Understanding Crypto 8: Tim O'Reilly: How does Web3 Compare to Web 2.0?
Tim O’Reilly has a history of convening conversations that reshape the computer industry. He's played a key role in framing and evangelizing terms such as “open source software” “web 2.0” “the Maker movement” and “government as a platform”. He is the founder, CEO, and Chairman of O’Reilly Media, and a partner at early stage venture firm O’Reilly AlphaTech Ventures (OATV).
He is also on the boards of Code for America, PeerJ, Civis Analytics, and PopVox. His book, WTF: What’s the Future and Why It’s Up to Us, explores what technology advances teach us about the future economy and government as its “platform.” He is a Visiting Professor of Practice at University College London’s Institute for Innovation and Public Purpose, where he is researching a new approach to regulating big technology platforms by limiting their ability to extract economic rents.
Welcome back to another comprehensive and informative episode in this limited series of the Rational Reminder Podcast, a weekly reality check about the world of cryptocurrencies. We are lucky to have another respected figure in the crypto world on our show, technology veteran Tim O’Reilly. Tim is a well-known educator and publisher in the crypto community and the Founder, CEO, and Chairman of O'Reilly Media. He has been involved in the technology industry since the inception and rise of Silicon Valley and, with over 35 years of industry experience, he is known for popularizing the terms ‘open-source software’ and ‘Web 2.0’ and is a respected commentator on the space. In today’s show, we do a lot of looking backwards with Tim and draw insight from his vast experience in the industry to discuss what the future of cryptocurrencies holds. We also talk about what separates companies that survive from those that do not, problems in society that blockchain technology can be used for, and changes that have taken place in the technology industry during Tim’s career, as well as what the definition of true innovation is, and much more. Don’t miss out on a truly mind-expanding and informative conversation with special guest, Tim O’Reilly!
Key Points From This Episode:
What Tim’s definition of a technology unicorn is. [0:03:07]
The problems that Web3 could solve and what makes a product innovative. [0:04:39]
Differences between companies that survived the .com bust and those that did not. [0:11:05]
Whether Tim thinks there are promising projects in the crypto or Web3 space. [0:18:36]
How society would change if Web3 became the new normal. [0:22:19]
Examples of the applications blockchain technology could have for record-keeping. [0:24:36]
Tim shares his thought son whether or not we need blockchain technology for aspects where the state is involved. [0:27:24]
Ways in which the evolution of crypto has changed Tim’s mental map of the future. [0:31:37]
The role public blockchains could play in breaking the recentralization trend in decentralized technologies. [0:35:02]
Tim tells us if he thinks centralization is a bad thing. [0:37:15]
Past mistakes and innovation within the tech industry. [0:41:39]
Tim’s opinion on the support of Web3 from venture capitalists. [0:44:02]
Whether or not building technology outside of the government's reach is the right way to improve society. [0:46:46]
Advice Tim has for those who are looking to get rich from cryptocurrencies. [0:50:35]
Read the Transcript:
*Quick note: We recorded this episode back in March before the recent crash in prices and DEFI insolvencies. Please keep in mind while listening to the episode.
Tim, how do you define a technology unicorn?
You know, I have to say I'm not terribly fond of the term, unicorn. To me, a real unicorn is one that is magical, something that really does something that is unexpected. If you think about it, the iPhone was a unicorn. I use this phrase from, this definition from Tom Stoppard's play Rosencrantz and Guildenstern are Dead. Where he talks about you see this marvelous unicorn and a second person comes and sees it and it's truly amazing, and then by the time the third person comes, it starts to become, oh well it might be this, it might be a deer with one broken off horn, and so on until it becomes as thin as reality. And you think about that, think about how radio was a unicorn, and then it became trivial. Television was a unicorn, it became trivial. The iPhone was a unicorn, now everybody has it. You know?
And so that I think is a really useful definition of unicorn. The way new technologies are amazing, and then they become commonplace. And I think about Tesla's rockets going on, Blue Origin going out and landing, like when we read science fiction as kids, that's what they were supposed to do. And that's a magical unicorn. A startup that doesn't do something terribly interesting valued at a billion dollars, I don't consider that a unicorn. First of all, there's too many of them, you know?
What does web three allow us to do that was previously impossible?
You know, that is the question that we all should be asking. What does web three do that was previously impossible? And I am struggling to find an answer to that question. There are a lot of things that it could do potentially, but realistically I see some incremental improvements right now. But the idea that, for example, we're going to have a new currency, the most widely used cryptocurrencies are so unstable in value that there's no way they can be used as a currency. You have no idea what they're going to be worth tomorrow. So, and yes they're used a little bit. And I guess I would also just say, if you look at the history of banking in the U.S., every bank issued its own currency, and then eventually it was like, dude, we need one currency. I think the future is going to be in government-backed crypto currencies.
But more than that, I think the promise of cryptocurrency that you hear all the time is A, it's going to be somehow decentralized, free from the power of big centralized institutions. And I don't see any evidence of that happening. First off, cryptocurrencies are incredibly centralized already with a relatively small number of holders holding the bulk of the value. So the reality and the promise just don't match up for me. And I do think that there's a set of applications that are waiting to really be done where you go, yeah, that really would be better. And I'm thinking about things like, I recently bought a piece of property and they had to do a title search, and this property hadn't changed hands for 50 years so it took weeks because they wanted to make sure that it was nothing had happened in the last 50 years.
And you go well, eventually if we have all that stuff on the blockchain, you go, okay, we're just adding another update to the ledger. And something that was hard, becomes easy. These assets have ledgers associated with them that are imutable, and that's a really good use case. And it will make a certain set of things easier. I think Dows, if they could be made to work in the real world, I think are really intriguing to me in the sense that... But I look at will they work in the real world of regulation? I mean could you, for example, raise capital to do something with this decentralized group? That'd be kind of awesome. But how likely is that when the reason you can't do that with ordinary currency is because there's so much fraud that it became a regulated activity to raise money from investors.
So why do we think that somehow crypto is going to route around this artificial constraint that's brought from outside that is not intrinsic to the old technology? If you contrast that with the internet, which was this truly decentralized technology that just innately routed around the bell system and the long line telephone system with packet switching rather than with literally point to point transmissions hand... It was a different technology. What I see so much in crypto is, we're building this alternative financial system whose chief virtue is that the people who get in early, get a lot of it and get very rich. And so everybody's hoping, well my preferred currency will become super valuable.
And I go, that's just, I didn't see any of that kind of thinking in the early days of the PC or the early days of the internet. There were people who wanted to get rich, but they thought they'd get rich by delivering real value, not by just being in the right place at the right time and getting a bunch of other suckers to adopt this thing.
I look at so much of the activity around crypto as being speculation rather than real invention. And that's not to say that there isn't real invention there, but my point is that when I wrote the piece that I wrote, why it's too early to get excited about web three, is because I think you can only call something like that in retrospect. Web two, we didn't say, oh there's this new thing coming that we're going to call web two. I mean there were people who did that actually, but when I defined web two, it was like, why did some things stick around after the .com bust? So the exercise, I think we should be going through is saying, okay, there will eventually be a bust of all this speculative crypto activity. And after that bust happens, somebody will say, wow, Solana's still there. Why did Solana make it and Dogecoin didn't? Or whatever, you know, or why did Ethereum make it and Dogecoin didn't? I don't really know who the winners will be, but there will be some winners.
And then we can ask ourselves, what were they doing that the other people weren't doing? And I think we can now, by looking back and saying, wow, what was Amazon doing that Pets.com wasn't doing for example, which is what we did for web two, we can do a similar exercise prospectively and ask ourselves, are there projects that are doing real stuff and really executing on it? And we look at operational metrics. Like if you're in the business of money transfer, say hey we built a cryptocurrency system for remittances say, I go, well how is our actual remittance volume growing? And how does that compare to somebody like wise.com that's doing it with just plain old, existing money? Is it demonstrating that it's meaningfully cheaper, that it's meaningfully easier to use?
When Amazon came along, everybody went holy... I mean I still remember, that was when I first really started paying attention to Amazon when a friend who would never use the computer, told me that she had bought a computer so she could use Amazon. But I guess all I was saying is we should be going through and saying, which of these things are really enabling this new kind of behavior that we... They really are enabling something that was previously impossible.
So what were the characteristics of companies that did come out successfully after the .com bust? What made them different?
There were three things that I, three or four things that I defined as central. And the first was what I, at the time I called harnessing and collecting intelligence. Later on, we called it big data. It was this idea that Google and Amazon, they said that they're fundamentally engines for getting smarter by harvesting massive amounts of user data. And at that point I was just seeing it as a very positive thing, because it really was. Google does things like, wow we can improve search by noticing what people click on. They clicked on the third link instead of the first one, more of them did that, you know? And so that maybe means the third link was better. And so they were, or Amazon did the same thing. They took all these factors into account, so that was collective intelligence at work.
You know, Wikipedia was another one. It's like wow, they're harnessing thousands of people. I used to do this animation of the Wikipedia page for that big earthquake in Japan that the Tōhoku or Sendai earthquake that ended up having all the nuclear plant trouble. And you watch this thing grow from a one line entry in fractured English to this massive 20 page essay over a period of time from thousands... That was magical. That was a unicorn. Linux was a unicorn too I mean, going back to open source. People marveled at the fact that wait, you could build an operating system with thousands of people who have never even met face to face? So that collective intelligence aspect was a big part of what I thought was central to web 2.0. Although I did also warn I said, the other side of that is that data will be the new source of lock in and monopoly control.
One of my web 2.O principles was using the Intel inside slogan, where Intel had become one of the dominant monopolies of the PC period, I said data will be the next Intel inside. So, but there was also this idea of much more flexible user interfaces, the idea of the programmable web, the idea that the internet was becoming the platform. Of course we take that for granted today, but at the time people laughed at it. That wait, you'll make API calls to remote internet services? And those services will often be about data? Those were the things that I called out as really interesting. If you look at web three, you'd certainly say, well this unique factor, I think of the blockchain is the imutable ledger. And it's really a challenging one, because it's both a bug and a feature.
And so people are going, oh well we'll use web three to displace the big internet giants. And then you look at, it was Dries Buytaert, the founder of Drupal did his first static website using web three technologies. And he was like, dude, it was really hard, really primitive. And you go, really you're going to build a website on an imutable blockchain? No, this wants to be dynamic and no record kept. So I would just say one of the things that I would be looking for is, are these people taking advantage of the actual strengths of this technology? And I would say, yeah, there are going to be opportunities.
Could you, for example, put together an alternative to the big central role of the investment bankers, for example. But if you think about that, what would that look like? And it wouldn't be decentralized in the end, because if you look at financial markets, you know SPACs kind of looked like an alternative to all that stuff. Oh we'll just go do these funny things. And then they don't get any analyst coverage, they don't get any of the big guys looking at us so they don't get any of the institutional investors buying, and these people could go public, but they got evaluations.
And so in a similar way, what's the infrastructure that it will take? If you look at the infrastructure of the current financial system, it's extraordinarily complex. It's more complex than even Google and the internet. And you think about what's happening today in the global economy because of Russian invasion of Ukraine and you say, okay, wow. We never really thought about the way that central banks would have to compensate for the fact that oil came from Russia and wheat came from Ukraine, and it's driving up inflation on those things.
There's so many entanglements, this is the lifeblood of our entire system. It will be a ton of time before we have a new blockchain replacement for all that infrastructure. Particularly because I don't see anybody really working on it. It's as if when... What I've often said about web three is it's as if we had the .com bubble 10 years earlier. You know, useful, but nowhere near prime time 1980s internet. There was, hey, cool.
There was email, there was FTP, there was Telenet. These were cool internet applications. They were pretty magical, but they were not ready for prime time. Imagine there had been a huge financial bubble associated with that, it wouldn't have been healthy. The web came along was the killer app, and then it was what got built on top of the web. And I think it's also really important to understand. People go, well it's still early, we're just building the infrastructure. I was just talking with some investors, for example, recently and they were like, well we're building this free infrastructure. And I said, yeah but don't you remember the history of the free infrastructure of the internet? It wasn't Tim Berners-Lee who became a multi-billionaire, it wasn't Vince Cerf who became a multibillionaire, this stuff grew away from all the money, Then it became super big and then people solved the problems of it getting really big. And I just see that there's a lot of cart before the horse kind of thing here, where things are being valued before the thing has actually scaled.
And some of that is maybe a fault in the reporting. Because people will talk about, well the number of transactions on some exchange. And you go, well wait. What kind of transactions are they? Are they just transactions as in something got written to the blockchain? Are they transactions as in somebody transferred money from one party to another? And everybody's just kind of throwing a lot of fud around. Again I talked to an investor and he said, oh but it's still early, you know, it's hard to know these things. I go yeah but, in the early days of the internet, it wasn't hard to know these things.
Mike Prettejohn started the web server survey pretty early on, I think in 1994, we had a count of how many web servers there were. And we could see it going up into the right. We knew who the players were and who had what share of that web server market. We knew how many browsers there were, we knew how many active users there were. Same thing with the smartphone market. And yet you come to web three and there's this huge cloud of unknowing, it's too early, it's too... So I guess I feel like there's this real thing in there, and I am just so excited for the bubble to burst and for all the people who are hoping that they get rich, as opposed to, well we're really solving this interesting problem and we're really getting traction with it. And then we'll see who's who. I'm just waiting for that.
On that, are there any projects that you see in crypto or web three now that make you kind of think, wow, that could be something like what we saw when we called web 2.0 web 2.0?
Part of it is, I'm not as close to this market as I was... I mean we literally were, we did the first commercial website, the first web portal, first internet advertising I mean, I was right in the middle of that one and now I'm kind of an observer watching from the cheap seats. So, I don't. And I'm starting to spend more time in that space and trying to understand it better. But mostly what I've been trying to do is just say, look. If there's some there there, start talking about these things instead of those things. Stop talking about valuations and start talking about users. Stop talking about valuations and start talking about what actual real world outcomes are happening. And again, I think what statistics I've been able to find, haven't been that encouraging.
I just look at something like, if you look at the valuation of the Bored Ape NFTs, for example versus, okay well how many people are doing this? It's just like, no a bunch of people got a bunch of valuable crypto and they're basically turning bubble crypto into other bubble crypto. And I'm still looking, that's all I can say.
And again, I think there are some real projects that give me hope which are where people are building infrastructure, but I don't think those infrastructure players are necessarily going to be big financial winners. So again, there's the people who are doing the real work aren't going to be getting rich, and the people who are getting rich now are getting rich in the same way that people were rich on paper before the .com bus. And I think the people who are really going to get rich from crypto, they made a lot of money and then got out, which and I have talked to some people who are like, I made my money and I'm out of this now, I think will be this next generation that builds on the leftover infrastructure.
And that goes to these question of productive bubbles. I believe that there's a pretty serious crypto bubble. And the question with any bubble is, what's left over when it bursts? And how useful is what's left over? And it may be... And then I'm very influenced there by Carlota Perez's work, her book Technological Revolutions and Financial Capital kind of makes the case that there's always a bubble with every new technology. But there are sometimes bubbles that are much more productive than others. Like if you look at the mid 2000 real estate bubble here in the U.S. Or for that matter, the Chinese real estate bubble that's been going on, or the Japanese real estate bubble that happened in the 90s. They were building out stuff that nobody needed. We had this massive bubble through this innovation of collateralized debt obligations that everybody said had all these magical properties that they didn't actually have, and there was a lot of pump and dump of mortgages.
There were a lot of builders were building homes that were sold to people who couldn't have really afford them, and it was basically all taking. And so that was not a productive bubble. Versus you look at the internet bubble, the .com bubble, those left over that people built out Fiverr, they built out data centers. I look at what's happening right now. There's something of a space bubble when Space X is valued so highly, and guess what? Elon is building fricking... I just became a Starlink subscriber, he's putting up satellites is, this is left behind, if the value comes down, the stuff will all still be there. Whereas the value of Bored Apes comes down, what do you got?
That's really interesting. So you mentioned new normals with tech unicorns. How do you think things would be different if web three did become the new normal?
I don't know what you mean by the new normal. And I don't know what you mean by web three, you know? So I guess I would say, kind of coming to the parts that I feel like I can really say, yeah that would be really useful, I would say, it would be really awesome if chain of title was something that you did not have to go figure out. Let me give you an interesting story. Way back when, I was talking to the mayor of Jerusalem. And he said, one of the real problems in Jerusalem is nobody knows who owns stuff, so nobody can fix stuff up. And I don't know if that's actually true, but I remember that assertion. Yeah part of the rule of law and property rights are the knowledge that this is really yours.
And we have a lot of parts of the world where chain of title and who owns something is not clear. And I go, that's a very real and very big problem that blockchain is uniquely suited to solve, but there's also a bunch of others. I go, right now we have this... You think about something like medical records. Like dude, I want my records. I'd like to have them associated with me, not with my hospital system or my doctor or whatever, and I go, is somebody building that?
Again, it's an application that's well suited for blockchain. It's an application where you really could say, yeah we're going to build a new decentralized architecture that's centered on the person. And then some people are thinking about that in a generic way, but nobody's building the company that says, oh yeah yeah we're the health records, the blockchain... And maybe they are. Again, there could well be a startup out there that I just don't know about. I'm not like I've done this big wide search. But you go, there's a hell of a lot more chance of a blockchain startup dethroning Epic than there is of a blockchain startup dethroning Google, right?
Yeah. That's interesting. I don't know enough about the healthcare records business, but on property title. I think that one's interesting because you have by definition, the state as a trusted authority. So do we need a public blockchain or a blockchain as a database to control those records?
I think there's a lot that could be done there. And a lot of invention. You know when you think about it, there's going to be a lot of skew amorphism in our initial approaches. Where we'll kind of go, we'll just build the blockchain version of things as they are. And then somebody go, well wait. We don't really have to do it that way. Let's just say you have a title company that says, well we now have our new blockchain system that keeps track of title. And then along comes somebody who says, well, and again it might well be an existing. Along comes Google and says, well we're just going to associate a blockchain record with every address. We know every address and we're going to go buy those records off from whoever has them, and we're going to put them on the blockchain and associate them with every address on the face of the earth. Suddenly they've routed around whoever does that stuff. Back in the day, the post office could have done that.
I mean there was a time when the post office, in the early days of the web, had an opportunity to be... When I first started talking about the internet as platform with these big database subsystems, I said location is going to be one of those subsystems. And who had the best, it was this huge database of here's an address and here's who's associated with that address, and they never did anything with that.
So I would be looking for things like that in particular where you want to have a record. That would be, again that's one class of application. But there's a lot of clever things. And in some sense, I feel like the whole Bored Ape phenomenon has done a disservice to NFTs. Because of course an NFT is just basically the notion that there is a token that's associated with a bunch of data, and it's not, unlike a cryptocurrency a fungible token, i.e. like money. It's a non-fungible token, and that token could be anything. It could be a legal contract.
It could be this chain of title associated with an address. It could be my medical records associated with me. It could be my DNA records from 23andMe and any other DNA tests I take. There's all kinds of things that you could do that we go, okay just here's my NFT chip for my DNA records. Here's my NFT chip for my bank account, whatever it might be. Those things are much more interesting uses of NFTs than trading cards.
It's interesting. But the part that I still struggle with with NFTs, like in the examples that you're just giving there is, if we're talking about like the Ethereum blockchain, for example. Do we need a public blockchain for stuff that ultimately has the state involved or we have a central authority by definition. If it's in a country where they don't have a robust legal system, maybe it's interesting.
Maybe, maybe not. I guess what I would say, again, I always look backwards. And I would say a good use case in Google. Where Google became the authoritative source. Again, the post office was the authoritative source, and now Google is with Google Maps. I don't see it as, even though it is public, that it has to be... That it will necessarily be owned by a public entity. In fact, it's more likely that because the public entities are slower, that it will not be owned by a public entity. And again, if I were trying to look prospectively at web three opportunities, one of them would be to identify some class of data that is currently owned by someone else who has not figured out how to put it effectively onto the blockchain. And when I say effectively on to the blockchain, I don't mean just, well we put it on the back end.
It's like, we figured out what the front end looks like. And again, that's an area where I think that web three has really fallen down on the job. If you read Moxie Marlinspike's piece, that he does this great job, we're just kind of going, I'm trying to figure this out. And I go, wait oh my wallet's not actually talking to the blockchain. It's talking via a traditional API to some big centralized player. Whoa, I thought you told me this was decentralized. It's on the blockchain, but the big centralized guys can just shut off my access to it, who cares? You know? And so I think in that sense, thinking through those hard problems of how are we going to, like if we want to build a decentralized user-centric version of some of the things that exist today, what will drive adoption?
What are those killer apps? You know? And I do think that there are certain kinds of things where somebody's going to figure out, oh yeah actually, I built a great biometric ID system, for example. Because again, these are markets that don't seem that big. Then there are companies doing non blockchain things like ID.me, who are kind of trying to be a provider of identity for, they actually serve government because government's so crappy. You know government would actually be a really good place to do that, but we have a political thing in the U.S. against having a government issued ID, which they don't have in Europe.
But again, you look at that and you go, right now there's this sort of whole set of random biometrics that are collected in various places. Some people might have your retina, somebody else might have your voice print, somebody else has your face, somebody else has your fingerprints. And you kind of go, would it be good if there were the most authoritative version of this is me, was an NFT that was associated with me that I own, that I could update.
That would be a cool thing to have. I still struggle with, does it need to be on a blockchain, on a decentralized blockchain? Or can it, is that just a problem somebody needs to solve with a database?
I agree. Well again, the question is, the thing that I think could be interesting is, could you make a unique association between essentially decentralized data store and the thing itself?. If you take this example of an address or a person, you do identity resolution. And my friend Jeff Jonas has this company called Senzing, which is an identity resolution company. He started out his career in identity resolution working for Las Vegas casinos. He'd just go this Tim O'Reilly used to live with the dealer. He's winning at blackjack on table five, and he used to be roommates with the dealer. Those are the kinds of problems that he solved. But in order to do that, you need to know that it's the same Tim O'Reilly. And that's usually done by, you find enough data points and you go, yeah. Anyway, I don't know. I just think there's a lot that potentially could be done, and there may even be people doing it, but I can't see it because of the fog of the bubble.
You may have partially just answered this question, but you talk in your book about having a mental map of the future. Has crypto changed your mental map of the future?
Not meaningfully. And I have to say, the thing that I... We're in a bit of sort of an AI trough of disillusionment, but I think the thing that will really potentially be destabilizing for the current big tech order, are large foundation models in AI.
I say that based on our own experience at O'Reilly where we basically have a search engine now that's based, a natural language search engine, where we took Google's Bert... We were working with a company called Miso who do this. They basically train it on our content, and now we have search that's better than Google for our content. That's amazing. And their whole business is helping small companies or e-commerce companies search their own content.
Oh, your search is insane. I've used it. It's crazy how good it is.
Yeah. It's so good. You know like before we were doing stuff with Solar and it was like clearly we don't have a big enough search staff to do all the whackamole that's needed. You change this parameter and then it bulges out over there, and then you change that parameter, you fix that and then something bulges out over there. It was a huge problem. With a machine learning model, it just adjusts thousands of things at once and kind of comes to a pretty optimal thing, and it's just a fundamentally better technology. And I think that you look at these models and their evolution, and we're just at the point where they are going to be able to integrate it, be integrated into various kinds of services. And I think there's a huge disruption waiting to happen there.
And while many of these things are, the model itself is... Right now I mean if you're using open AI and you're calling the Codex API over at open API, you are at the mercy of their big centralized. But there's other cases where somebody has actually, like in the case of Bert was released as open source, and as other people who are building these things, releasing them as open source. And so there is this interesting potential where that could be a decentralized market. The other thing that I find really interesting on the decentralization front is synthetic biology. There's a model of synthetic biology... And this comes from conversations with Drew Endy who's a professor at Stanford who thinks a lot about these issues. Drew makes a point, there's a model of synthetic biology that looks like the chemical industry.
Like you're going to make biofuels out of yeast, right. And big chemical plant. He said, that's not how nature works. You basically have the information in a seed, the materials arrive, energy arrives, and it makes the thing. And he said, we could build a synthetic biology industry that looks like that. So that's a really fascinating thing. What would it take for it to be like that? And of course, Monsanto is working very hard to make nature look like the big centralized thing.
So there's a real opportunity for people to get ahead of that and go, no, no, no. We're going to actually have a decentralized bio industry of the future. Or same thing, decentralized solar, how does that work? How do we get that out of the hands of the utilities and make it easier for energy to flow like packets flow? We could have an energy internet.
Do you think that public blockchains will break the trend of recentralization that decentralized technologies tend to experience?
I have a very hard time predicting the future. I think everybody does. What you can do is you can see trends, and then... The way I operate is I look at things... Well first of all, I look back at history a lot and I go, well what does that tell me? Partly history gives you a set of mental models for what happens. When I look at the future, I look at some trend, like there's a trend for us to build bigger and bigger living molecules. Drew thinks, well we're going to be able to build a cell from scratch. You can kind of look out along that trend line and go, if that trend line continues, certain kinds of things will be possible. And then you go, okay well, I don't know if those things are going to happen, but I know what some of the sign posts are going to be, and if they happen, then I'm looking for the next one.
And same kind of thing. When you saw Elon Musk's rockets taking off and landing and being reusable, you could go, well that will enable a bunch of stuff. And none of us necessarily said, oh he's going to use it to build a satellite network. That was a little unexpected. Jeff Bezos says, oh I want to do space manufacturing and mining and stuff like, who knows? Will he do it? Will he not? But we're going to see data points that tell us whether that future is coming true. We now know that it's increasingly possible.
So you look for... There's a great line from the poet Wallace Stevens, he said, searching out the possible for its possibleness. We're all doing that. And so you watch for something becoming possible and then you say, okay it's possible now, is it actually happening? And what are the signs that it's happening? And again, just think about you're traveling along the road or hiking somewhere, and you're looking for the next signpost or the next pile of rocks. And you go, okay I found that next turn, now I'm looking for the next one. And so you kind of know that there's a destination out there, you're not able to predict any particular tree along the trail.
So fascinating. So tell me, is centralization a bad thing necessarily?
No. I often think of decentralization and centralization as phases. And every technology that I've looked at has gone through a series of decentralization and recentralization. And this was really the thing that introduced me to this idea was really looking at the history of the computer industry. I entered it really in the mini-computer era, when companies like Digital Equipment Corporation and Data General and Prime were trying to wrest the mantle from IBM. And then along came the PC, the personal computer, and the PC was kind of a mistake on IBM's part. There were microcomputers, the Sinclair and the Osborne and the Apple. And they were like, hey there's this little market dudes. Let's just, you know we got to catch up, it'll be 100,000 computers, and so we'll just make it out of commodity parts. And we'll license the operating system from this little company Microsoft, we'll catch up that way.
And then it turned out, of course, because it was made out of commodity parts anybody could make one. They had this requirement that everything had to have multiple sources, but when it came to the 386 Intel made the decision they were like, no we have enough a market that we're just going to tell IBM, screw it. We're the sole source for the 386. We're not licensing it to anybody else. And they won that bet. So all of a sudden that was the first step towards centralization. And with Microsoft, with Windows, they suddenly go, okay, everybody had to write their own drivers for devices. And Microsoft says, hey dudes, we'll make the driver problem go away. We'll write all the drivers, you just call our APIs. And nobody quite understood that how much power that would give to Microsoft. Right? And suddenly Microsoft is the new central hub.
They've taken the... The PC takes the over from the mainframe, and Microsoft takes over from IBM. So then along comes the internet and open source. Everything's open, everything's decentralized again. You know what I mean? The PC had been a totally decentralized market, got re-centralized. And the internet comes along, totally decentralized. Everything from ISPs to websites, everybody's building their own websites. And you know here we are within 10 years, you've got Google. Well first you have Yahoo, then you have Google, you have Amazon centralizing eCommerce, before long there's some big dudes who are earning most of the profits. And so I guess I would say, having watched it there three times, that just says, yeah decentralization happens, and then somebody figures out how to re-centralize. When I described it originally, I was looking at this idea that I had written this article called hardware software and infoware. This idea that hardware was the center of the computing universe, and IBM was the big dog and they controlled the hardware specs.
And then we went to commodity hardware and then Microsoft figured out how to control it via software. And then there was this we're entering this new age of what I called infoware at the time, which was basically active websites that were database backed and services. Later on I dropped that term. But really it was the programmable web that I was talking about. And cloud computing. You think about well, what happened was that data was used to become the new locus and centralization. We ended the big data area, the people who had the most data got market power. And so what's going to decentralize big data is, I mean big data is the thing that needs to be decentralized now. And as I say, that's why I'm kind of interested in this idea that it may actually be large foundation models that decentralized data. If you look at something like GPT3 or any of Google's large models, they've basically encapsulated in a single AI model, a huge amount of the knowledge that used to be Google's alone.
And it's now encapsulated in this AI model. And so I think that's actually going to be a decentralization of data in a really interesting way. Anyway, I just see them as alternating stages. The thing that's interesting about it is that you have the possibility of a change of regime, but it's not guaranteed. Actually I really recommend Tim Wu's book, The Master Switch.
Oh, it's a phenomenal book. I agree.
Yeah it's a phenomenal book because it's really about the architecture of industries. And you kind of get, here's radio. And what's his name, David Sarnoff, it's kind of like an early radio enthusiast. Kind of like Larry and Sergey, be no evil. This is this liberating force. And then he kind of builds up RCA and CBS. And I forget the guy's name comes up with FM, and which is just a much better technology, and he's like, oh man, this is going to really screw us. This could be low power, it can be lots more stations, could be decentralized. And he just does this big PR campaign, it doesn't really work, it's not ready for prime time, and holds it back for 30, 40 years. Television, it's like it's not ready for prime time.
And you know, and he has the FCC in his pocket, and it's not ready for prime time until RCA has basically figured out how to reverse engineer what the independent inventors have done. And then suddenly they reveal, we're in the television era, but it has the same structure as the radio era, with the same networks and the same stations, and we end up with 30 years of that. So somebody can really engineer this stuff wrong. And I think there's a good chance that... His whole point about the, once you have industry concentration, it tends to persist for a long time.
And I actually have to say that usually the decentralization happens somewhere else. It's never a direct attack. And that was one of my takeaways from the open source community. Where it was, everybody was so focused on Linux and GNU Linux, and we're going to build the free operating system for personal computers. And Linux was important, but not anywhere near as important as the internet. The internet was just like some other layer that kind of expanded out into something completely different. And I guess you'd say that some of that is happening with blockchain and web three. It is certainly completely different. But my sense is that the chief use case so far is financial speculation, which just doesn't bode well, you know?
What do you think about the venture capitalists... I mean, not naming any names, but venture capitalists that are really backing and getting behind and promoting web three as the future.
I don't know quite what to make of it. In one sense, it is a continuation of what's been happening for the last decade, which is that the market is no longer picking winning technologies, venture capitalists are. Along comes Sunil Paul, who invents what we now think of as Uber and Lyft. He raises $35 million, he's cautious, he doesn't want to raise more money, thinks he'll keep more ownership, doesn't go very fast. Lyft and then Uber copy what he's done, raise a ton of money. And if that had not happened, if there had not been so much cheap capital, ride hailing would probably have developed more like the internet. I.E., it would've been lots of different small companies, you would've had taxi companies go, oh, actually that works. And you would've had a decentralized version of Uber. And then it would eventually have gotten rolled up and centralized, but it was pre-centralized.
It was centralized from the get go. And that was the piece I wrote, you know my critique of Reed Hoffman's blitzscaling, this literally became the venture capital playbook. Can you speed your way to monopoly? David Rosenthal, the Stanford professor, wrote a really interesting critique of web three. He said, this is never going to be decentralized because it's funded by venture capitalists, and venture capitalists want monopolies. And they're not funding a decentralized web three, they're funding a centralized one.
The thing that's really bad about it is, in the case of the internet, you had this period of 10 years of lots of experimentation where the end users picked the winners. And now you have a period where there are people with the biggest funds, a couple of big egos, pick the winners. And again, returning to the Tim Wu book, it's a little bit like we're building a future where a few very, very large funds are basically making all the decisions and they can be wrong. Whereas like, if you have thousands of people making decisions, many of them will be wrong, but some of them will be right. But we basically created an environment where the venture capitalists are deciding what will be done and what will be funded. And there's not that many of them. And they're not that smart. It's possible that you have somebody who's in such a position of power that their self-interest actually trumps the actual advance of the technology. They pick the winner that's best for them.
Crypto censorship resistance is attractive to people who don't trust the government. That's kind of one of the things that is made Bitcoin and crypto popular. Do you think building technology outside of the government's reach is the right way to improve society through technology?
That's a very interesting and difficult question. On the one hand, I believe that government is an incredible force for good. It is in some sense the ultimate platform for society. The rule of law, kind of like a reliable financial system, these are things that are engines of prosperity, and where you don't have them, things go very seriously awry. And governments are uniquely positioned to do that.
Governments though do get sclerotic, governments do overreach, there are real risks in government power. And so, are there circumstances in which you want to have the ability to keep secrets from your government? Absolutely. I'm a big fan for example of Signal, for that reason, you know? I go, I want to have encrypted communication sometimes. And at the same time, I don't believe in this sort of crypto anarchist society. I mean, we have to go forward and think, how do we work together?
And that's why I will say, there's some wonderful stuff happening. I think, the Ethereum community seems very forward looking that way. I know that they're very deeply engaged with economists like Glen Weyl, who is really going, wow. We could do markets differently. And, let's experiment with that. And those kind of things give me a lot of hope. Not with having the crypto wild west that a science fiction writer like David Brin would write about or whatever, or Vernor Vinge. But instead, come up with better governance structures. And there's something really powerful and wonderful in the idea of basically a blockchain ownership as a governance structure. Which is why, again, this intersection with Glen Weyl's work is really, really interesting. Like how would we govern a sort of a shared, in some sense it's some shared digital space that everybody owns.
And that part of the decentralization story I really buy, and is really interesting. And the question is, what do you apply it to in the real world? Like if you think about the decentralized governance of, which turned into again, eventually centralized governance, of open source projects, what was it enabling in the real world? It was enabling decentralized software development. A real thing. It wasn't just for its own sake. Nobody was just kind of going, oh, this is so cool, you can do decentralized software. Somebody had to actually be doing software. There was a thing that you used it for. And so again, I just keep coming back to that, okay we have this decentralized capability, what problem are we solving with it? And I do think, see some small glimerings, like you kind of go, okay, we're going to buy this thing collectively.
Okay. And effectively, what we're going to do with it is effectively our, a result of our votes with our currency. Well, is that good? Is that bad? That's kind of what the Coke brothers are doing or, you know we have more currency, so we have more votes. Is that really what we want? I mean, I pick out the Coke brothers, I should just say you money and politics.
It's a really good point though.
Yeah it's like, if we decry it in the current financial system, where people who have more money have more political power, and then we're building a system in which the people who have more cryptocurrency have more votes, aren't we just recreating the thing we like least about the current governance structure?
So a lot of the crypto discourse is really focused on amassing exorbitant wealth. So do you have any words of wisdom for people focused solely on trying to get rich with this?
Yeah. You can't take it with you. I won't mention his name, but you know, talked to one, a guy who had amassed quite a lot of crypto wealth and he's like, I don't need more money, I'm just trying to figure out how to spend it usefully now. And that is a real goal. It's like, okay, what are you going to do with that money? And that reminds me of something Brewster Kahle told me a lovely story. Brewster Kahle founded the Internet Archive. He first became rich when he sold WAIS to AOL, which is in 1993, about the same time as I sold GNN. He says, I go to this financial advisor and I say, what do rich people do with their money? And the guy says, well, mostly they count it. And Brewster said, I don't want to do that, I want to do something with it.
And then so that when he had his next exit, which was of Alexa, the original Alexa, but not this Alexa search engine that he built, which he sold to Amazon, he founded the Internet Archive. He said, I want to take that money and turn it into something I want to keep doing. And so my hope is that the crypto wealth that gets amassed, that it's put in service to build a better future. And that it isn't just wow, we amassed massive amounts of money, let's buy some yachts. Which is kind of, you look at the failed states of the world, that's what people do with their money.
That's a great answer. All right, Tim, this has been a fantastic conversation, we really appreciate you coming on the podcast.
All right well, thanks. Glad to chat with you.
Participate in our Community Discussion about this Episode:
Links From Today’s Episode:
Rational Reminder on iTunes — https://itunes.apple.com/ca/podcast/the-rational-reminder-podcast/id1426530582.
Rational Reminder Website — https://rationalreminder.ca/
Shop Merch — https://shop.rationalreminder.ca/
Join the Community — https://community.rationalreminder.ca/
Follow us on Twitter — https://twitter.com/RationalRemind
Follow us on Instagram — @rationalreminder
Benjamin on Twitter — https://twitter.com/benjaminwfelix
Cameron on Twitter — https://twitter.com/CameronPassmore
Tim O’Reilly on Twitter — https://twitter.com/timoreilly
'Hardware, software, and infoware' — https://dl.acm.org/doi/10.1145/253671.253687
'Why it's too early to get excited about Web3' — https://www.oreilly.com/radar/why-its-too-early-to-get-excited-about-web3/