[Editor’s note: This is part of a series examining the internet’s first 50 years and predicting the next half century. Join Xconomy and World Frontiers Forum on July 16 for [email protected], an event exploring the internet’s past and future.]
Right alongside Moore’s Law, which describes the exponential growth in computing power since the 1960s, there’s Metcalfe’s Law, which describes what we’ve done with all that power: namely, use it to connect the world.
The law proposes that the value of a network increases in proportion to the square of the number of endpoints—an idea that turns out to be roughly true, according to a couple of quantitative studies.
Before he got a law named after him, Bob Metcalfe earned his renown as the co-inventor of the Ethernet networking standard and the co-founder of 3Com. Today, he’s director of innovation initiatives at the Cockrell School of Engineering at the University of Texas at Austin.
Few people have a fuller perspective on how quickly the internet has grown and why it’s set to keep disrupting new industries. Metcalfe will join a panel predicting the internet’s next 50 years at [email protected] on July 16. I interviewed him for a pre-conference look at the forces shaping the internet’s future. Here’s a lightly edited transcript of our conversation:
Xconomy: If you had to guess, how long will it take to reach a 100 percent internet penetration? Will that happen for all 10 billion people who will be around by the year 2050 or 2060?
Bob Metcalfe: Well, we’ll never get all the way. It’s asymptotic. But it will be on the order of another 50 years to complete the adoption cycle, I think. It started 50 years ago at zero, and reached about halfway in just 50 years. Assuming a certain symmetry to the adoption curve, that would say another 50 years.
X: There’ll always be some people who aren’t on the internet, probably out of choice or just because they’re so remote or cut off by their governments.
BM: Yes. And you saw how I did that—there was no magic. I just assumed that the adoption curve is symmetrical.
X: So what follows from that? I mean, I think that we’ve discovered in the last few years that the wonders and glories of the internet, the things that everyone had stars in their eyes about for the first maybe 25 or 35 years, have their counterparts in the form of fake news and all the other pathologies of the internet. And I wonder whether having full connectivity, or close to full connectivity, would make those pathologies even harder to deal with?
BM: I think that the thing that’s created these pathologies is that the connectivity has grown so quickly it has temporarily outstripped our ability to handle it. So, we’re catching up. It’s happened so quickly, in just 50 years, depending on when you want to start the clock. And I think it has less to do with the penetration level than it does to the suddenness of it all. We’re still trying to figure out how to deal with this connectivity, and we will. We will figure it out.
X: Do you think that will happen on a case-by-case basis, coming to grips with specific problems and coming up with a fix, and then waiting until the next one?
BM: Yes. I think we’re making [a mistake]. One that has attracted my attention is the fake news mess, where we all decided to ask Google and Facebook and Twitter et al. to filter out fake news. We asked them to do that, and then they made the mistake of agreeing to [filter] it out. And so, it’s going to take a while for us to recover from those errors. But we’re going down that dead end right now. It’s an example of an issue where we haven’t really figured out the correct answer. I think I know what the correct answer is, which is the branded composable filter, where we are in charge of filtering our own information.
X: You mentioned that in our emails before this interview. I wanted to ask you what you meant by that?
BM: Imagine that there is a workbench for building filters, and you have a bunch of branded ones like New York Times and the Wall Street Journal and Bob Metcalfe down the right-hand side. Then you take your mouse and you select some, and you pull them over and you compose [a set of filters].
What I mean by branded is, there’s a lot of them, and they have someone’s name on them. I like the New York Times. I suppose that they would have some filters to contribute to this. So, you then compose them, that means you can imagine this workbench allows you to add them and subtract them and complement them until you can compose a filter of all the existing filters. Then you could put your name on that and make that your branded composer. And then whenever you then come to this service, Facebook or Twitter or whatever it’s called, you would then look at it through your filter, so you would decide what’s fake news and what’s not. Or you can borrow other people, right, by choosing to put their filter inside of your composition.
X: I love this idea. It sounds adjacent to ideas like the old-fashion RSS reader, which has nearly disappeared, but there was a time only 10 or 15 years ago when you could set up a really cool personalized RSS reader and just pull in the feeds you wanted.
BM: Oh. I had forgotten about RSS feeds, but I’m imagining this composition tool being much richer than that. Anyway, do you want to start this company? We could just do it. You know, build the composition tool. I know Tim Berners-Lee has decided to start his own company, with pretty much the same idea, which is the people own their own information. In this case, people own their own filters.
X: Well, the problem with that would be, how do you get a VC to fund it, if it’s not driven by advertising?
BM: Why does it have to not be driven by advertising?
X: I feel like the examples we’ve seen today of advertising-based business models in social media and internet-based information have been imploding and kind of backfiring on us.
BM: I don’t remember you East Coasters complaining about the New York Times being advertising-based back before social media. I don’t remember you being all upset about how CBS and NBC were all advertising-based long before the internet.
X: Fair point, fair point. I really should have said “psychographically targeted advertising à la Facebook.” But this actually leads to a related question, I think, that gets back to our theme here. I like this idea of branded composable filters where you’re in charge of assembling your own way of filtering the news. But it feels like giant companies have been setting the agenda in so many areas, including creating the gateways that most people go through to find their news and information. Facebook is one of them. Google is another. So, is there a way in which we’ve allowed these giants to take over our access to information? Or maybe you could have said the same thing about Microsoft in 1992.
BM: And in 1985 you would have said the same thing about IBM.
X: I guess these monopolies die off eventually.
BM: They do! That’s the point. They come and go. And right now, it’s Google and Facebook screwing up.
X: So, I guess the question I was working toward is, does that worry you at all? Or do you think these big entities that have too much power will ultimately fumble and be replaced by other entities—who will eventually have too much power and then fumble?
BM: I think it will happen just like that. You know the old rule of adult supervision, that when you start a company and then the company outgrows the founding team, they bring in the adult supervision? Both Google and Facebook are now run by their founders—Larry [Page] and [Mark] Zuckerberg. And that’s the problem. They shouldn’t be running those companies. Google did a good thing by bringing in Eric Schmidt [as CEO in 2001], the adult supervision. And they gave it back to Larry. And Zuck did a good thing by bringing in Sheryl Sandberg to be his adult supervision. But he then made her COO instead of CEO. And he’s just not competent to run that company, and he’s just running it right into a buzz saw, by just not knowing what he was doing.
X: But that’s just kind of the drama we see over and over again I guess. These companies get reined in eventually by someone or by some force. And the next thing happens.
BM: Right. The last time, remember, it was the government that helped us break up the AT&T and IBM monopolies so that we could build the internet. That might happen again. Or it could just be competitors. Microsoft could come back. Bing is going better than I thought!
X: You made a remark in our pre-interview emails that I wanted to come back to, about how the internet will continue to disrupt industries, but even bigger ones. You know it’s already disrupted commerce and news.
BM: Journalism’s financial model has been completely disrupted, and we haven’t found a new one yet, and that’s a problem. Journalism needs a business model.
X: From your lips to God’s ears. But now you expect a similar thing to happen in education, energy, and health?
BM: Yeah, they’re the big ones. Much bigger than the ones the internet has already disrupted.
X: To me, the common fact about those three particular sectors is that they all have extremely sort of persistent built-in inertia, like they all have their own systems and their own technologies. And I think they may be resistant to change.
In healthcare, for example, they’re stuck with these electronic medical record systems that were written in some cases for billing and insurance purposes and can’t live up to the goal of becoming repositories for our personal health data. You have to kludge so much into an EMR to make it even usable for a modern hospital.
BM: I’m currently a registered patient on three different Epic systems, Epic being one of the leading brands of EMRs. And the three of them are not connected.
One thing that’s happened that’s bad, and it sort of agrees with you, is that the term “healthcare” has come to mean administration of records and billability and payables. And the whole healthcare industry has taken its eye off the ball, which is health! We should be improving everyone’s health, not their administrative health records system.
So, I think the emphasis is going to shift. The internet is going to allow us to shift our emphasis back to health instead of being preoccupied by the ideology and the administrivia of healthcare. And those record systems are going to cave. The internet is very good at standardizing databases. And the health records are being standardized even as we speak. So, I’m not as pessimistic as you just indicated you were. So, there are two things happening. One is, the internet is very good at connecting things. So, all those incompatible health record systems will be crushed into a decent system or several of them. And second, the emphasis on health instead of healthcare administration.
X: What about energy? How do you see the internet disrupting the energy sector?
BM: Energy should be viewed as a network, and we will be exchanging energy. We’ll be switching power packets around from random sources—from wind and solar—to meet random demand, at any time of day. So, matching the randomness of time of day and the randomness of supply will require storage. And so, we will be putting energy in storage and then switching it around. Energy will be solved not only by the science of thermodynamics, but also by using the internet model for exchanging power.
X: So, literally having a smarter grid that’s internet-based, but also just applying the idea of things traveling around in small pieces, on an on-demand, as-needed basis.
BM: With storage. So, remember when the internet started, there was no [data] storage in the internet. The telephone company had been built to be sure that the telephone call got through in less than 150 milliseconds, so you couldn’t have any storage. But look at the internet today: The internet is a big, distributed storage system. The servers, even the storage on your cell phone is huge. So, energy storage is the big, needed technology. And then data will be switched among these storage-enhanced energy consumers and producers.
X: OK, what about the last one, education?
BM: That’s the easy one. It’s happening already. Internet learning is replacing bricks-and-mortar education. You can get MBA and masters’ degrees in mechanical engineering without even showing up on campus now. And the price, the cost of an education, is plummeting because of these online systems. And predictably, the education system is in denial, with professors running around saying how important it is that they have personal interactions with their students. And meanwhile, more and more students are getting their degrees online.
X: That’s true. I was thinking of examples like Udacity and Coursera and edX that were all the rage five years ago, but haven’t grown as fast as people expected. MOOCs, massively open online courses, hasn’t turned out to be the be all and end all that people thought it would be yet.
BM: Yet! Oh, I’m so glad you said “yet,” because I was going to add it on.
X: So, you think that technology is just so appealing, in terms of its ability to help people learn remotely, that it’s going to succeed? Someone will find a way to make it profitable and successful?
BM: Well, they are already. I think you need to review the data again. But anyway, the costs are low. Production values are high. It’s lifelong learning. It’s wonderful. It’s happening. You should see all the degrees at Georgia Tech and Arizona State there. Those are two big leaders, offering their entire degrees set online.
X: You know, you just mentioned something a minute ago that makes me think about a conversation earlier today with Don Norman. We were talking about this question of whether you can rebuild the existing internet on the fly and patch it as you go along. You said you thought we could make up for the rapid connectivity and the problems that that’s created. But can we fix things like the lack of a proper system of identity and privacy within TCP/IP and HTTP and all these protocols, or do we need something fundamentally new? Do we need a second internet? Do we need to start over from scratch? That tends to Don Norman’s point of view—that the existing internet won’t go away, but somebody will have to build another one, a new one that has a better fundamental design or a design, and not one that sort of randomly grew out of university culture of the late ’60s. Right?
BM: I think Don’s onto something. Of course, he always is. One of the early flaws in that ’60s design was anonymity. The intelligentsia of the internet has from the beginning believed that anonymity should be the default, and that’s backwards. Anonymity should be the exception. You should be able to have anonymity if you want it, right? It should not be the default, the way it is today. And that’s the cause of our privacy problems, our security problems—that the internet’s default is anonymous. And the way a plumber like me notices that is that I know in my [Ethernet] packets I carefully put both a destination address and a source address, and no one ever pays any attention to the source address.
X: Because they didn’t care, or because they weren’t trained to think that way?
BM: They were 1960s grad students. I was. We were. And you know I wrote the first RFC [request for comment] in 1973 about a security breach of the internet, and no one until that time had imagined that anybody bad would ever get on the internet. And that turned out to be two high school students in Los Angeles. But the people who were screwing with the network now are much worse than that. Anyway, having anonymity means the bad actors can operate with impunity.
X: Right. And so, is that the kind of thing that we can fix by monkeying around the edges of the existing protocols that make up the internet? Or is that something that requires a fundamental rethink?
BM: I can’t answer that question. But Don Norman has raised it perfectly. Can we jack it up and fix the anonymity underneath, or do we have to start over? I guess I don’t know the answer to that. Starting over is really hard.
X: That was his point.
BM: Yeah. There’s just so much invested in standards and infrastructure. You know the TCP/IP/Ethernet standard was made in 1973. TCP on top of IP on top of Ethernet. All in Palo Alto, CA, in the summer of 1973. And here it is, it’s still ticking. There’s just enormous momentum in these compatibility standards. And you’ve watched how IP version 4 is still in the internet, not IP version 6. Just making that small change has proven so difficult. It’s not even accomplished yet. And it has been decades since we started. I used to write columns in InfoWorld in the ’90s about how IP version 6 was imminent.
X: Yeah, Don pointed that out as well, that IPv6 has been on the docket since like 1996, around that time, and it’s still not complete. You know there are portions of the internet that still aren’t using v6.
BM: Big portions of it.
X: That was actually the point I was trying to make when I asked you about education and energy and health care. Those are also sectors that have infrastructure that’s really hard to change, that’s just as hard to change as things like the fundamental protocols of the internet. Which is why I’m a little skeptical that they’ll be disrupted.
BM: The reason that they’re going to be disrupted successfully is that the internet is so powerful and connectivity is so powerful that it’s going to overcome those resistances.
X: Let’s talk about the Great Firewall of China for a minute. Can the internet
stay a unitary thing? Or are we in an age when more countries are figuring out that if they control the internet at their borders, they have much more control over their populations? Is there a chance that the internet will kind of fragment into multiple national internets or regional internets, and would that make any difference?
BM: It will definitely be a loss because it violates Metcalfe’s Law. Small, partitioned networks are worth less than unified ones. And I think it’s possible that, for a time, that partition may work, but eventually it’s a losing game, and the value of connectivity will overpower them. So, the internet is going to finally cure us of communism because of the reduction of friction of commerce and the enhancement of freedom of information which breeds democracy. I’m very optimistic.
X: That comes through. And that’s kind of where I wanted to end. You’re very optimistic, and most of the time I’m optimistic too; I feel like it’s part of my job. There’s no point in being a sourpuss about technology. We need to help people figure out how to make the best use of it. Where do you get your optimism?
BM: I guess I’ve learned over the years that cynics are often right, but they never get anything done. Cynicism and negativity are a dead end. Every once in a while, things look bleak, but then they get better. Like this current mess with social networks—it’s a passing thing. It too shall pass. And so, my optimism is based on having watched connectivity blossom over the last 50 years, and in my book, there’s no sign of it slowing down.