is no wire. Yeah.
Speaker 6:Yeah. Yeah. Yeah. I remember
Speaker 2:I got that first titanium power book or whatever the fuck it was. And I was very, very excited about being able to sit outdoors and type. I don't know why that was so fascinating to me, but, it it was everything.
Speaker 1:It was amazing and I don't know I don't know why my brain conflates these 2, but this was also about the same area era in which you broke all your hands.
Speaker 2:That was a little bit later, and it was only one hand that I broke. Actually, it was just one metacarpal, but yes.
Speaker 1:Are you sure you didn't break everything? I just remember you only being able to type with one finger from each hand. Is that not correct?
Speaker 2:It is. It so it's almost correct. So, I broke, 4th metacarpal in my right hand. So my whole my whole right hand, like, there I had 3 pins in in my hand, and I could basically type with one finger on that hand. And my left hand was fine.
Speaker 2:I would note we were building DTrace at the time, and you can type D trace entirely with your left.
Speaker 1:That's right, it was originally called JK LIO.
Speaker 2:That's right but in to accommodate me, we're like, what do we do? Only the left hand.
Speaker 1:Jakoya has to change its name. Yep.
Speaker 2:And I I know you you refer to it as robot hand because it looked like I was being assisted by a robot on, like, a pick and place machine
Speaker 3:on the right side.
Speaker 1:A meat stack pick and place. World's worst but most entertaining pick and place machine. We I did used to what like to watch you type. It was it was I mean, on the one hand, it was personal
Speaker 2:That's morbid.
Speaker 1:Yes. Yeah. More like that. I I knew I was I was ashamed for not being ashamed of watching you watching you type. But so I I I order business that with Wi Fi, and that the
Speaker 2:yeah. Alright.
Speaker 1:And I don't even recall ever being like I don't know if Ygrene said it was Ygrene hype, but Ygrene was like a really big deal. That was a really deep transition. The, okay. Where does Java land?
Speaker 2:Yeah. Right. I mean, it it it certainly felt crazy overhyped at the time. Crazy crazy over like, right once run anywhere. You know, swing UIs were somehow going to not be the ugliest thing ever, but going to be pervasive and fine.
Speaker 2:I
Speaker 1:mean, I think we got into level 3 on Java and at least Java on Silicon is arguably level 4.
Speaker 3:Yes.
Speaker 1:Yes. Where
Speaker 7:I mean, write write once, run anywhere has kind of one, and your eyes that don't look like the native host system have also won with Electron. So it's was kinda like right idea wrong implement Well, and
Speaker 1:I think that this is what kinda I I asked about.
Speaker 3:I mean,
Speaker 1:because it's like Java also, like, definitely succeeded in many different angles and, like, you know, I think it I don't know.
Speaker 2:Yeah. Absolute I mean, say what you will about Java. You can like it or hate it as a language, but the fact I mean, it's gotta be, you know, in the top 3 or 5 of the most, you know, most lines of code per year or have everyone that measure it. I mean, it
Speaker 5:It it also runs in a really just unconscionably large number of places. Yeah. Including somehow SIM cards all running Java.
Speaker 1:So but I think but but the hype level of Java was was, exaggerated for sure. And we're hitting some utopian futures. But it also ultimately, lived up to at least some some aspects of it, if not, maybe even many or most.
Speaker 2:Yeah. Certain certainly some piece of it. I was wondering what you thought about Windows NT. Oh. You know, something that we've talked about on the show, you know, read a lot about.
Speaker 2:But both of our sort of college, high school, early adulthood was was tainted with the the specter that, at least the utopian future Yeah.
Speaker 1:It's I of Windows 3. I was just gonna point out that level 3 is supposed to be Utopian features futures, not dystopian futures. I don't even think, like, Microsoft is pretend pretending that it was the utopia future. Microsoft is just like, no. Our dystopia is inevitable.
Speaker 1:I I just don't.
Speaker 2:That's right. Just we we all need to come to terms with it. Like, it's gonna our terrible future is gonna win. Yeah.
Speaker 1:And I don't know how much of it was, like, hype versus dread, honestly. I mean, I did people are just kind of acceptance.
Speaker 2:Of the inevitability of it.
Speaker 1:Of the inevitability of it. So I get a a couple of storage related ones to ask you. Dedue. Where does dedupe deduplication, where does that come out in all this? Yeah.
Speaker 1:I don't know. I mean, so how I
Speaker 2:mean, that feels more like an exaggerated returns, I think.
Speaker 1:I right. I think so too. Yeah. Yeah. Yeah.
Speaker 1:Yeah. Because because I think
Speaker 2:it was pretty mildly hyped, and I feel like kind of as a busted flush right now. Like, I I don't think like, Dedoop doesn't really feel that real to to me right now.
Speaker 5:Do do you actually know anybody who runs DDoop anymore?
Speaker 1:I don't feel I know anyone who runs, like, a SAN nerd anymore. Yeah. I'm not I'm not the very first to have.
Speaker 2:I'm not gonna hang with the right kind of nerds anymore. Yeah. It's like it it's clearly important for some things. Like, I I think that folks, like, if you're if you're, hosting a bunch of Windows VMs, that's still important.
Speaker 5:But, like, disk is cheap.
Speaker 1:Yeah. Agreed.
Speaker 3:It is.
Speaker 5:You know, it's just so comically cheap that, like right? I mean, a a petabyte is, like, 50,000 Well,
Speaker 1:it's or maybe the kind Yeah. Right.
Speaker 2:The kind of things that were most effective are are becoming diminishingly small parts of the what you store. That's right.
Speaker 1:And it's like, well but if out of curiosity, like, if I have smarter software systems or certainly in a world where a lot of that looks compressed or encrypted or looks more like a like signal or looks like images or looks like video. It's like well, all that stuff is not right.
Speaker 5:Or or or or in a world where you have smart snapshotting on your VMs and or and or datasets, like like ZFS and dedupe don't really make sense as a We're sorry.
Speaker 1:We've actually put a, a dialogue box to to to try to attempt people to to not actually enable it. That was a a classic example of something that was done a little too elegantly. That was done in in in insufficient number of lines of code.
Speaker 5:Yeah. I I also have to ask since you guys were there, the ZFS copies attribute. Mhmm.
Speaker 2:Copies. Yeah.
Speaker 1:Yes. Go on. Ask your question.
Speaker 5:Just I mean, I have many questions.
Speaker 1:Adam, do you recall the wager that we had in the office?
Speaker 2:No. I don't. Did I win?
Speaker 1:Something like that. Maybe maybe for purposes of this story, you know. I can't. No. No.
Speaker 1:No. Okay. So I well, then you're not gonna be able to check me if I get this wrong. But so we had analytics coming from our storage product telling us which features were used when. And what we felt that end copies setting copies to it so copies equals a number more than 1, creates effectively spurious copies of your metadata.
Speaker 1:I mean, am I I don't feel I'm pitching at this very well, but I don't I I I don't there's a great way to pitch it.
Speaker 5:Yeah. I think that's about right.
Speaker 2:I think I think it was of your data. Right? Because end copies of, because we actually, as I recall already implicitly had end copies equals 2 on some internal meta
Speaker 1:Yeah. I thought it was and it we it was taking spurious copies purely of metadata, but it was copies above and beyond what you ever you're doing with the Brazy 2 of your
Speaker 2:Yes. Yes. Absolutely. It was it was it was above and beyond the act the, like,
Speaker 1:was we had, IB as a a target. And what were the 2 ISER and what was the other IB? ISER and ISER is ISER. And what was the other ISCSI that that that is ISER stands for ISSCSI extensions for RDMA. So literally, 2 of the letters in in that acronym stand for other acronyms.
Speaker 1:And, actually, because one of them is ice, actually, this is the that that letter expands twice.
Speaker 2:Yeah.
Speaker 1:And then I cannot remember the other rival IB standard. But the the the question that we had
Speaker 6:I'm sorry.
Speaker 1:SRP. SRP.
Speaker 2:Oh, for the word. Is.
Speaker 1:That is. It's SRP. And we put I think we had SRP in the product if I recall correctly. We thought that both of these things were not well informed. And so the question that we had was and we knew that 0 customers had turned on either of these things.
Speaker 1:That 0 customers that were connected to us, and that's basically all the all of our customers. We had we had really good data on what our customers were doing, which is really
Speaker 2:bell, but I don't remember
Speaker 1:So we if I recall correctly, Shrock and I had a bet about which one was gonna be used first. And I can't remember who's on which side of the bet, and it doesn't usually matter because both numbers went from 0 to 1 on the same day. And, like, has the same customer are we being pranked by a single customer who is turning on both of these questionable features? And as a result, they were, like, just different customers that were turning different single, But yes. So no.
Speaker 1:I I don't know. I don't think it was hyped because I don't think people knew about it. But, below InfiniBand. Where does that come out on the on on Kai? InfiniBand was very hyped in its
Speaker 2:era. Yeah. I I feel like my my lens on InfiniBand was so tainted inside of Sun because it felt like such a big deal inside of Sun, and I don't really know how how big of a deal it was outside.
Speaker 1:I feel, Tom, I feel it was it was pretty people Intel was certainly making a big deal out of it.
Speaker 6:Yeah. It it was overhyped. It's it's firmly entrenched in the a HPC world.
Speaker 1:Firmly entrenched in the HPC world, which is kinda where it landed. And so it's hard to know whether it's like, well, did this thing, like I I think it is fair to say
Speaker 5:Well, I mean, it's also tough to pitch InfiniBand in RDMA if your architecture is microservices. Right? If your architecture is giant shared memory compute, InfiniBand starts making much more sense. But if if everyone's throwing rest APIs at each other with reckless abandon, InfiniBand doesn't really do much of anything for you.
Speaker 1:It doesn't. And then I I would I would add to that, it's a lot of, there's a lot of work, a lot of yak shaving to set this thing up. And that the, people who don't value their labor, which is, say, National Labs, seem to have a stronger affinity to InfiniBand than commercial customers who actually, like, listen. We have a job to do. We're actually just trying to get this thing to kinda do do stuff for us.
Speaker 3:Well
Speaker 6:I I mean, I I forget the category of a system area network where all the endpoints have to be very homogeneous for anything that makes sense.
Speaker 1:8.
Speaker 5:Well and, also, I mean, for a big enterprise network, you assume that you have heterogeneous machines continuously being upgraded. The machines may be diverse. They're definitely running diverse workloads. If you're building a supercomputer and every node is identical and you care about being able to treat this thing as one giant shared memory machine, so your software people don't, like, completely have their heads catch fire because TCP is not reliable. Or Yeah.
Speaker 1:Those are the conditions in order to make sense. Those were not the conditions under which it would it it definitely was at the level 1 or level 2, at least, of of hype, I would say. I don't think I don't think InfiniBand ever rose to Utopian futures, but I definitely think it had marketing claims that exaggerated returns for sure.
Speaker 2:Well, it it for for another Intel technology.
Speaker 1:Oh, yes. Which one are you which one on my list are you picking?
Speaker 2:I I know. I know. I know. I know. I I'm I was wondering if you could guess, but I'm gonna go
Speaker 1:first with Optane? 3 d cross Optane. Yes. Nice. Yeah.
Speaker 2:Where where where do you think that reached? Because I feel like it it it kind of, blew a DUI on the, on the
Speaker 1:high school. Absolutely. Definitely. Yes. What no.
Speaker 1:It it it definitely was at a level where it's not even driving for sure. It was at a hype level where it should have gotten a designated driver. Yeah. No. I think it was getting, I mean, up there.
Speaker 1:Certainly, up to Utopian Futures. I'm not sure quite the magical thinking. But I Optane and it is, you know, it's kind of a sad saga to watch unfold because it was so heady. And then every time you kinda ask for an update for it, it was pretty clear that it was just not not fulfilling the the it's certainly not fulfilling, it the great ambitions of Intel with respect to and and we don't even know, like, how it actually works. Intel refuses to say that this is phase change memory even though it clearly is.
Speaker 2:Well, for the longest time, they couldn't make it.
Speaker 1:Yes.
Speaker 2:Right? Like, it it was not available in any
Speaker 4:kind of reasonable quantity.
Speaker 5:I mean, if you look at the pricing today, I'm pretty sure they still can't make it. And if you look at the performance, I believe what everybody thinks they're doing is they're running a DRAM cache in front of an SSD, and they've just built a really nice integrated controller that sits in a DIMM slot.
Speaker 1:Well, and it is interesting how they started off being like, we are gonna be faster than DRAM and persistent. It's like, okay. Great. That's exciting. It's like, we're gonna sit between DRAM and Flash.
Speaker 1:Like, we are now faster Flash. Like, wait a minute. I thought, wait. What what just happened? It's like
Speaker 2:It's like, except, you know, the economics don't make sense
Speaker 1:for me.
Speaker 2:So as long as you have a problem that needs slightly faster SSD and you're willing to pay a 4 x premium.
Speaker 1:But this actually does remind you something I did not have written down, the memristor and the machine from HP. Oh, boy. I mean, that's got to be I mean, we've all kind of done HP the dignity of just let's just forget that ever happened. What the hell happened to that thing? I mean, that was definitely did you see the ad for that?
Speaker 5:I I looked this up the other day because I started wondering whatever happened to this thing too. Apparently, they they just said, you know, turns out building a big new architecture from scratch is kinda hard, so, like, we're just gonna send it to the printer today. And they might be able to use memristors. So, like, they're they're they're
Speaker 3:they're they're they're they're
Speaker 6:they're they're they're they're enterprise
Speaker 5:copier out there somewhere with a memristor.
Speaker 1:Yeah. I think Tom, so you you you think it's level 4 of the machine. I think it might be level
Speaker 3:4 too.
Speaker 1:Adam, did
Speaker 6:They were clearly deluding themselves in a big
Speaker 1:big way. Did you see the ad for the machine, Adam?
Speaker 2:Oh, no. I know.
Speaker 1:This is what I gotta look
Speaker 6:this up.
Speaker 1:If you are wondering, like, wow, I don't know. Like, level 4, like, that's a magical thinking. That's a pretty outrageous claim. It's, like, alright. Well, just like, that's fine.
Speaker 1:You sit through the ad, and then you tell me that there will be, like, exaggerated returns or marketing claims. It's a it they're all wearing, like, Star Trek costumes. A hailing yeah. Yeah. It's, it's it's it's bad.
Speaker 1:And I think the is you gotta kinda put in the kind of I mean, the machine, I think, arises the magical big thing. The memristor is
Speaker 5:Oh, that thing is just I mean, that it's hard to determine the hype level of something that never actually existed. Because if it if it's never existed, it is sort of by definition magical.
Speaker 1:I also think that there
Speaker 5:Like, I I find it very hard to hype rank things that never, like, actually ship. Like, Optane, we can compare results against theory. But, like, when you defund the project before you ship your first unit, like, it it's just kinda like a one over 0. We're we're
Speaker 1:we're we're NAND on the hype scale. Yeah. There needs to be a NAND over here, for the level NAND. So okay. That's an interesting point.
Speaker 1:I do feel that, like, one of the things that you see a lot in these things that get way overhyped is not providing technical detail. Like, I'm gonna provide you the emotion of a revolution, but not really the technical detail to support it. Not yet. Like, it's coming. Or it's a secret, or you have to sign an NTA.
Speaker 1:I feel like that's a common trend among these things that got, like, just
Speaker 2:Okay. I mean, case in point there has gotta be the results of the recent trial. Right? I I I mean, like like, I feel like you're talking
Speaker 1:about So you know what? I am actually not talking. I I mean, obviously, Theranos for sure. Yes. Yes.
Speaker 1:But, actually, what I have in mind is something that I think absolutely hit level 4, and that's the segue.
Speaker 2:That's a really interesting one. Because because you're right. That that that there was, like, this brief feeling that of ubiquity.
Speaker 1:A and of not well, so do you remember? Okay.
Speaker 5:Hang up. Full the segue is a level 4.
Speaker 1:God. Okay. Are you like, don't take this the wrong way, but where were you in 2,000?
Speaker 5:I was not in San Francisco by any stretch of the imagination. So, like, I feel like I feel like you guys just
Speaker 1:have to, like, upgrade. No. No. No. No.
Speaker 1:This is not a San Francisco thing.
Speaker 5:For the rest
Speaker 3:of the world. This is
Speaker 1:not a San Francisco thing. This is like so this is an this is a mesmerizing story. There is a great and we'll link to it. There's a great podcast, called Dakota Ring. Dakota Ring has got an episode on Who Killed the Segway.
Speaker 1:It's part of an article about the, that talking about the Segway. No. So what happened on the Segway is Segway is invented by team Kamen who's amazing, a really interesting technologist. Also, couple screws loose. I I don't I don't think that's I'm not saying that pejoratively, but it's got some wild ideas.
Speaker 1:Dean Kamen has this idea for what what becomes the segue, but also is paranoid that it is gonna be taken by someone else. So they have absolute secrecy. And what they do is they signed a a someone's writing a book about the Segway. And they were pitching this book proposal, but they would not give any details about what it actually was. And then they that book proposal got leaked.
Speaker 1:And there was a very brief period of time. This is not a Silicon Valley thing. This is like the whole country went cuckoo for what is this thing. This is like any I think they're like putting modern terms. This is like I don't know.
Speaker 1:Maybe Elon Musk or I don't know. What what would be the example of where someone today would be like, I've got an idea. And you got Steve I remember Jobs saying that this is, more important than the bicycle. I believe it, it was Jeff Bezos saying or maybe it was it was Bob Metcalf saying it was gonna be bigger than the Internet. So, like, Matt, this is not just like a and inside the 101 kind of thing.
Speaker 1:This was this went local. And then there and there's a great moment, which I had actually read about but not heard until listening to this podcast, where, Diane Sawyer so they go to announce this thing on, like, the Today program or whatever on ABC. And they do this unveiling, and it's the segue. And Diane Sawyer, who they really should have briefed on what it was before because she could not she's like, is that it? Like, that's not it.
Speaker 1:That's that that no. Like, she's literally like, that's not
Speaker 2:what we're waiting for.
Speaker 1:Like, bring out the thing that we're that the entire nation is waiting for. So, no, I'm not making this up. I know I sound like I I know I sound like a this sounds like a vivid dream, but this actually happened.
Speaker 2:Okay, Brian. In terms of other, I think, magical thinking, maybe YouTube Utopian future, I was thinking back to to my childhood and and tell me if this rings about because there was a period in, like, the early nineties when, like, we thought that maglev trains were gonna be everywhere and that, like, power was gonna be free. And, all of this came, I think, because of these either false or falsified conclusions about cold fusion Yeah. In, like, 1989, I believe.
Speaker 1:I I don't know how much that's the Utah group, the cold fusion group. The, and I don't know how much of that that but there's definitely a lot of enthusiasm. And there still is a lot of enthusiasm around nuclear fusion. But, yes, it was definitely fee for Petro. I feel the seventies and eighties.
Speaker 1:I mean, other folks can can kinda chime in here. But, yeah. Infinite free power was definitely something we were gonna have.
Speaker 2:That's right. And then and then, you know, the magical thinking surrounding that and, like, all of the things that it would enable that it, you know, never could.
Speaker 1:Yeah.
Speaker 8:I mean, yeah.
Speaker 6:Cold cold cold fusion was a specific hype hype thing for
Speaker 4:a few years, but, you know, just turned out
Speaker 6:to be bad science, I think. But but every time there's some energy breakthrough, you get this utopia cycle.
Speaker 1:Yeah. When I think it's it it's kind of interesting because, you know, Adam, I'm I know you read The Code Breaker recently, and I'm I'm reading it now. Interesting book. It's certainly the underlying science is super interesting of of CRISPR. I kinda feel like the human genome project is a little bit in the overhyped category.
Speaker 1:I mean, I would love I love it. I market biologist weighed on that, but it seems like that's a that got a bit overhyped.
Speaker 2:Yeah. Certainly from from the code breaker book, it it seems like it was a huge expense, a lot of hype, certainly exaggerated returns in that I don't think it turned out to be particularly useful or perhaps some of the technology that spun out of it was useful, but the, you know, concretely, it it was, you know, quickly deprecated, that dataset.
Speaker 1:And I think it also it highlights the huge project highlights something that happens in this kind of level 3 of Utopian Futures where people begin to think, oh, with this technology, I can do this. And something that, like, clearly the technology cannot do. Like, and people thought that the that the Human Genome Project would allow us to do all sorts of things. That's like, no. No.
Speaker 1:No. We're not all all we are doing is disassembling all of the instructions of the program. We have no idea how the program works, and that's an exaggeration, obviously. But there's there's a lot that we just by sequencing all the nucleotide base pairs pairs does not mean that we've actually cracked the code on all this stuff. But people were clearly, like, drawing that inference.
Speaker 1:And it's kind of our responsibility as technologists to kind of reel people back in at that moment. And it can be tough, actually, when I mean, Adam, surely, you have been with someone talking about something that DTrace can do that it cannot, in fact, do.
Speaker 2:You mean they're No.
Speaker 3:No. You have
Speaker 1:to do that. They're saying that. Right? Yeah. Right.
Speaker 1:No. I mean, they're saying I think probably probably both of them. Exactly. Right. But did I push Right.
Speaker 1:Well, that and that depends if you are merely in level 2 of exaggerated claims, that would be you, or level 3. That's them. But, I mean, have you you have been around people who are, like, saying, oh, I I can do this with this. And you're, like, you can't actually.
Speaker 2:Yes. Absolutely. I mean, absolutely. You hear technology pushed into areas where it's being misapplied or or people are drawing the wrong inferences. That that one of the questions I had was how can we tell when we're part of Yes.
Speaker 2:Hype cycle and magical thinking? You know, when you're you know, if I'm in my if I've got my 3,000 purse person space where I am talking about, you know, Huddl and and how it's going to the moon, How can I hear myself there? And and and and how can we identify like, what are the shibboleths of of these hype cycles and and identity Yeah.
Speaker 1:How do you know you're in 1? Sorry, Todd. Go ahead. David, try it again here. We're
Speaker 8:we're we're talking about several different types of hype cycles here. Right? I mean, like, you're talking about technology hype cycles and also the Human Genome Project. I guess, on on your comments on the Human Genome Project, I would point out that's that's funded science. Right?
Speaker 8:So, like, yeah. Sure. It it just maps the human genome, but that was pretty novel at the time. And a lot of these things these other things you're talking about wouldn't be here without that project. And I guess I I mean, to some extent, they have to hype it to get it funded by congress.
Speaker 8:Right?
Speaker 1:Yeah. That's a good point.
Speaker 3:Don't do
Speaker 1:Yeah. No. This is a good point. This is a good point.
Speaker 8:If you don't do that, it never happens, and
Speaker 1:and none of this
Speaker 5:happens. Right?
Speaker 8:So so that is part of the hype side.
Speaker 3:Yeah. No. I yeah. I don't know.
Speaker 5:And I'll I'll I'll throw up for a second counterpoint in there. Right? Human genome project itself did not deliver anywhere near what it promised, but it turns out gene sequencing is kinda useful.
Speaker 1:Totally. It totally and, Todd, you're making a very good point about the that there's a degree to which you have to point people to a distant future to get them excited about the sacrifices they need to make in the present. Am I phrasing that? I'm trying to phrase that in the the most charitable possible way.
Speaker 8:Yeah. And and you have to keep the I mean, you you have to motivate congress to the point where they allocate enough money for now so they keep allocating money over time. Because if you told them the full cost of the thing that you're proposing, they would never fund it.
Speaker 1:Right.
Speaker 8:And and it is a very delicate balance. But, I mean, I don't I work at a laboratory where we are working on fusion, and we've had some gains recently, and we've had some pipe cycles in the past that didn't turn out that great. And, I mean, it's still going lower key than it was in the past. And I think there's there could be some big gains. I don't think they'll be as big as, you know, what what it was hyped initially, but I think it'll be pretty cool.
Speaker 1:Let's go oh, wait. So tell me about that. That's what you're
Speaker 3:telling me.
Speaker 9:Wait. Thorium is not gonna save us all?
Speaker 1:That's it. What but but but to tell you about that because I think that I I one thing I I think you do see is for people that survive one of these big hype booms and busts, they are much more sensitive about not making exaggerated claims. Is that true where you are in terms of the fusion space?
Speaker 8:I think it yeah. I mean, there was so, for people who don't know, I work at Lawrence Livermore. We had NIF, the biggest laser in in the world. It's the size of 3 football fields. It has a under 92 beams.
Speaker 8:It fires at something the size of a BB and implodes it. It and it's a little piece of deuterium. There have been different I mean, this has been going since, what, the late nineties or something. And and like I said, you know, they've had to keep the budget going. They finally got ignition this year.
Speaker 8:It was, you know, less dramatic than than we thought it was gonna be. I mean, basically, what that means is they they got energy out of the capsule which is cool.
Speaker 1:That's very cool. Yeah. Wow.
Speaker 8:Yeah. So and that's a big deal because now it's in a controlled environment. They can experiment with it. More things can happen. I I think there was definitely, like you know, when I first started the lab maybe 13 years ago, there was a lot of hype around NIF for energy, and and, like, the life program, if people ever heard about that, where, you know, they were gonna build reactors out of this thing.
Speaker 8:There were reactor designs that came out of it and stuff like that. There was sort of a timeline. There was a deadline for ignition. Everybody was hyping that shortly after the machine came online. Didn't hit that.
Speaker 8:And then after that, I think, you know, there there was sort of a real negative reaction and kind of a backlash, to a lot of the hype that was coming out around energy. And, basically, what happened is that the focus of the machine, it sort of has three purposes. 1 is sort of astrophysics research, 1 is energy, and the other is weapons research. It got pulled back into weapons research after that. And the interesting thing about that is that that focus actually helped because those are the people who know how to make fusion happen.
Speaker 3:So
Speaker 8:because there's one place where it does work.
Speaker 1:Yeah. Right. Right.
Speaker 2:One place
Speaker 1:where we actually have demonstrated it. Yeah. Right. Right.
Speaker 8:And so I think, you know, ultimately, that was helpful, but, you know, there's, I don't know. Could I I don't know how to evaluate all of that. I mean, I think it was certainly out of control at some point or another. But on the other hand, did it get funding? Yeah.
Speaker 8:Did it keep the thing going? Did did it help? I I think ultimately it'll be a valuable thing, and it's a lot less expensive than some other things that we do. So I I don't know. I think ultimately, you and it's totally valid.
Speaker 8:Real apps don't look like LINPACK. They don't just do flops all the time. On the other hand, LINPACK boils down to one number that you can tell and you can compare to China and Russia. And you can say, our number is smaller than theirs. Therefore, we need more money.
Speaker 8:Right. And that helps keep things funded. Right? So, like, there's there's a huge balance there that has to be struck. So I, you know, I I can see both sides of it.
Speaker 1:Well, totally. And I, you know, it is this challenge to really sell people on a vision for the future, get them to endure with you when it's stumbling, and at the same time, not succumb to getting things overhyped. It's like it's a really hard balance to to strike. And I think that, like I don't know that fusion ever I think that there was a time when fusion felt inevitable, to me anyway, as a you know, as a kid moron kid. It felt like there it just felt like the progress of things.
Speaker 1:I don't know if you're talking about, like, maglev trades. It felt like that was going to be the the what progress would yield. And then there was a period of time where it's like, okay. Actually, maybe that's never gonna happen. But it now it feels like, well, if it could happen, it would be an untaught.
Speaker 1:I imagine this is the the the the tenor at the lab that, like, boy, if this could happen, this would be humanity changing. But it's really hard.
Speaker 6:I think I think every generation goes through a cycle of fusion utopia.
Speaker 3:I
Speaker 6:mean, I think I've seen about 3 of them now.
Speaker 1:And, Tom, are you using fusion utopia as a metaphor there, or do you mean actual fusion utopia?
Speaker 6:Actual actual fusion utopia.
Speaker 1:Is that fair enough? Okay. Yeah. Yeah. Well, you
Speaker 9:asked about signs that something's getting hyped.
Speaker 1:Yeah.
Speaker 9:I think it's usually a bad sign when people start, you know, naming their companies after the hype cycle. So some, like, blockchain iced tea, that's a bad sign. So I'm curious what all you think about say the Rust programming language in a company that may
Speaker 7:or may not be named
Speaker 3:for it.
Speaker 1:Oh, that would that would just that that would just disgust me, honestly. That would be so ridiculously Pandering. That the pandering. Absolutely. That company would be doomed.
Speaker 1:I I I find that question, like, nauseating. I can't let alone the answer.
Speaker 6:I I just
Speaker 2:To jump on that kind of hype cycle.
Speaker 1:Oh my god. How ridiculous. No. Absolutely. No.
Speaker 1:This is this is fair. I would like to believe that we are but a tip of the hat. I actually I watched Rust is kind of interesting.
Speaker 9:You would say there are no Rust Utopians?
Speaker 1:There are no well, so I actually I actually don't think that there are. So this is gonna be okay. I'm gonna, like, maybe I will I'm gonna prove myself to be a case study of how you don't realize when you're inside of a hype cycle. But I feel that Rust is actually very boots on ground. And the reason that I've actually been surprised at how just how good Rust has been for just how many different kinds of jobs.
Speaker 1:And part of the reason for that is that Rust is very reverential of history. So there are a lot of things that it it has learned from the things that have come before it. And I feel a lot of these things that have failed, they haven't learned from the things that came before it. And, you know, Adam, to your kind of your question. I feel like when things are not looking at what came before it to really understand them.
Speaker 1:I mean, there's a you don't wanna be, like, incarcerated by the past. But you also, like, when you're totally oblivious of it, it is problematic. And I felt like we've seen this in a couple of different times where people have been absolutely oblivious. And then they just repeat the same mistakes. So and I don't see I mean, I do not think that Rust is overhyped, for whatever it's worth.
Speaker 1:Silence. I'm with you. I was just trying
Speaker 2:to think about, you know, is that is that what we would say if we were part of the hyperlink? Or if we were part of the problem? And certainly, like, you go you're if you go to the Rust blog or the Rust language guide or any of these things, they won't tell you it's the right answer for all problems for all people. Although, you know, plenty of it's adherence will. So, you know, just because the main although you you go to the Web 3 Foundation website, which I did for the first time today.
Speaker 2:Oh, will. You you you sorry. I just wasted, like, 6 hours of your life now.
Speaker 3:But, yeah, I mean, it it it
Speaker 2:there is a a certain humility there.
Speaker 7:Right. The the company, I mean, phenomenon is mostly to take advantage of less sophisticated diagnosis. Right? Whereas, there would be an argument to be made that you had trouble getting money from sophisticated investors. So, like, I don't know if the oxide naming scheme really was
Speaker 2:a ploy to get
Speaker 7:get money. You know?
Speaker 1:I think we can say with absolute confidence that there are I that it was not done really, certainly not our institutional investor.
Speaker 2:Or or or how about this? It was if it was successful, it was only mildly so?
Speaker 1:That's right.
Speaker 6:That's right.
Speaker 2:Brent, I, I, you know, I had another question about, about a hype cycle that you are part of. And I'm, I'm just gonna give a little bit of history because, so this is about DTrace and DTrace obviously achieved, you know, I'm very proud of what we did with D GTrace. But back back to when I joined Sun in 2001, what I heard from, you know, I think you were underway by the point, you know, building, actually building DTrace. But what I heard was that DTrace first existed as pages in yours and Mike's notebook for at least a few years and that there was a little bit of a hype cycle at least in the serverless kernel group where problems
Speaker 1:would come up
Speaker 2:And say, oh, well, you know, DTrace is gonna solve that.
Speaker 1:Yes. So so That is true.
Speaker 2:What what level do you think it it achieved? Like, was it to the point of of othering, you know, the debuggers that that might have existed?
Speaker 1:I I I I like the fact that you're giving me the choices you're giving me me are level 5 or level 5 or a different kind of level 5. I mean level 5 or level 4 like what do you what do you think you were at personally well and because because even though I I mean,
Speaker 2:I think at least some, if not all of the claims were eventually backed up. I think there was, you know, it's kind of a meme at the time, although we didn't have memes back then. We just
Speaker 3:know you
Speaker 2:know stories we told on our grandfather's name.
Speaker 1:Yeah exactly we were not even on our belt. No, you're right, you're absolutely right And we and I think actually but, you know, this goes to kind of a part of Todd's point too that, like, you I think the difference is, like, we never misrepresented that we had not yet written
Speaker 2:a single one.
Speaker 1:But we know we got it was definitely magical thinking. Certainly utopian. No question. Quite arguably magical thinking. Was it othering?
Speaker 1:I hear what you're saying there. I feel like this is a little bit of an invention. Maybe. Maybe.
Speaker 2:Like like, did did Mike go and kill other projects?
Speaker 1:No. Well, not boy. Yeah. Please. This is the most aggressive and annoying form of hype.
Speaker 1:Notably, the time of arrival is less important at this level. Simply that the technology will arrive at some point in the future. I gotta say, like, reading that paragraph, like, that does feel awfully on the nose. I know that we well, I mean, because ultimately, we were insufferable about it. In in particular, a colleague a senior colleague was working on a problem and was complaining about not being able to debug it.
Speaker 1:And I did helpfully volunteer. I'm like, Dtrace will actually does actually solve. I actually said, Dtrace does solve that problem for you. Use the present tense, which is ridiculous. And he is this is, Tim Marcelin, who is not necessarily prone to blow ups, but definitely blew up at me, in his very dignified English way, telling me that I needed to go write Detroit's.
Speaker 1:That's all. So
Speaker 3:I I
Speaker 1:I think it seems fair. Seems fair. It seems though it it was And
Speaker 2:the and the use of the present tense does kind of put you in in some some of the upper echelon.
Speaker 1:Upper echelon. I think we can definitely see where we we have left behind level 2 and probably level 3. We are definitely it is a level 4 versus level 5.
Speaker 3:Yeah.
Speaker 1:Yeah. Yeah. Mhmm. No. That's that's a good point.
Speaker 1:That's a
Speaker 3:good point. I I don't know
Speaker 2:if it's a good one, but it you know, it's self contained at least. At least we weren't telling, you know, too many people outside of the, of the reservation.
Speaker 1:Well, I do and I think that that's it. I think that you gotta, like the the challenge and the balance is you gotta make clear that, like, no. We haven't done actually any of this. Like, this is all something I think is possible. And I do think that, like, things really I the I mean, because another, like, big one, a in a big, big bust is nanotechnology.
Speaker 1:And, you know, Aaron, you were using your by by your metric of, like, companies named there were definitely plenty of companies that had to rename themselves because Nanotech was such a big bust. And are you Adam, did you like, the character stuff that you No.
Speaker 2:No. I I I think I think this one completely washed over me.
Speaker 1:Oh my god. This was and so the idea was and this goes back to a Feynman kind of quip, that we could build robots that would build smaller robots. This is kind of like a thingy and sing singularity for small robots. And the idea then is that we would build these nanomachines, these molecular scale machines. And then the kind of the pitch, especially made to to generals and admirals, was that it would be weaponized.
Speaker 1:And because a and the kind of the big proponent of this, K. Eric Drexler, was famous for saying infamous for saying that a cow is just a machine that turns grass grass into steak. So we can, like, make a machine that turns you into gray goo. Gray goo was the have you ever heard the term gray goo? This is like okay.
Speaker 1:Yeah. Right. So the the like, we we're gonna build the machines. It's gonna turn into gray goo. And it's like which of course, you know, everyone the military gets very excited about this, like, turning everyone to Gregoo apparently is what they wanna go do.
Speaker 1:But it does seem like, okay, this seems plausible. And I remember reading the book, Nanotechnology by and I got like man, I got like 3 quarters through the book before I realized, like, wait a minute. None of this is real. This is all just like stuff that maybe we could do. And that was not featured very prominently on the tin, I would say.
Speaker 5:Well, you know, the the interesting thing about nanobots is nobody ever mentioned where you would hide the software inside of them. Right.
Speaker 1:Where's the firmware? It's like,
Speaker 5:well, yeah. Firmware will figure it out.
Speaker 1:Firmware will figure it out.
Speaker 2:And How do I firmware update
Speaker 1:this killbot? I I quickly. But but but my right leg is gray goo and my left leg is going fast. Yeah. No.
Speaker 1:I there and there were lots of practical problems as it turns out. And that was, you know, Todd, that was an example of, you know, trying to use hype to to get funding and interest, and it definitely succeeded on on on by that metric. But by not representing where it was properly, it allowed people to get way ahead of themselves. And then it was such a deep bust that that was I don't know where it is now. I guess people just haven't heard of it now.
Speaker 1:But that was a term that you literally could not use, and that became a scarlet letter.
Speaker 5:I mean, Nanotech is still
Speaker 1:Of the the kind of the AI winter of the eighties, the expert systems were gonna be were gonna save us all. And that that would really into microcomputers, and we'll
Speaker 9:see whether that Microsoft company actually goes anywhere.
Speaker 1:That's right. Exactly. Well, exactly. Like, plenty of this stuff actually, like, lives up to it. The I did have a couple others that I that I wanna make sure that I get I get out there.
Speaker 1:Actually, I was I went back and read, First Round, the the VC company, pulled their, their founders, asking their founders what they thought the most overhyped, domains were most overhyped sectors were. And they started doing this in 2015. The most overhyped sectors in 2015 included cryptocurrency, AR, which I think ARVR, which I think is definitely one of those things that's been I I don't know. We'll have to talk about that one separately. I don't know how you kind of chalk that one up because there's a lot of a lot that's real there and a lot that's not yet where people think it's going.
Speaker 1:And then the other one and, Adam, this is what I wanted to bounce off you because I don't think I remember this. Chatbots.
Speaker 2:Yeah. I I think chatbots were, like, weirdly huge. I mean, still are something. But, yeah, my last company, there was this there, like, there you know, I went to the, the Slack conference. I mean, obviously, like, they're gonna be filled with with different vendors, but there were tons of different vendors.
Speaker 2:All who wanted to show up with their own both, you know, ML based chatbot, natural language chatbot, and this is both for, like, chat ops, but also for when you, like, you know, DM, mister Peanut and and tell them about your allergic reaction or whatever it is. You know, like, having having, you know, replacing humans with with these chatbots.
Speaker 6:Yeah. Has anyone has anyone ever met a chatbot that they like? I
Speaker 5:I There's an alarmingly large number of companies that still think this is an okay way to interact with customers.
Speaker 1:But is this like an but this is like a sector. This is what I didn't realize. This is like a I mean, Adam, you were at, like, chatbot con.
Speaker 2:Yes. I was at bymanchatbot
Speaker 1:con. And
Speaker 2:and yeah. I mean, I I I, like, I agree, Matt, that it it persists. But there was this, like, this summer of chatbots or whatever where where it did feel like, this was gonna be a much bigger thing. I no. I don't know about you know, I'm trying to to taxonomize it.
Speaker 5:My my my bigger objection is that in 2020 to 2022, they are still continuing to proliferate.
Speaker 1:And where are chatbots? Are chatbots magical? They do not
Speaker 3:have this.
Speaker 5:They were promised, and they will not stop until they turn the entire Internet into chatbot gray goo. Like and, like, it seems to be contagious between
Speaker 1:together, like, big chatbot? It's like, I you know, are you wondering why I called you on here?
Speaker 6:We we we'll get into other angler. There's nothing left on the Internet except chatbot.
Speaker 1:Well, okay. So, Dan Olson, who did have this great, 2 hour video on, this kind of NFT take down over the weekend, he did report, getting into Discords where chatbots are trying to sell Bitcoin to other chatbots. So I do think that this is the thingy and singularity is gonna happen with chatbots like
Speaker 2:oh, that's amazing. It's alright.
Speaker 1:So that was a sector. That that is not the the the founders from circa 2015 from 1st round are are not wrong. There are chatbot comp and then what happened to the chatbot companies? They're still around?
Speaker 2:I think that a lot of them are still around. I mean, they're, like, you know, the powering, the pop ups that, like, whenever you go to any web page, it's like, hey, I saw you've been on my web page for 30 milliseconds. Can I sell you something? That like, those are all powered by chatbots.
Speaker 1:Okay. I've got another one to ask you about, and this is prompted by Aaron talking about when companies name themselves after a trend, serverless. Is serverless where's serverless in all this? That's interesting. Because clearly real, I mean but also clearly in the exaggerated claims from what I was doing, but not I mean, you gotta say there's some Utopian futures around.
Speaker 1:It's terrible as
Speaker 2:I think that that's right. I mean, because just because I I mean, I even almost wonder if it it it went to level 4 to magical thinking because there was this this snake oil of serverless as this panacea. That like Right. The, you know, the, you know, the problem with your microservices environment is like too much state. And somehow, I I I mean, maybe I was hanging out with the wrong kind of nerds because I certainly was, like, I was very interested in Lambda and functions and and all the different flavors of this.
Speaker 1:What's there's definitely, like, there's a lot that's real there. It's just got it got over its keys, I think.
Speaker 2:Absolutely. I mean, and it's like we can't all be stateless services. Like, at some point, we can't just pass the buck infinitely.
Speaker 6:Well, the restaurant, the block
Speaker 1:chain. Right. Just ask the chatbot. I mean, it's I I feel like that's it. Like, can we somehow, like, combine the various hype cycles into, like, a super cell of of
Speaker 2:And then name a company after it.
Speaker 1:And then name a company.
Speaker 2:Where you're going.
Speaker 1:Yeah. Exactly. So then the other just from the drawing from, a deeper past, when you were talking about Intel, I didn't know Intel's done a couple of these. I want I was just trying to go Optane or Itanium. I think before.
Speaker 1:Yeah. You got you gotta put that one in the it I mean and VOIW more generally.
Speaker 2:Yeah. So so I'm trying to place exactly the time of the optine upcycle, and it's gotta be sort of, like, very late 2000s, very I mean, late, nineties, early 2000s. Is that right?
Speaker 1:On Itanium? Yeah. Yes. So the I would say Itanium, and I would say it reaches its peak with, like, the Monterey and the this is SCO and IBM and Intel, I guess, on their, I would say 99 is gonna be the peak of peak IA 64. And with it, this is where I got that I bought a this is one of my most treasured possessions.
Speaker 1:My book on IA 64 that has the graph of the adoption of IA 60 4. And in 2020, it was going to be the late adopters in agriculture that were going to
Speaker 3:be adopting IA 60 4.
Speaker 2:Well, they still have yet to adopt i64
Speaker 1:for their credit. I know. I had to I had to I'm looking for I I kinda when I found this thing about moving, I kinda tweeted that out. I'm looking forward to the laggards in agriculture FCOE.
Speaker 3:Yeah.
Speaker 1:I had the fiber channel over Ethernet. I feel like that kind of I mean, maybe we were just in a little like closet where people were where this is getting hyped, but if you were in the right spot, I thought this was getting definitely hyped.
Speaker 3:Yeah. It was Yeah.
Speaker 6:I thought it was creating with it.
Speaker 1:It it also did Tom did you work at this? I thought FCOE just made no sense with I I I just why would I
Speaker 2:it was a solution in search of a problem, it felt like.
Speaker 1:Well Okay. That I feel is also an interesting attribute of a lot of these things.
Speaker 3:Yeah.
Speaker 1:As a solution insert as I'm just looking down the list, we got a lot of solutions in search of problems.
Speaker 6:Yeah. Yeah. The the one I was biggest into was a ATM.
Speaker 1:Yes.
Speaker 6:It's definitely into the magical thinking and, maybe even othering because there there were definitely definitely personal wars going on between ATM and non ATM world.
Speaker 1:There definitely was. Right? And that red herring cover with Andy Bekkerscheim on the red herring cover of that the ATM, I remember very vividly.
Speaker 6:Yeah. Yeah.
Speaker 1:So in the end, I mean, obviously, like, you don't run that as a cover if there's not other ring going on for sure.
Speaker 2:Hey. Reg regarding FCOE, do a company that took the the success of FCOE and made it was like SATA over Ethernet or something like that. It was like a anyway, it was it was Yes.
Speaker 6:That was good. 8 HCA over Over Ethernet.
Speaker 1:That's right.
Speaker 2:That's right.
Speaker 1:Yeah. Yeah. Which
Speaker 6:which made a certain amount of sense too. And then there was a a SCSI over Ethernet from Singapore. Right. But,
Speaker 4:The important thing about all these is, you know, it's an argument that you only need the one cable. You know, you well, you need power, and then you need the cable unless you talk to other things. And you don't need a fiber channel cable and ATA cables and serial ports and all sorts of stuff hanging off the back. I think that's important.
Speaker 1:It's all on cable in terms of the problem that's
Speaker 5:lost. And and, like, a lot of this stuff, it seems like was not wrong. It was just early, which, I mean, if you're Wall Street, it's the same thing. But, like, NVMe over fabrics seems to be pretty legitimate at at risk of being shot down by, you know
Speaker 6:The other example is is RDMA transiting from InfiniBand Ethernet. So that that's a legitimate thing now. But, you know, Cisco could have pushed FCOE, but, you know, they they had a f fiber channel switch business to protect, and they didn't wanna lose that. Yeah.
Speaker 1:And I think I mean, no. NDME OF, I think, is is I think I think it's with Jetpack. I think you're you're in a safe space. There's, like, San Diego. Right?
Speaker 5:Okay. Good. You know, I mean, I I know that, like, oxide's still doing direct attached SSDs to your processors, and I wasn't sure if I was getting shouted down by a mob for saying VME over fabric. But, you know.
Speaker 1:I mean, direct attached NVME. But yeah. I mean, it and I think well, so actually, or where are smartNICs in all of this? Are smartNICs in the I I I I thought
Speaker 5:I mean, smart nicks are legitimately extant.
Speaker 1:Do you but do you think that smart nicks are becoming to exaggerate claims at all?
Speaker 5:They're suffering from the fact that they're usually FPGA adjacent and therefore a real pain in the ass to program.
Speaker 1:Right. Fair. Okay.
Speaker 5:But, like, if you are trying to do some hypervisor hypervisor level functions, like segmenting NVMe or VLANing, etcetera, and you're willing to put in a little bit of time on some generic code, like, I I mean, Amazon got great returns out of them in the beginning just because they could arbitrage the price between ARM and x86. Right? But, like, I could see these being real to, like, real people and not just Amazon.
Speaker 1:Yeah. I think so too. So I I agree.
Speaker 6:Well, it I my my theory is that it's a lot more real if you think of them as if you rename it to be front end processors, which existed forever.
Speaker 5:Right. Like, I mean, SSL offload is just probably the easiest one just because they do a lot of it.
Speaker 3:And
Speaker 1:another I mean, the Internet started
Speaker 9:that way with a dedicated processor that was doing the network stuff and then handing it to the computers.
Speaker 1:Yes. And certainly, that has been a cycle that I think we are absolutely gonna repeat, especially as we get kinda post Moore's Law here. Another one from, Adam, and I think that's how you even make reference to this on online, but, Talogen and Yeah. And pink.
Speaker 2:Yeah. Yeah. Yeah. I was I was thinking about that one because this is another one from from when I was a kid. It was a huge Mac booster.
Speaker 2:I went to the conferences and had Mac user magazines, which were, like, ran to 300 pages and, like, I'd fill out the form to get, like, free swag from companies and stuff. And, yes, so, like, Apple, their their next generation OS, they they were building system 7, but then system 8, which they codenamed Pink or Copeland, was gonna be, like, the the, you know, preemptive multitasking, like, we need next generation operating system. That was the answer to everything. Like, it was the coming messiah.
Speaker 1:And you know that Taligent, the company that that that Apple and IBM set up as a portmanteau of talented intelligence.
Speaker 2:I've forgotten that, but yes.
Speaker 1:Which I I feel is another, like, red flag. And when you were talking about SmartNex, Tom, you were saying that if they maybe just rename themselves, and I say this is someone who work on SmartOS, but did not name it. That, putting, I think, smart, intelligent, creative in your your name is probably as much of a red flag as putting the name of your technology in your name. In was your company name. Exactly.
Speaker 1:It does demonstrate hubris. It definitely oh, god. Who speaking of which, who would name a sister that? But that's just Gollig. Yeah.
Speaker 1:I'd say, alright. I feel do you I think we got to most of the ones on my list. Oh, another one for you, Adam. As a, I think, b.
Speaker 2:Yeah. That that was super well hyped. I don't where did it land in terms of the hype cycle? I don't know. Maybe a bit of a utopian future, especially because all of Apple's previous efforts had failed.
Speaker 2:Yeah. So, folks were actually getting desperate. And so, it it felt like the future that that could not otherwise be delivered. I say this is the owner of 2 books on b programming and, like, a stack that has, like, a 16 of CDs that to install Bee that is a 16 year old. I like lifted from their booth or something.
Speaker 2:I don't know. I I I apologize.
Speaker 1:Did you go with me to the Bee bankruptcy option? We there together?
Speaker 2:Yes. I did. Yes. I did. I I got a mouse pad that was not part of the lot, and I know that you I don't know if you still have those whiteboards.
Speaker 1:I will the oh, you know, I do not still have these whiteboards. So the the whiteboards, there were, I bought I got it, like, at home I mean, I how many I bought, like, 12 whiteboards for $10. It was a it was a lot of whiteboards for not very much money. And then do you recall that one of those whiteboards, we could not erase?
Speaker 2:That's right. Because it had, like, their their vNode architecture
Speaker 1:on it. It it it had, like, a 3rd of the BOS VFS architecture on it written in permanent marker. So you know that someone had, like, picked up a Sharpie thinking it was a driver's marker. And I felt like I feel like I'm right there with you. I can I can feel it all happening in front of me?
Speaker 1:So that no. I think we had that one in. So when we moved, we lived on Potrero Hill in moving to North Valley. And Bridget is like, this whiteboard has gotta go. There's a huge whiteboard that had a bunch of, like, stuff in part.
Speaker 1:Like, this is a historical art
Speaker 3:like,
Speaker 1:you get this, like, the you know, this is a an important historical artifact.
Speaker 9:So you donated it to the Computer History Museum?
Speaker 1:I well, no. I I feel I did one better actually. What I did is I wrote the saga of this whiteboard on the whiteboard. I explained the whole thing. The the bankruptcy auction, that it was from b, that here's this BFS architect 1 third of the BFS architecture that comes free with this free whiteboard.
Speaker 1:And I just, like, put it up on the street on on Missouri there on on on Tarot Hill. And, you know, God bless San Francisco. You know, maybe 15 minutes later, we see a truck, like, slow down. Guy gets out, reads the whole whiteboard very thoughtfully. It's like I can see him kinda, like, thinking of himself, like and puts it in the truck and drives off.
Speaker 1:So I've got no idea what happened to the whiteboard, but I would like to believe that it's in good home right now. I or at least somebody is arguing to not throw that thing out wherever it is.
Speaker 2:That's right. It's being moved yet again.
Speaker 1:It's being moved yet again. Exactly.
Speaker 2:Yeah. When we were at that, bankruptcy auction, I had turned down an offer from Bee, maybe like 6 months before we went to that auction. So I felt pretty smart.
Speaker 1:You were not actually being auctioned off. And you remember that I got the I talked about other prize possessions that I, I found a box that had an AT and T Hobbit menu on it and an exponential x-seven format on it. So I'm like, I will bid. Like, this is what I'm here for. And I will bid, you know, how do I bid on this lot?
Speaker 1:And I'm thinking, like, I'm immediately doing the calculus of, like, how much am I willing to pay for this thing? And it's, like, a lot. And I find the auctioneers. I can't find the lot number on it. And the auctioneer looks at it and and looks at me and says,
Speaker 3:just take it.
Speaker 1:I was like, wow. Okay.
Speaker 2:He's already been paid, not on commission.
Speaker 1:I think that clearly not. Clearly not on commission. I think he's just, like, box of crap. I'm like, alright. Well, I would have paid a lot more than that.
Speaker 1:That's fine. Yeah. I will definitely
Speaker 3:take it. Yeah.
Speaker 1:But but so yeah. B, I feel got into I think some of that utopian future is magical thinking, honestly. Yeah. Maybe not. Yeah.
Speaker 2:But At least utopian future. I mean, because they were such a dystopia at the time and there was no way out.
Speaker 1:Yes. And I wanna think of that. This is one of these challenges. I've been just going back to Todd's point is that like you got to be you do have to point people to a better future to get them excited to be involved in a nascent technology effort. To be to develop it, to fund it, to use it.
Speaker 1:Like, you do have to, like, get them out of, like, the the the present.
Speaker 5:Thinking about it more, I think one of the key distinctions between 34 is interoperability with legacy systems.
Speaker 6:Like, interesting.
Speaker 7:Like, you
Speaker 5:you once you say, oh, I'm I'm Steve Jobs building the next machine, and it does not mount literally anything from anyone else anywhere that has ever been written. Like, this is just flipping magic at this point.
Speaker 1:No. I think that that that's kind of the interesting way of looking at it. So you think about, like, the whole web 3 thing, you're like, no. Like, no. No.
Speaker 1:We don't need to interact with we are banks. We actually replace the financial system. It's like, oh, okay. Alright. Yeah.
Speaker 1:Woof. We're definitely goodbye level 4. Hello level 5.
Speaker 2:Yeah. But it's not necessarily the technology, you know, producers itself who who create that hype bubble. You know, I was thinking about our previous shows, and we talked a lot about Docker.
Speaker 3:Mhmm. And,
Speaker 2:like, the Docker hype cycle was out of control.
Speaker 1:Out of control.
Speaker 2:Right? Like and, in terms of, you know, venture money and so forth, both in Docker and anything in its ecosystem. And there, there was at least a UTAE Utopian future there, if not magical thinking, but I think a lot of it, at least initially didn't come from Docker themselves, or at least they, I mean, maybe I'm rewriting
Speaker 1:the script. I don't think
Speaker 2:that they, they were not advocating it in the way that their their, you know, disciples started advocating it.
Speaker 1:Well, so Docker is an interesting kind of use case because the what actually happening, people were actually using it. Yeah. I mean, it I mean, that's what was happening. It's like developers were downloading it and using and it was useful. And so then, like, the magical thinking around Docker or the utopian future was much more on the business that they would build rather than the tech.
Speaker 1:Because containers are still very much. But that's right. I don't mean containers then
Speaker 2:Yeah. Absolutely. Right. Right. The the the the notion is clearly there, and, like, there's a there's a real there there.
Speaker 2:So I I I think there's some questions utopian
Speaker 3:like,
Speaker 5:it is the utopian future and all the ridiculous money that got thrown at Docker more just about the idea that for a while, there was a very strong thought that you could make a lot of money off open source software somehow without an appreciable business plan? Yes.
Speaker 3:Right?
Speaker 5:Yeah. And and, like, that was more of a generic feature of the industry than a doctor specific, although they're certainly the elephant in the room.
Speaker 1:Wait. Are hold on. Are are we the elephant? Who's the elephant? Is Oxide the elephant again?
Speaker 1:I I just feel
Speaker 5:mean, last I checked, you're selling hardware. Right? You
Speaker 3:you have
Speaker 5:the Ox serverless as Oxide?
Speaker 1:Frank. Yeah. That's right. We're gonna pivot to serverless. Like, yeah.
Speaker 1:What? What? Yeah.
Speaker 2:It's for your stainless service.
Speaker 5:Okay. We we are attaching them to weather balloons so you can run-in the cloud. It's perfect.
Speaker 1:Exactly. You know, I think so, Matt, that's an interesting point is that the actual the the the hype was not around Docker per se. It was caught up in a larger hype bubble around funding of open source projects effectively as VC funded businesses. And that definitely got it. When you think of these things as VC funded businesses, you get into a lot.
Speaker 1:Yes. You definitely get in to, a lot of level 3, level 4.
Speaker 2:Yeah. The the alchemy of converting GitHub stars into dollars.
Speaker 1:Yes.
Speaker 2:You know, that that that felt like there was certainly some magical thinking there. Yes.
Speaker 5:I mean, especially on the basis that most people who are running open source software tend to be incredibly cheap.
Speaker 1:Oh, I so I was actually talking to one of the VCs invested in Docker, and I I was saying, you know, one of the things I think is kind of interesting as a developer myself, that what I like to use are the technologies that I know are unmonetizable because I don't want to discover that we need to go pay royalties for some I mean, I'm not gonna use proprietary software in part because we can't afford to have royalties based on our technology stack. A lot of developers pretty similar. And so if as an investor, if you are seeking out those things that are being used by developers, if that's your metric, it's feels conceivable to me that you are gravitating toward to those things that are tautologically unmonetizable.
Speaker 9:But that's kind of where the hype from Docker came from. There's a there's a belief in the industry that developers have no budget, but ops has large budgets. And Docker was like, great. We can make the developers fall in love with us and then the ops guys actually pay for us.
Speaker 2:So so how did this VC take that, Brian?
Speaker 1:Nervous laughter. And then I was kind of like like kind of laughing along. And then I like, do you, do you think that could be true? I'm like, I don't think that could be true. Yeah.
Speaker 1:I think it could be true. Sure.
Speaker 2:Eric, to your point, I mean, I think they're looking for the Trojan horse. Right? The these cheapskate developers take the thing that is free and easy, and then somehow you can use that to roll in and spring spring out with your salespeople.
Speaker 1:And that's what the only
Speaker 5:way that I can think of to be able to do that is to do something like Oracle does with VirtualBox and its extension pack, which is just, like, completely a Trojan horse and also makes everyone hate the product once they find out about it.
Speaker 1:It it's not good.
Speaker 5:Like I
Speaker 9:mean, I know DB, your first key value store is free.
Speaker 5:My second one is not.
Speaker 6:Well, in some sense, Red Hat screwed up the the industry by being successful, so people think they can replicate that somehow.
Speaker 1:Which okay. So Red
Speaker 5:I mean, Red Hat's core strength was showing up at a time when Linux really freaking sucked and making it suck a little bit less.
Speaker 1:End disrupting proprietary units. Right? Then now there was a it was x86 disrupting the risk vendors, and there's a lot that's going there. I do feel that this is the important point, Tom, that you're making. Because people look to Red Hat as the model, And they there's this kind of assumption that, like, no.
Speaker 1:We like, Red Hat did this. We can do this. And I do feel that this is one of those things you get at that level 5 stage where people start kinda weaponizing history. And it's like, no. No.
Speaker 1:Like, you're on the wrong side of this. And certainly, I mean, god, the number of times that that folks that are asking tough questions about web 3 are being portrayed as, like, critics of the Internet. It's like, no. That is actually not what happened at all. But I think that it there is this kind of idea that, like, no.
Speaker 1:Because this thing succeeded, therefore, that implies that all possible things can succeed. It's like, no. That is that is not what it implies. Alright. Well, I think we got a we got a good, I think this has been a good good tour of hype.
Speaker 1:I wanna go see, what I'm wondering is who won the big mister Peanut account for among the the big chatbot? I'm wondering, isn't it? That that's I wanna go DM Mr. Peanut. Apologize for for muting him and see if I can figure out who's that.
Speaker 2:Who that vendor is, that lucky vendor.
Speaker 1:Who that vendor is. Adam, any any
Speaker 6:I'm disappointed we didn't touch on 5 g.
Speaker 5:Do you feel 5 g? We gotta come out there.
Speaker 1:Okay. We'll we'll we'll end on 5 g. Do is 5 g I feel like 5 g has been a long time coming. I can't tell if I is 5g overhyped? Can someone
Speaker 6:Oh, and and and it's the same damn playbook as 4g and 3g. And, you know, it's basically we need more bandwidth, so we're gonna hype this to the moon. And, none of that makes any sense, but more bandwidth would be nice.
Speaker 5:Well, I mean, hang on. Some of this makes sense. Right? You can go and you can take spectrum and you can divide it up more finely as opposed to the the comically large chunks that LTE did. And so you can plausibly fit more people into the same amount of space.
Speaker 5:That part is true. This is what they call 5gfrone, which is basically the low frequency bands of it where you take 4 g and like, if you're T Mobile, it's a software update.
Speaker 6:The the technology the technology is great. It's the it's the hyping that's out of control. You know? What what what the end user gets is more bandwidth for cheaper maybe someday.
Speaker 1:So, Tom, could you explain to me how the canceling of scheduled flights for 5 the 5 g rollout? Like, I I feel like is that real? I mean, is that like a
Speaker 5:That's the FAA being a bunch of
Speaker 1:Yeah. I mean, clearly, this is like yeah.
Speaker 6:That's the same same thing as not being able to power on your phones on the plane. They haven't proven decisively that it could cause could not cause harm, and therefore, they ban it.
Speaker 1:Right. That just seems like a very weird Eddie in this particular hype cycle on on on 5 g. Alright. Well, I think we'll we'll we'll leave it on on 5g. I don't know.
Speaker 1:Unless you got any, any parting thoughts?
Speaker 2:Nope. That's that's all the hype I can stomach for one night.
Speaker 1:All the hype you can stomach. I think been good. It's been a good tour. And I'm you gotta go watch that Machine Dad. I'm telling you.
Speaker 2:Yeah. No. That it's I'm it's gotta I gotta queue it right after this.
Creators and Guests
