Transcript: After the Revolution: A Review of the Tech Boom, Bust, and Rebirth
with George Gilder, Hance Haney and Matt McIlwainAfter the Revolution: A Review of the Tech Boom, Bust, and Rebirth with George Gilder, Hance Haney and Matt McIlwain
The Rainier Club – Seattle, WA, May 20, 2015 Sponsored by the Discovery Institute
Event Transcript
Eric Garcia, Vice President, Discovery Institute: …Our moderator for tonight’s event is Matt McIlwain, a partner with Madrona Venture Group. Matt and Madrona finds and funds the companies that bring to life the technologies of tomorrow.
Matt McIlwain: …This is a really momentous year in terms of anniversaries here in Seattle and so I’m going to take one minute to sort of frame that and use that framing to allow Hance and George to introduce themselves. So this is a big anniversary year: It’s the 40th anniversary of Microsoft, it’s the 20th anniversary of Amazon, and it’s also 20th anniversary of a little venture firm called Madrona Venture Group. Happily Madrona was here in time to be an early investor in Amazon — that one certainly worked out.
What’s interesting though is that roughly 20 years ago technology was in a very different place, and I think we’re all amazed at how little things seem to change year over year. But when we look back in decade time horizons we see how much things truly have changed. And just to give you two tangible examples from the technology world: those of you that were accessing the Internet twenty years ago — and I’m not sure if it was all of us — were accessing over a dial up modem at 56 kilobits per second. Today depending upon where you access the Internet you are probably accessing it at at least 100 megabits to possibly one or more gigabits per second. And you are often accessing it completely wirelessly. So that is a pretty amazing accomplishment from a telecommunications perspective.
George Gilder: Where can you access the Internet at 100 mbps wirelessly?
McIlwain: Well you know we’re getting there.
Gilder: That is the target of 5G which is the next wireless generation — which will come on in 2018 — and it will be at least 10 times more bandwidth than 4G today. And at the point a lot of the talk about wireline bandwidth becomes kind of irrelevant — wireline will be for backhaul not access.
McIlwain: Let me give George one more example to chew on. Anybody want to venture a guess how much a terabyte of storage cost 20 years ago? One million dollars is exactly the right answer. In fact, I have a friend who was the CIO of a cancer research center and he bought one of those terabytes for $1 million — so it wasn’t just an imaginary transaction. How much does it cost today? A hundred bucks. At Frye’s you could probably get it for $59 — or at Amazon.
Gilder: Now it’s worth understanding that that million dollar terabyte was hard drives and slow access and you couldn’t even process the contents. Well now it’s largely a solid state drive so it’s hugely faster. How much less space does the terabyte occupy? It’s not just a thousand times cheaper, it’s probably a million times cheaper when you consider the real utility advances, the size, the energy usage and the connections to even larger terabytes in the cloud. And petabytes in the cloud.
McIlwain: And I would guess that a petabyte will cost $5 twenty years from now. So with that in mind, we want to take a short journey back into the past and understand what the state of telecom was with the 1996 Telecommunications Act and where we are today with the FCC’s immodest grab at net neutrality. Hance, give everybody in the audience a touch of your background so they know where you’re coming from and then we’ll go from there.
Hance Haney: In 1995 I was a legislative assistant to Senator Bob Packwood of Oregon, who at that time was chairman of the Senate communications subcommittee. And in the previous couple of years beginning in 1992 Congress decided to regulate cable companies because Congress was alarmed that cable prices were increasing at a rapid clip. So the FCC came out with a set of rules that caused cable rates to increase for many consumers. Then we went back to the drawing board and came out with another set of rules that reduced cable rates on average 17 percent, and it took $3 billion dollars of cable industry revenues off the table. But that $3 billion was sufficient to leverage $15 to $18 billion of investment in new generation networks. So cable investment in broadband came to a screeching halt, and there was a recognition in Congress that something needed to be done. Meanwhile, the FCC was trying to encourage telephone companies to provide competing video services through a digital video dial tone type service. But the FCC wouldn’t approve the telephone company applications because it didn’t think the phone companies were reserving enough channels for independent programmers. So we had this impasse.
Republicans took control of the Senate and House in the 1994 election. And we had visions of completely transforming telecommunications policy to create opportunities for everybody to invest in next generation networks. We really failed to deregulate telecommunications but one thing that we were able to accomplish was to create a containment wall for services that combined communications and computing. The idea was that since no company had a monopoly in that area — and everybody was starting from scratch — that was going to be completely deregulated. And we’ve had a period of twenty to thirty years of bipartisan consensus on that. During that period of time — not because the FCC was capping prices or regulating rates of return, but because companies had profits and from those profits were able to invest tens of billions of dollars every year in constantly improving technology — so that in the mid-1990s a gigabyte of wireless data would have cost $100,000 and today it costs $10. So that bipartisan policy has been a huge success, but in the past couple of years we’ve been told all consequential innovation takes place at the edge of the network. At the core of the network we have nothing but “dumb” pipes. But just like it’s impossible to build new highways and freeways to keep up with congestion, network providers are finding themselves in the same situation — it’s impossible to provision dumb new pipes fast enough to eliminate congestion.
In fact the network is not just a series of dumb pipes. It’s constantly being improved and I would just cite two examples: Network functions virtualization (NFV) and software defined networking (SDN). SDN allows you to take things like firewall, load balancing, intrusion detection and reduce each of those from a unique customized hardware device with internal proprietary software into a piece of standard software — all of which can share space on the same server in the cloud. So it reduces operating costs dramatically. Instead of packets just being sent to various servers and bouncing around until they reach their destination, SDN and NFV will allow the network to be operated like the air traffic control system operates commercial aviation. They can find the optimal path to ensure maximum utilization of the network which offers vast efficiencies, and also it enables the operators to dynamically redeploy network resources in real time. So for example let’s say your office has a connection and the capacity has been chosen because let’s say you’re a retailer and you’ve got to be prepared for the Christmas rush and the Mother’s Day rush and so forth. The rest of the time a lot of that capacity is sitting there unused. Well right now you’ve got to pay for that capacity anyway. In order to redeploy that temporarily for somebody else’s use would require a lot of human intervention and doesn’t make sense. So in essence it is a dedicated resource for you and that’s how the pricing is established. But with the new technologies a lot of network resources can be redeployed in real time to meet demand so that what you’re not using can be sold to somebody else. In that kind of a model it becomes economically efficient to charge you only for what you use and not for what you don’t use. So the point I want to leave you with is that if network service providers can invest in constantly improving new technology that’s what offers consumer benefits — what ultimately offers lower prices and better services for consumers. To the extent that networks become more secure, more reliable, offer more bandwidth at lower prices that offers a platform for the innovators at the edge of the network operating out of their garage to create the consequential innovation we celebrate.
McIlwain: Both of our panelists need no introduction but you know George the accolades are nearly endless. Telecosm and Microcosm are just fantastic, and Wealth & Poverty — the list just goes on. Your stories about Israel, your stories about some of the innovation that came out of Carver Mead’s lab over the years — it is just a joy. We’d love for you to take us back to your perspective from 20 years ago and bring us forward to modern times.
Gilder: I just remember 20 years ago I was scurrying all over Silicon Valley covering people like Marc Andreessen who were launching Mosaic and transforming the network. Every couple years there’d be a major change in the whole architecture of the Internet. And I just imagine that if FCC chairman Tom Wheeler and President Obama had been in office at that point and defined the net of that time as the Internet and rendered it a public utility with all the price controls that are implicit in a public utility — and platform for litigation that it offers — we would not have many of the great gains that you dramatically presented at the beginning of this meeting.
McIlwain: One of the things that is interesting is how many people in the room — I think we’re kind of hinting around this topic of network neutrality…
Gilder: And they want to neutralize the network. Eunuch-net.
McIlwain: I like to think of it as Obamanet, but how many of you have been biased or harmed by the policies of the last several years? We had Google somehow emerge, we had Facebook emerge, we then had Twitter emerge. We then had Snapchat. Now we have Meerkat and Periscope. It seems like somehow the network less regulated — it is not unregulated — has been able to continually produce these edges of our…
Gilder: Well a third of all a value on the stock market — a third of the market cap — is directly involved in the Internet and dependent on it.
McIlwain: What are some of the the risks you see in the kind of government heavy-handed regulation that Tom Wheeler and the FCC have to tried to impose here?
Gilder: Because our companies are global platform companies now they will leave the United States and innovate elsewhere. If we seal off our Internet as a public utility they’ll have to move to places where the future is open to surprises. The chief thing that regulation does is halt surprise. Information theory — which is the subject of Knowledge and Power — defines information as surprise, and that inherently can’t be predicted. I think the reason capitalism works is because it’s open to entrepreneurial surprise — to new information. Every business is the entrepreneurial test of an idea. And if it succeeds it endows further investment in that path of progress.
To have a regulated Internet would have inhibited the great surprises of the last ten years. In 2000 when I was in the middle of this, all the attention was on wireline bandwidth of various sorts — fiber optics, and Telecosm was largely about fiber optics. Apart from my prediction about the teleputer — it would be as portable as your watch, as personal as your wallet, it would recognize speech, it would navigate streets, it would collect your mail, your videos and your photographs…It just might not do Windows. But it would do doors. It would open doors to your future… I went on and on.
McIlwain: I don’t know if you realized how prophetic you were that it wouldn’t do Windows. But I think I’m using that in a different sense of the operating system word.
Gilder: But still the general sense was that Internet bandwidth came over wires, but what happened was a thousand-fold rise in wireless bandwidth over the 10 years. It’s probably more than a thousand-fold now. Over a three year period there was a 152-fold rise. And that was completely unexpected; it couldn’t have come from a fully regulated regime. If every time you changed the Internet you had to go through a litigious process at the FCC, you couldn’t have that sort of a transformative eruption. At the moment we’re at a similar time of transition. The move to cloud computing is a complete transformation of the network, and hundreds of billions of dollars of investment to achieve that kind of transformation. And that kind of investment can’t be made under heavily regulated regime. That just doesn’t happen.
McIlwain: I’m curious what’s been the biggest surprises for you since you were there when the last big regulatory change went into law — through the legislation that went into law in 1996. What have been the surprises for you that you might not have happened otherwise?
Haney: First of all, there were few people in Congress — George was talking about the possibilities that existed for wireless at that time, and there were people who understood the telephone companies and cable companies would be competing against one another to offer both voice and video services — but one thing that surprised me at the time is how incredibly cautious most members of Congress were and what they were willing to accept. When the chairman of AT&T said at the time that wireless would not be a competitor for wireline in his lifetime (he is still alive, by the way), people believed that and there was just this enormous cautiousness about the potential for new technologies. One thing that surprises me is how there could have been a bipartisan consensus for so many years that has just been completely up-ended.
McIlwain: What actually is the FCC trying to do here, and what is your prognosis on what might or might not happen?
Haney: The FCC to its credit is trying to position itself as a referee in case egregious conduct occurs, and the chairman of the FCC has said he doesn’t intend to regulate prices or mandate unbundling — which is forcing the owner of a network to provide access at low marginal costs to competitors. The problem is that those commitments are completely unenforceable, so network service providers will in fact find themselves subject to whatever three out of the five members of the FCC think is a good policy of any given point in time. And of course class action lawyers can file lawsuits at any time, anybody can file a petition before the FCC and every time this happens it has to be resolved and there are potentially years of uncertainty every time. For example just whether the taxes are going to apply to broadband services is an open question. Telecommunications traditionally has been one of the most heavily taxed things — like beer and alcohol and tobacco — and broadband has now been designated as a “telecommunications” service. Well there are good arguments on both sides of that but the point is judges are going to have to figure that out and we’re all going to be in limbo while we wait for them to do that.
Gilder: Jerry Hausman of MIT has done a study of telecom taxes that shows that it costs $1.43 to collect each dollar of telecom taxes. Deadweight loss, he assigns it.
McIlwain: So what the FCC asserted was that “information” services need to be regulated under Title II of the 1934 telecom act, but that they’re only going to pick and choose certain elements of Title II that they think are appropriate and they need to engage in today. And then they’re going to open it up to people that believe that they may have been harmed. So the example that I’ve used is to kind of make it a little bit more tangible is if I’m flying on a plane and I paid extra to have Internet access through Gogo, could somebody else say no you you can’t pay extra to have Gogo internet access on the plane? Or conversely if Gogo says well you can pay us but we’re not going to let you watch streaming movies because we’re on a plane here; somebody else could say, no that’s wrong Gogo can’t restrict the access to that content. So these are the kinds of admittedly hypothetical arguments that somebody could choose to make, all under Title II of the 1934 telecom that was created for a regulated monopoly.
Gilder: Based on the railroad commission law.
Haney: From Victorian England.
McIlwain: So do we think that this is going to hold up from a policy perspective, or do we think that there’s going to be wiser forces that rise up and challenge it?
Haney: The 1996 telecom act has been referred to by the Supreme Court as a model of of ambiguity. It’s very vague in a lot of areas. It was specifically drafted…
Gilder: A million words of deregulation.
Haney: But in this one area, the definitions of “telecommunications” and “information” services I don’t believe are ambiguous. The FCC is counting on the fact that under a Supreme Court precedent that they’re going to get what’s called Chevron deference — which means if the statute is ambiguous then the agency gets to supply a reasonable interpretation of it’s choosing. Now the FCC came up with the opposite interpretation previously. And what the courts usually say is you can change your mind but you have to explain why you’re doing it, and if there are serious reliance interests you have to explain how they’re going to be affected and whether the benefits to the public outweigh the harms to the reliance interests. If the new interpretation is based on changed factual issues then you have to explain why your factual determinations were wrong the first time. None of that’s happened this time around. There are also procedural problems here. The FCC came out with a proposal before President Obama intervened. It was very different than what President Obama demanded. And so there is a disconnect between the proposal what the public was allowed to comment on versus what the agency actually chose and that could be a violation of the Administrative Procedure Act. But the definitions themselves in the Telecommunications Act I think are actually quite clear. The situation we have here is that communications and computing are becoming more intertwined every day and so caching, cloud computing, e-mail. And it doesn’t matter whether these services are provided by the network service provider instead of someone else — the statute doesn’t require that. And so I personally feel that the courts should overturn the FCC in this case.
Gilder: And probably will.
Haney: President Obama will not be in office when this comes back before the FCC, I think.
McIlwain: So notwithstanding all this policy uncertainty, George, what do you see as some of the greatest opportunities — telecom-related — going forward in terms of where innovation can happen assuming we can find a way to get the government out of the way?
Gilder: I think the whole Internet’s going to be transformed, and it’s going to take hundreds of billions of dollars of investment — some of that by some of us — and that among the most important transformations will be the provision of a transactional layer for the Internet. “Layer 8,” Matt Scholz suggested. Layer 8 will be conducted through the blockchain currently associated with Bitcoin, which is the leader. If it can get its act together and adapt it’s algorithms to improve its capacity without jeopardizing its security and stability, it can continue to benefit from its first mover advantage It can extend its sway from transactions to covenants and patents and any kind of contractual commitment that might be suitable for the Internet.
I think the most important result of this will be on automating transactions to the degree that you can have hassle free-seamless micro-payments. That will usher in the era that I am going to describe in my new book as Life after Google. Google gets 95% of it’s revenues from advertising and I think that that cannot last. This is a peculiar Indian summer for people who think they can make money and aggregate all content and control content through collecting money from advertisers. I think that anybody who downloads a YouTube video and gets an ad first is irritated. That model just isn’t going to compete with the model whereby you can pay some small micropayment for that download and directly compensate the producer of this content. I also think the advertising model on television — as I wrote in Life After Television — will finally go away because I don’t think anybody really wants to watch twenty five minutes of advertising to see thirty five minutes of content. So I think we’re going to have a transformation of the Internet that will make it much more robust. It’ll have a security model that is much more effective and expedient — a distributed security model.
McIlwain: I am fairly confident that not everybody is familiar with this blockchain technology. And what’s interesting about it from a conceptual perspective is it’s kind of changing something that’s generally centrally-controlled, and making it something that is highly distributed in terms of how it is controlled.
Gilder: Well the existing model for security for transactions is to make the customer authenticate himself by divulging lots of personal information — the last four numbers of his social security number, username, passwords, pins, mother’s maiden name, favorite dog, whatever it is — the user divulges enough information to be defrauded. The information has to be protected. So it’s put in various super secure databases that are surrounded by fire walls and which offer a very promising and attractive target for hackers. You know where the information is and if you can break in you can get millions of card numbers.
McIlwain: Why do people rob banks?
Gilder: Yes. And you have a separate network for major bank transactions. I could go on to the scandal of floating currencies. These monetary units are valued in a market with one hundred times more volume than all the stock markets in the world put together and twenty five times more volume than all the goods and services traded internationally. People think that the existing system actually works and it’s tenable — I think it is a monstrosity and is currently causing this secular stagnation that leading economists use as an alibi for the failure of their fifth or six quantitative easing.
The Bitcoin model is a completely different security model. It says we minimize the amount of disclosure for each transaction, but we publish it on a public ledger that is propagated across the net — thousands of computers hold it. And so to attack one computer and to falsify a bunch of transactions is totally ineffectual because unless you can capture half the computers in the system you can’t change anything. All the transactions in this public ledger are brilliantly integrated by mathematical hash technology so that to change one — the latest — transaction you’ve got to change all the transactions back to the Genesis block of the blockchain. Each of these transactions is irreversibly timestamped. It’s a distributed model of security and it’s more secure than the existing system. It also offers additional uses beyond micropayments overthrowing Google or television.
McIlwain: As an example we’re looking right now at companies that will simplify and reduce the cost of one of the most gouged services you’ve ever experienced: currency exchange. So do you ever think you pay too much when you go from here to Europe, say? If instead you could take your U.S. dollars and push them into a Bitcoin or a blockchain-based currency even if it’s ephemeral — just for a minute — and then it transfers into Bitcoin and then it transfers into euros. Now all the sudden I’ve gone from US dollars to euros. I’m not sitting with anything in Bitcoin but I’m using that as a medium to do a currency exchange in a different kind of distributed, trust-based network. So there’s a lot of possibilities here — mind you, you need a massively distributed global telecom network to do something like this. You need these computers all at the edges of that network, you need the math, you need the algorithms and the math model that was developed to create a concept like blockchain and Bitcoin. I don’t think this would be the kind of thing that a centrally planned government would be instituting any time in the near future. In fact, I think that’s the biggest risk to blockchain — especially as it relates to finances — that governments will not like the power-distributed aspects of it as it relates to the financial system.
Haney: I’m just curious to know what are the implications for tax collection?
Gilder: The blockchain is public. Even though it does allow transactions that resemble cash, nonetheless, as Dread Pirate Roberts of Silk Road discovered, the blockchain is published and every detail can be scrutinized and various encryption tools can be decrypted. We aren’t going to win this as a war against the government. We can win this as a new transactions layer for the Internet. Currently the Internet comprises about six or seven percent of total transactions. Bitcoin is some infinitesimal portion of that. So there’s a long way to go before Bitcoin becomes the general transactions medium for the Internet. But as that happens Internet commerce will be facilitated and improved. I think at some point people will start wondering how many Bitcoins a dollar is worth rather than how many dollars a Bitcoin is worth. That’s a future, the end of a long path. But until then it’s new infrastructure for the Internet — for Internet transactions. And why would the government oppose it — it publishes all these transactions. It has further uses as a ledger — I call it a cadaster, because a cadaster is a ledger that encompasses lots of other items beyond mere money.
Hernando De Soto believes that one of the most valuable uses of the blockchain will be for timestamped titles in third world countries that allow people to mortgage their houses and thus monetize huge amounts of capital that’s currently illiquid because of contestable and murky titles. There are a lot of benefits that I think sensible governments will not attempt to suppress. The Republicans in Congress are quite friendly for Bitcoin and even Democrats actually. I don’t think that it’s unpopular. Putting the Silk Road people in jail was probably good for Bitcoin — Wired has very dramatic articles about that, and just how guilty these people are is hard to tell — but nonetheless it’s good for Bitcoin that it’s shown that it can cooperate with law enforcement when you have actual “hit” contracts being at least contemplated explicitly on the Silk Road site.
McIlwain: Another big topic is cyber security. The reality that I think any commercial data center today has got vulnerabilities, and already probably had some kind of security infiltrations that have taken place. I know some of the biggest companies in the world — of course we’ve seen some very public examples like Target and Home Depot on the retail side — but big companies of all kinds: financial services, etc. What’s the future of this whole area of securing data? Apart from that the conversation we’ve just been having, what can be done about it and what’s the role of telecommunications and possibly other data driven approaches. Hance, do you have any perspective on this topic?
Haney: Cloud based services I think inherently are in some ways easier to protect but also present a more inviting opportunity for hackers. One bit of good news is that Google is migrating all of its internal administrative business functions into a cloud based environment and relying on a system now where they authenticate the device that their employee uses to access the company database. They then verify the identity of the employee and then all transmissions are encrypted. I hope that that kind of optimistic view of the potential for cloud based services to improve security is a good sign. I also think network providers have a much greater role that they could play in this — George and I visited a AT&T a couple years ago. They’ve got very sophisticated tools available for them to monitor manage and view traffic loads and look for anomalies and to sample transmissions to look for to find signatures suspected…
Gilder: Is that metadata? That’s probably illegal, right?
Haney: Deep packet inspection is very controversial, but there are opportunities in the network to vastly improve security. I think some of the network service providers are very hesitant to move in that direction because they’re concerned about the regulatory implications — which certainly now that the FCC is treating everything like a telecommunications service has the ability to regulate it in every way — is not going to encourage network providers to play a more active role, which I think is unfortunate.
Gilder: I think this Google system won’t really work unless they use the Trusted Platform Modules that allow autonomous execution of BIOS pre-boot authentication of the device…
Haney: Technology that’s already available today.
Gilder: Technology that’s already available today in billions of computers but is very slow to be adopted although it’s been endorsed by all the major computer companies. Because if it’s all on the network, if it’s not isolated execution of the boot process it could be manipulated and hacked and so it would not achieve the better security that you’re describing.
Haney: In this space I think people have been wondering how to monetize better security and better privacy. There’s been a lot of uncertainty around that. And obviously money is going into that now because hacking has become so common and so ubiquitous. Major corporate resources are going into security. And hopefully that’s going to be helpful as well.
McIlwain: We think this is an area where the network and the ability to transmit at least metadata in real time across different data centers can be used in an interesting way together. Said another way, if each one of us in the room had our own data center and we could tell on the network that there was some anomalous behavior — this server kept pinging that database and grabbing information out of customer records, and then that similar technique was being used over here and used over here — we can stream back the information and have services that can identify those behaviors. You could actually much more proactively — and in an automated way — take action against that. I think this is similar to how some things are done today in some of the Black Ops areas and beginning to see that apply into the commercial world is an opportunity for greater security.
Gilder: Peter Thiel’s Palantir — which derived from the security functions of PayPal that made PayPal the dominant system for Ebay — and his point is that if it’s fully automated it doesn’t work because a fully-automated system is low entropy; and you can’t can’t capture really unexpected strategies that begin with the hacker understanding the defenses. So what works with Palantir — a multi-billion dollar market cap, billion dollar revenue company that supposedly was crucial to finding bin Laden — they discovered that it wouldn’t work if they tried to do a totally automated big data system — the function of the computer system was to present human operators with actionable responses.
McIlwain: For those of you that like old movies, go watch a 1983 movie called WarGames. It’s on this theme. It’s got some politicization in it that’s a little bit goofy, but the movie itself is quite a good one for dealing with this theme that we still — and I think for a long time to come — will be relying on human intelligence even if it’s data- and it’s analytically-informed.
The last two decades have also really created an era of empowerment at the edge of our community, at the edge of our corporations such that through mobile apps, cloud, software-as-a-service (SaaS) and even these services that combine an app and something that happens in the physical world — take Uber as an example — we as consumers are incredibly empowered. How do you see that not only changing how technology continues to push innovation but also how it can change the role that government has and maybe limit the role of government — which is my hypotheses as we’ve seen with Uber time and again in cities where local jurisdictions have wanted to hold on to some of the power and revenue, and then the empowerment at the edges of our society have overwhelmed it?
Gilder: In general we see movement from hierarchies to heterarchies — that is, top-down control (TV is the tool of tyrants, as I described it) to distributed heterarchical systems with the control on the edge. That’s a general trend in technology. It’s the law of the microcosm, and I think that that will continue to advance and prevail and it will favor governments that allow it to happen. Governments that try to suppress creativity will become weaker and their influence will decline. Their greater immediate regulatory power will come at the expense of the prosperity of the country and its ability to function in the world economy.
Haney: I agree with George.
McIlwain: The genie is out of the bottle, as it were.
Bruce Chapman (chairman and co-founder of Discovery Institute): I have a strong sense of timeliness here, because it was 25 years ago that Discovery first got started, and George was instrumental in that; and Hance Haney — who was then on Bob Packwood’s staff — was soon involved — and a lot of other people here that I could mention. In fact, I want to mention one. This is Rex Hughes. He was one of our first interns, and he sort of told George insistently about this thing called the Internet. George was resistant at first and then he told the whole world — it’s a long story, but it’s a wonderful story.
Rex Hughes: …a late evening — I think in Aspen at the Progress & Freedom Foundation — a couple months before you interviewed Mark Andreesen for the legendary Forbes cover…
Gilder: Yep.
Chapman: Rex is now a professor of cybersecurity at Cambridge University — and he’s from here. And we have Carver Mead (Gordon and Betty Moore Professor Emeritus of Engineering and Applied Science at the California Institute of Technology), with us — is it Burien you live in, Carver? — the question is what is it about this area? And Matt — you’re invested literally and figuratively in this area — what is it about the Greater Seattle area and our role in what is going to happen in the future? What are we doing? What should we be doing to play a part in this better development that you all are describing?
McIlwain: Ten years ago we started investing in these things called cloud computing companies. In fact, 8½ years ago we hosted an event at a little less tony place up on Capitol Hill where Andy Jassy who runs the now $6.5 billion cloud computing business for Amazon talked about this thing called EC2 — which was an Elastic Compute Cloud — and that’s something that startups might want to try to see if it might be useful to them, and then I got up and made fun of venture capitalists. And look at it today, here in 2015. We have two of the top ten — by market cap — tech companies in the world in Microsoft and Amazon. They are the two market leaders in cloud computing — they both claim to have $6.3 billion in run rate revenue in cloud — that’s well over half of the market. I know they define cloud differently, but the main point is that they are big, big players. And for more reasons than one, I think we can rightfully say that Seattle is the cloud capital of the world. Now, what are the next things is I think the question, Bruce. My view is that, leveraging on cloud computing, the next really big area is data in what we talk about as dataware or the intersection of algorithms, math models, data driven formulations that combine with software programming and deliver realtime intelligent applications — either to a system (like spool me up another instance on Amazon because I need more compute) automatically; or if it’s a human-involved interaction, it makes a prediction or a recommendation to me of which as a human I can choose to take or not to take action. And of course based on what I do the system learns again and it becomes a continuous loop. And in a place where we have the Allen Center for Artificial Intelligence, professors like Jeff Heer and Carlos Guestrin at the University of Washington who are leaders in these fields, we really have a critical mass opportunity to be the leaders in dataware just like we are the leaders in cloud computing and I personally think that’s one of the most exciting opportunities amongst many.
Gilder: …Carver Mead is currently engaged in a challenge to Einstein’s general relativity, and it’s a fascinating cause and promises a major simplification of physics. That is happening also in the Seattle area, between Caltech and Burien.
Haney: …I think it’s something more than the fact that this is a nice place to live and the people are really nice here. I think entrepreneurs feel welcome here and feel supported here…as an Oregonian I wish I could bottle it and take it home with me.
McIlwain: There is a special cultural aspect here, and there’s also no state income tax which is also something you don’t have in Oregon.
Questioner: Two questions: We obviously are not fans net neutrality nor any more regulation, but is it possible to imagine — I know it’s difficult politically — is there any need for regulation at all of the telecommunications industry? Can we just get rid of the FCC? And then the other question: You know we just learned about this hacking of an airline recently, and you brought up WarGames. How much do we need to worry about Skynet super computer?
Haney: …Before the FCC designated broadband as a “telecommunications” service, it was under Federal Trade Commission jurisdiction and was regulated just like every other sector in the economy. When the FCC designated broadband as a telecommunications service and turned it into common carriage broadband was taken out of the FTC jurisdiction. So if this FCC net neutrality order were to be overturned by the court, you’re correct that broadband would still be subject to minimum regulation comparable to what is in effect in most other sectors of the economy.
Gilder: Peter Huber — one of the most brilliant people I know — has done a book on why the FCC can be abolished. The heart of the matter goes beyond his book. The idea that the electromagnetic spectrum is a sort of public commons or a natural resource that somehow is scarce and somehow resembles beach front property — and the peril of monopoly — is just plain wrong. The spectrum as it used in communications is created by engineers, inventors and creators. It’s an effect of lasers, masers, klystron tubes, antennas, computers — all these inventions of our age. And the idea that somehow it’s out there and the government should own it and auction it off to big companies — who then use various tranches exclusively as if new technologies of software-defined radios and other devices don’t permit the shared use of spectrum and much more efficient use of spectrum — it’s just wrong. The FCC drastically reduces the efficiency of the use of spectrum and thus results in a scarcity that requires network non-neutrality. So I think it would be very good to abolish the FCC.
Questioner: We have the example of how Comcast behaved with Netflix nine months ago — where as an intervening carrier not as a direct provider for Netflix — Comcast ecided to investigate and then essentially tax Netflix for that and Netflix ended up writing a check so that their consumers at the far end wouldn’t suffer this lag. So how does an open market deal with that — and this by the way is kind of one of those quintessential sort of things why people claim net neutrality is needed. So what’s the market response to that and will it be effective and fast enough to make sure the consumers — and in this case the provider — aren’t held hostage by someone who is screwing around?
Haney: As I understand it Netflix paid Comcast in order to improve the interconnection that Netflix got, and so it was for a service improvement. And personally I don’t have a problem with that. If broadband providers can make money at other places in the network I think that benefits consumers; that reduces what consumers have to pay for the service and that makes broadband service more affordable. In telecommunications we see that as service improves, cost decrease, usage increases and overall industry revenues increase — and so I personally feel that broadband providers should have flexibility to experiment with new business arrangements and that there’s nothing inherently wrong with Netflix paying Comcast to improve the quality of the delivery of their service.
Gilder: The whole Internet — the whole telecommunications — is full of service level agreements. Every business that buys bandwidth has a service level agreement that specifies the properties of the bandwidth and prices for various guarantees and features. And content delivery networks — Akamai, Limelight and Highwinds (where my brother works) and the like are in the business of offering better access for a price. That’s the reason I believe all these Silicon Valley people said network neutrality is meaningless, and so ridiculous an idea that we can deal with it — ‘it went through our lobbyists, and we can get waivers’… And when it translated into Title II all of a sudden there were expressions of ‘we didn’t mean that!’ All of the major people retrenched, and I think that episode is edifying and I think may improve telecom policy in the future.
McIlwain: There’s a whole common sense principle here. Does it feel like common sense that somebody should be charged more if they have a 10 gigabyte mobile plan every month, does it seem like common sense that somebody should be charged more for first class versus coach class in an airplane? It seems like there are some practicalities that overwhelm the net neutrality arguments.
Questioner: How does the Internet of Things — the idea that of all these different interconnected devices forming their own Internet — how does that play into, first of all, the evolution of the Internet as a whole, as well as what would some of the ideas such as net neutrality and the regulatory framework that we have now, how will that affect the Internet of Things?
Gilder: There’s a company — where is it, Carver, where is Impinj?
Carver Mead: Fremont.
Gilder: Near here, right?
Mead: Yes.
Gilder: Impinj is really a crucial Internet of Things (IoT) company. It does RFID. It’s based on a chip the size of a grain of sand that has a couple kinds of memory, a little processor, an antenna and it runs by incident radiation. It can be ubiquitous. It’s a really major breakthrough in semiconductors as well as systems. It provides a first step for a really ubiquitous Internet of Things apart from just connecting big electrical equipment plugged into the wall like refrigerators and TV sets or whatever. That’s a Madrona company.
McIlwain: That’s a Madrona company.
Haney: They’re also adding mobile wireless connections to things even like freight pallets now. Containers that are on ocean ships down to smaller… So it’s not just passive RFID, but it can also be mobile wireless connections so that there is a visibility into the supply chain from beginning to end. And that’s one example. Uses in the industrial area are absolutely huge. The medical devices that we so often read about as well that are going to be continuously monitoring our heartbeat and other vital signs so that the goal of which is that if we have a stroke or a heart attack or something like that it will have been predicted. The hospital will have been made aware that we’re on the way. We will have that incident in the hospital and not where we can’t be immediately treated. These are obviously vital functions. Also part of the Internet of Things. This would be an example, I would say, of where we would want that traffic to be prioritized. We would want very high levels of quality of service accorded to that traffic. The providers will have to go to the FCC, file a petition and get permission to do that. I assume the permission will be granted, but that is a bureaucratic process, that’s not permissionless innovation it is bureaucratic approval.
Gilder: It works in the Food and Drug Administration.
Haney: That’s what we’re talking about.
McIlwain: It’s really about the data that’s been able to be collected — whether it’s the heart monitor, whether it’s the sensor — it’s really about that data and the timeliness — i.e., real time — data and actionable things that you can do as a result of that that is the promise of the Internet of Things. On the regulatory side I think that you would see that people would cloak it in privacy if they were trying to make a case to apply the forbeared portions of the 1934 act apply to information services. That’s the risk that Hance is pointing out.
Questioner: You talked about Bitcoin, Venmo and things like that. When is this going to become more ubiquitous in a way that we’re actually using it to purchase consumer goods? It feels like it could happen at any time, but not.
McIlwain: This is the $64,000 question in venture capital. My college soccer coach had a great expression. He’s a Scotsman, and he would say, ‘ah lads, would you do the same thing only different?’ And it’s exactly that. It’s trying to figure out why things that somebody tried ten years ago or 20 years ago and failed are actually ready now. You take out this device — this Teleputer that George predicted 25 years ago — even ten years ago people were super skeptical about the smartphone. Here we have this incredible device with so many robust capabilities. So I think the exact timing is hard. I will say one thing just to throw out a specific area. I’m not as much an investor in the payments area, so I don’t have a specific answer on that, but if you had asked me a year ago if I thought that virtual reality was going to become a reality I would have laughed at you. I would have said I would rather be wrong than actually bet on that category…
Gilder: That’s why you didn’t invest in Otoy?
McIlwain: George pitched us on a related company. But what has happened is that the cost of all the components, and the volume at which those things are produced that work in this device have now created a set of experiences — and by the way, right here in this town the second largest office for Oculus, which is the company that Facebook bought is right here in Seattle. The best shot at a competitor to Oculus is a partnership between Valve — the makers of Steam and a whole bunch of really cool games — and their partnership with HTC. I’ve done that demo. That is unbelievable. It would blow you away, not just from an entertainment perspective but from a productivity perspective. And we could go on to Microsoft, and I have no doubt there are things going on Amazon, etc. So the whole areas of virtual reality and augmented reality — and again, Seattle, Bruce, to your question, as an area, that I think the time has come and it will be very exciting to see in the years ahead how that gets adopted.
Gilder: Otoy, incidentally, just made a deal with the National Hockey League to do holographic light field broadcasts of hockey games. That is dependent on a really robust, low-latency Internet, and that’s the kind of capabilities that future progress in this technology will enable — but which a regulated Internet can’t really accommodate.
Haney: According to one estimate ten years from now we’re going to have fifty billion connected billion connected devices with the Internet of Things. We’ll need devices that can be useful for ten years. There need to be improvements in battery technology for that, and the engineers are working on ways now where these devices — which are only transmitting little kilobyte streams of information — but they have to be as efficient as possible. Even though each device is just transmitting kilobytes at a time, there are going to be so many of these devices that it’s going to require a lot more spectrum — so that has to be accomplished, too.
Questioner: There are two technologies that really interest me: the autonomous car and drones. We’re talking a lot about regulation and about efficiency, and I notice that Amazon already has a better drone than the regulations for previous drones. I can understand we’re all very scared of drones because we’ve used them to kill a lot of people in other places. But how are we going to create technology and regulations that allow companies to actually iterate faster than regulations? Can we have a efficient and flexible regulations? As a millennial that sounds really normal, but maybe it’s not the same to others — to have flexible regulations.
Haney: It’s possible in theory, but we’ve never figured out how to do it right.
Gilder: We want simple, low-entropy regulations that can accommodate high-entropy creativity, and you want simple, regular regulations. Flexible regulations is noise in the channel — it actually obstructs creativity that actually results in greater opportunity and capability.
Questioner: Mr. Gilder, you mentioned about the focus ten or 15 years ago being wireline broadband. And then wireless, who would have thunk, right? Why do you think that was not foreseen? A misunderstanding of Moore’s Law, or was it just pure innovation…
Gilder: I think people didn’t understand the spectrum issue. They still think that spectrum is scarce. One of the companies I’m supporting now is an Israeli company called ASOCS that make software defined radios that allow you to move the baseband processing in smart phones and cell phones away from the base stations to the cloud. This will allow you to have very simple antennas without environmental conditioning and a lot of computational capabilities and specialized functionality and have huge multiplication of essentially coat hanger antennas anywhere. Each of these antennas has access to the whole spectrum. So there isn’t any limitation really. But there is a belief — I mean, if you fixate on the existing technologies and imagine that spectrum is something that the government allocates, you can imagine that it’s scarce and that it can’t accommodate the kind of broadband world that we’re currently moving into.
Questioner: So you’ve made a very good case that regulations could be a great limiting function on the potential of the Internet. So as we look at our algorithms and applications of the Internet — health care, transportation, education, etc. — every one of these areas has a regulatory component. It’s not just not the FCC. The FDA is going to be involved in medical applications, telemedicine. Transportation — Uber — being subject to all sorts of regulatory things. So the question I have is, if this is the problem — about creating potential or enabling the potential of the Internet — what are your suggestions on a practical method for trying to manage a more effective regulatory environment? How should that be organized? Who should get behind it and how should it be financed? How do we make it happen?
Gilder: Mancur Olson wrote a book about the rise and decline of nations, in which he showed historically how nations rise and they create increasingly complex and burdensome bureaucracies and finally fall into a kind of sclerotic condition where they can no longer grow. Then they decline. We’re in that position right now. Other countries, however, and the United States has been in this condition before. The example I like to contemplate most is after the Second World War. All the leading economists said that unless the United States maintained its government spending at the same levels that it maintained during the war — as Paul Samuelson predicted — we would experience ‘the worst dislocation and depression in the history of economics.’ Other economists had similar visions. Instead, the people had a wild, irresponsible moment and they elected a troglodytic Republican Congress in 1946.
Did it leave government spending at the same level? It reduced government spending 61% and laid off 150,000 regulators. It dismantled most of the price control bodies and panels and commissions that had emerged during the Depression and through the Second World War. It laid off a million government workers. It cut taxes in half by establishing the joint return, and it also cut all sorts of taxes on businesses. The result was not the worst depression, it is what we now see is a golden age. And it saved the West. The U.S. didn’t decline as it surely would have if had retained the apparatus that nurtured the stock market crash into a Great Depression.
There’s a great new book by Jim Grant — who is a wonderful writer — about the crash of ’21-’22, which was just as fierce as the crash of 1929. But Harding and Coolidge were in office, and they raised interest rates and cut government spending and did everything wrong from the point of view of any conventional economist. And that crash led to no depression; it led to the roaring twenties. Countries around the world reach a point where they deregulate. New Zealand, I tell that story in Knowledge and Power. New Zealand just zeroed out its government. It had gone from the third richest country in the world to a third-world country that couldn’t even feed itself despite being an agricultural center. New Zealand went to zero-based budgeting and eliminated its Department of Agriculture and became so productive in dairy that it elicited unfair trade practices suits from dairymen in Wisconsin. It became one of the biggest agricultural exporters in the world. It can be done. You wipe them out and start again. Peter Drucker said this: every ten years all regulation should be gone in a jubilee. You eliminate them and then you judiciously regulate as needed.
McIlwain: Well on that note I think we want to thank Hance and George for being our panelists tonight.