The following is a transcript of Randy Picker’s conversation with Tim Wu on how antitrust shaped competition and innovation in computers and chips, held at the 2024 Stigler Center Antitrust and Competition Conference.
Filippo Lancieri
We’re about to start with the first keynote of the conference.
Good afternoon from Chicago. Thank you all who are joining us online and here in person. My name is Filippo Lancieri. I’m a post-doctoral fellow at Zurich and a research fellow here at the Stigler Center. This afternoon, we are very happy to host a keynote with Professor Randy Picker on “The Quest for Next: How Antitrust Shapes Competition and Innovation in Computers and Chips,” in conversation with Tim Wu.
Before I begin, remember you’re on the record and we’re going to post this on our YouTube channel afterwards. As usual, the views of Randy and Tim are not those of the University of Chicago. We hope that you’re going to join us for more events online, so check our website and also check ProMarket.org and the Capitalisn’t podcast, which are part of the Stigler Center.
Now let me quickly introduce our speakers.
Randy Picker is the James Parker Hall Distinguished Service Professor of Law at the University of Chicago, and (other than being a great PhD supervisor) he also received his BA, his MA, and his JD from the University of Chicago, and was Associate Editor of the University of Chicago Law Review. His teaching focuses on antitrust law, intellectual property law, competition policy, and regulated industries. He’s the author of the forthcoming book The Quest for Next, which this keynote is going to be derived from.
Tim Wu is the Julius Silver Professor of Law, Science, and Technology at Columbia University. He just recently served the White House as special assistant to the president for technology policy and competition from 2021 to 2023, among many other public positions. He also teaches and writes on power, antitrust, and the growing power of Big Tech platforms. He is the author of The Curse of Bigness: Antitrust in the New Gilded Age.
So, without taking much of your time, Randy, the floor’s all yours. We’re very happy to host you.
Randy Picker
Thank you so much. Thanks for the invitation. What a great setting. I really appreciate the chance to talk. Tim and I are old friends. We taught a seminar together years ago. I’d teach a seminar with Tim every year if I could, so I’m thrilled to have this chance to do this with Tim.
When I agreed to do the talk, I wasn’t quite sure what the conference was going to be. I knew we were going to chase Chad’s question. Then I saw the agenda, and my reaction was, “Oh my God, they’re going to have said already everything I want to say!” That made me a little nervous, but now I think where we are is…there will be a nice mix of agreement and disagreement. Then we can try to sort that out.
So, how to run at Chad’s question? I think there are a couple ways to run at Chad’s question. One way is hardcore empiricism, the type of things that Tommaso was talking about on the first panel, the kind of work Simcha does, and others. That’s not my comparative advantage at all, so I won’t try to go down that path.
The other one is to focus on the technologies that produce growth. This GPT technology, what I now call the old GPT since we’ve got a new GPT in town. Those technologies are the places that growth, enormous growth occurs in the economy. That’s a rich, complex literature. Chad has participated in that literature. Looking at how antitrust might shape those technologies strikes me as a sensible path for us to pursue. That’s what I think we’re doing with regard to these case studies in the morning and then the rest of the afternoon. So I’m thrilled to have a chance to participate in that. If you look at these companies that we’re talking about over time, they have at various points in time been the most valuable companies in the United States or in the world. If we want to see where the money is, if we want to see where value is getting created, these seem to be the right companies to look at.
This first slide, as you can see, goes back to IBM and AT&T in 1982. If you go to the current era, the dominance of the tech firms—Microsoft, Nvidia, Apple, Alphabet, keep going down that list—these are the firms that are the most valuable firms on the planet. Understanding those firms, and understanding the technology associated with them and then how antitrust might play into that, that’s a great topic. And that’s why we’re here.
I want to go back to transistors. We heard about transistors on both panels this morning. I think that’s a richer story than we’ve seen so far. Here’s where we start.
Two Bell Labs researchers, John Bardeen and Walter Brattain around Christmas of 1947, discovered this thing that they will eventually call the transistor. They know what they have. Bell Labs, AT&T, goes to get patents, and then they go public with this invention. This slide is the paper they published [“The Transistor, A Semi-Conductor Triode,”] in the technical literature, the Physical Review, in July of 1948. The title of the paper was, I think for everyone, really the tip off as to what they had done and what they had accomplished.
This word transistor? Completely new at this point; they made it up, internal Bell discussion: “What should we call it?”
Semiconductor, well, that’s a… We’ll need to go to the periodic table of elements and look around a little bit.
But the word that was really the powerful word here is the word triode.
The triode was invented by Lee de Forest in 1906. You need amplification in these industries. The early radio industry and the long-distance industry needed a technology of amplification. Make signals bigger, make them louder, make them go longer. The triode was the great initial device to do that.
A triode was instantiated in vacuum tubes. Vacuum tubes were an enormously successful general purpose technology, maybe. Radio emerges. The early era of digital computers. John Atanasoff in the United States at Iowa State in roughly 1940. The Great British cryptography machines, the Colassi, during World War II. And then ENIAC sort of at the edge, in 1946. All of it situated in vacuum tubes.
These guys say, “We’ve built something that is a much better device that will replace the vacuum tube.” That’s what they saw. “It will use less power. It will be less brittle. It will do all the things that the triode has done for us for decades. We’re going to replace that with a superior technology.” They knew they had that.
Now, over here, it’s interesting. They thank someone else, a gentleman named William Shockley who ran the lab. Shockley is a difficult, tortured soul over time, if you know Shockley’s history. Shockley was thrilled that the transistor had been invented and frustrated that he wasn’t the one doing it. So, Shockley sort of goes off by himself for the next month, and he invents the next transistor. The first transistor we call the point-contact transistor. The next one is the junction transistor. That’s what Shockley goes off and invents.
There’s a footnote here that’s sort of interesting, too. They talk about the effects they’ve seen, found in both silicon and germanium. All of the work they’re doing right now is basically in germanium. If you have full mastery of the periodic table (as I assume everyone does, right?), germanium is sort of here, and silicon’s right above (I had to go look; I don’t do that kind of thing). That’s where they are. So, they’ve got both of these elements, but right now they’re working in germanium in 1948.
This immediately gets a lot of attention.
We talked on the first panel about diffusion of technology. This slide is a 1951 letter from Bell Labs, in response to a letter that came in from a gentleman named Olsen at Geophysical Service. Geophysical Service would eventually rename itself as Texas Instruments. They want to go into a transistor business, so they write to Bell Labs and say, “Look, what can you tell us? What can you send us?” And they say, “Well, you know, it’s really early stages. Go look at the Physical Review. Here’s the press release. Western Electric will be back in touch eventually.”
But what that really meant was that Bell Labs started licensing this technology in 1951. Bell Labs had a general policy of doing lots of patent licensing. I don’t know what to make of that. Regulated monopoly, they don’t have the normal incentives, so I don’t know what to do. But they had hundreds of licenses.
Perhaps more relevantly here, they had obligations under military contracts. They’d signed a bunch of military contracts. They were a huge military contractor. The military was effectively creating a kind of common pool in this technology through these contracts, which bound all of their contractors.
This book, I have a copy of this book at the law school. This is known to the trade as the $25,000 book because you got the book plus a license to do the IP. (I did not spend that much on my copy.) This book says November of 1951. And then as you go inside the book, they say, “Look, we’re doing all of this to get the technology dispersed to aid the military effort. The Korean War.”
So this technology is spreading long before the 1956 final judgment. There’s no question that the 1949 lawsuit has been filed, right? So we’re always sort of dancing in the shadows of antitrust. That’s tricky, I think. But they clearly had these obligations under these military contracts to do this licensing. And they seem to do a lot of his licensing generally, anyhow.
A publication of the era, Radio Electronics, says, “We’re studying the market. Here’s what it looks like in July 1952 if you want to go buy transistors.” Western Electric is making 6,000 point-contact transistors a month, selling almost all of those to the military. The military wanted this, obviously, for mobile communications. Hugely valuable for that. Shrinks it smaller. Raytheon’s making them. RCA’s making them and is ramping up. GE’s making them. There are lots of companies that are making transistors. Again, this is July of 1952. This technology is spreading pursuant to the licensing and know-how transfers that Bell Labs is doing pursuant to these.
We have a next step forward.
There’s a conference in Dayton Ohio on transistors in May of 1954. A gentleman named Gordon Teal, who used to work at Bell Labs, has moved to TI, Texas Instruments. People in the morning of the conference say, “Silicon transistor? It’s a few years away.”
Teal’s talking at the end of the day. He says, “I’ve got a few silicon transistors in my pocket. Let me demonstrate them to you.”
Silicon deals with heat much better than germanium does, so that’s what made the silicon transistor attractive. So, the next step forward, or I should say, the next public step. After the fact, we know now that Bell Labs actually had some silicon transistors but had not gone public with those. But we moved from germanium to silicon (it’s called Silicon Valley not germanium Valley for a reason) in May of 1954.
Then the Wall Street Journal has a front page story in May 1955 about the transistor, the devices it’s moving into. They talk about the possibility of a Dick Tracy radio transmitter. I guess we know what that looks like. And the units here are millions of transistors being produced by roughly 18 companies.
Again, this is all before the 1956 final judgment.
IBM says, “We’re going to build computers next year using transistors. Those transistors will be faster and use less power.” If you’re watching (as I hope you are) the world of AI and chips, the chips that are rolling out week by week, where everyone’s chasing Nvidia, where Google releases new chips, and Google and Apple are releasing new chips. All of those chips are about more computation and less power. That’s where we were in 1955 as well.
Eventually we do get front page news. The AT&T settlement. But what I want to establish, I hope, is that it comes later. It comes after we’ve created this world.
Indeed, the relatively contemporary criticism of this settlement, as to what it does, is that AT&T was already doing this licensing, so what kind of give is that? Yes, it blocks them from going into other businesses. But did they want to do that, anyhow? So, they had to stay in telecommunications. They also could continue to work for the military. DOD would never have blessed this deal if that hadn’t been happening. The criticism of the 1956 settlement is that it sort of didn’t accomplish that much, given what was already in place and given what AT&T otherwise wanted to do.
The next day they did a settlement like that with IBM. Tim was talking earlier about industrial policy. Carl asked if they have the power to do that. This is exactly what they’re doing here.
This is really sort of a visionary DOJ that sees that we’re on the cusp of the next. We’re on the cusp of next. That’s the electronic industry that is going to come. And they have the power to create a kind of common pool in the patents of that era. Not necessarily because they knew that when they filed the suits. The AT&T suit is filed in 1949, the IBM suit in 1952. But when they get there, they get to this point and they settle to take advantage of the moment. Fiona asked this question on the last panel: What about the current cases? I don’t have a clue. We could do that. But what I want to say is, settling those cases is not an embarrassment. Settling cases, and changing things now, given the pace of technology. That’s so important. I think that’s exactly what the DOJ did at this remarkable window in 1956.
There is work that looks at some of this. Nagler, Schnitzer, and Watzinger have a couple of papers. There’s a fourth author. I think you referenced the broad paper. They have a specific paper on the technology of the transistor. They say, “Look, there was a lot of licensing going on. We don’t think anything happened as to the transistor with regard to the 1956 settlement.”
Okay. That’s point one.
Point two. I want to go to software.
The IBM System/360 in 1964 is a remarkable machine. The first computer platform. I think someone said that. IBM’s computers did not talk to each other, so they were incompatible. It’s not a question of them talking with their competitors and that kind of interoperability. They didn’t talk to each other. IBM spends a huge amount of money to try to run at a new platform. That’s what the System/360 is. It’s got different levels. That turns out to be remarkably successful. That’s what gives rise eventually to the 1969 suit that we heard about on the last panel.
I want to show you three developments in software. Then our job is to try to disentangle them a little. That’s what I want to do.
One development is, IBM is engaging in bundling, where you get the software and the services and the hardware all together. Their mainframe competitors were doing that as well. I don’t know what to make of that. It was not distinctively the leading firm in the industry that was doing it. Others were doing that as well. That had the effect of creating a kind of common pool in software. Software is sort of a zero marginal cost good. Everyone freely sharing software? It sounds like open source. That doesn’t sound terrible, right? That’s the world that was getting created.
If you were outside that industry and wanted in, you had a problem. Marty Goetz said, “I know what I’m going to do. I’m going to get a patent on my software. That’s what I’m going to do. When I get a patent on it, then IBM can’t just scoop it up. I can proprietize it.”
This slide is his patent, April 1968. The leading journal of a day said, “Full patent is issued for software. Full implications are not yet known.” Yes. They had no idea. Right? Software patents, if you do this kind of thing, there’s a lot of criticism of them in a later era. But in this era, they start as a way to allow entrants to come into an industry where they face the problem that the incumbent was effectively scooping up all of the innovation. They wanted to proprietize that innovation, and so, let’s go to patents to do that.
That’s development one.
Development two we’ve heard about. IBM is facing potential antitrust action. They know this. In December of 1968, they announced, “So, we’re thinking about unbundling.” Right? So again, how do you tell? Would they have done this independently? I saw a reference to the Fischer book. The Fisher book’s fantastic. Frank Fischer was the IBM consultant. I think he says he bought a boat off of IBM. [inaudible comment] Is that what it’s called? That’s fantastic! I love that! Okay. In any event, he claims of course that they were going to unbundle anyhow. How do you tell? The complaint very much focuses on these bundling issues. And, lo and behold, after the suit was brought, 6 months in, they unbundle. Not a lot. They say, “We’re going to take down hardware prices by 3%,” so I don’t think the industry was thrilled by that. But they’ve unbundled.
Marty Goetz gets a patent.
IBM unbundles.
Those are two effects, and then we have a third. The third is this. This is known in computing history circles as the trillion-dollar magazine cover. I own a copy of this as well that I bought on eBay that I have in my law school office. We are staring at Popular Electronics for January of 1975. The ALTAIR 8800. They call it the “world’s first minicomputer kit.” “Kit” should frighten you. They would send you the pieces if you wanted, and you could build it on your own, as a true hobbyist would. They’d also charge a little bit more and build it. Critically, in this is an Intel microprocessor.
Paul Allen sees this magazine in a bookstore in Harvard Yard and goes to his friend Bill Gates, who’s a Harvard undergraduate. They’d worked together on software projects back in high school in Seattle.
Allen says to Gates, “This is it. This is the moment. We have to move.” Have to do what? Have to create a version of BASIC for the ALTAIR. Well, BASIC was created at Dartmouth in 1964 on a General Electric computer, okay? There’s no sort of IBM path there. Gates and Allen set to work to create their version of BASIC, which they eventually sell for this device.
On the west coast, this magazine is being passed around at the Home Brew Computer Club. The Home Brew Computer Club is where the young people who are interested in these kinds of devices are meeting in the Palo Alto area. And someone sees this named Steve Wozniak, and he says, “I can build a better version of this.” Wozniak then, with his friend Steve Jobs, builds the Apple I, eventually the Apple II, eventually also creates a version of BASIC for the Apple II.
This is where personal computing, in some sense, starts. It starts, importantly, because of the effects on Gates and Allen and then Steve Wozniak.
But if you ask what actually makes the Apple II a success, and Steve Jobs is crystal clear about this, he says it’s when VisiCalc was released. VisiCalc is the first spreadsheet. This slide is an ad from Byte magazine, September 1979. Jobs says, “Our sales just completely took off.” Finally. Programming in BASIC? You’ve got to build your own software? Yeah, okay, that’s a delight for the Wozniaks of the world, but not for the rest of us. This was a program you could actually just run.
Lots of people study these industries. I don’t have a handle on exactly what happened, but Martin Campbell-Kelly has a paper from 1995 where he looks at software revenues. Here’s what he thinks happens. He says that runs from roughly 1964 through maybe 1985, he says there you can see a couple changes in slope, and then a sharp inflection point. Okay. My suggestion is (I don’t know this) that the creation of the software industry is really driven off of the creation of the personal computer. That has basically nothing to do with the IBM unbundling. Yes, there are some slopes there and level changes. Yes, absolutely, we’ve got to disentangle the Marty Goetz patent. But I really think it’s the personal computer that drives this.
I want to talk about the IBM PC. That’s a great story, too.
IBM’s watching what’s happening with Apple and has to figure out what to do. You could do presentations without PowerPoint in the old days, and you did those with these very large charts. This slide is the presentation that Bill Lowe at IBM gave to the IBM corporate management committee on August 8th, 1980, to say, “Should we do the IBM PC?” It’s a pretty interesting thing to look at. If you look at it right at the top, you’ve got this question of closed versus open system. To what extent is IBM going to open this up, and how would they do that? You see VisiCalc over there. You see Microsoft BASIC. Microsoft was a languages company at this point. IBM hardware. Down in the corner, you see processor 8088. They’re going to use an Intel processor.
They’ve got a one-year timeline. IBM has had a poor record with smaller computers heading into this. They’re not confident about their ability to execute on their own, so they sort of do this outside of the normal chain. They did have their own processor that they could have used. That’s a fantastic story. (There’s this whole world of RISC vs. CISC processors that, if you know what that means, you do; and if you don’t, you don’t. You could read a book and find out. It’s a great story. An important story.) They had a processor that they could have used, but they didn’t.
Then the question is, what operating system is going to run it?
They go to Bill Gates and Paul Allen. They’ve already talked to them about getting computer languages. And they say, “Do you have an operating system?”
Bill Gates says, “No. You should go talk to this guy up the street. He’s got an operating system, not us. Gary Kildall. CP/M.”
There’s many stories about how that particular meeting goes, but eventually they work their way back to Microsoft. And when this computer is released, it is released with three operating systems.
Audience member 1 (Chad Syverson)
You’re not going to say the thing about him being on his airplane or whatever?
Randy Picker
Supposedly, Gary Kildall was on an airplane and flying, and there are stories like that. Gary Kildall has an unfinished biography. You can go read his story there. Okay, but in any event, make sure you see the critical point: IBM as we think of it, as offering the keys of the kingdom to Gates and Allen, something that would make them billionaires.
And they say, “No. Go somewhere else.” So that world changes.
The release of the IBM PC was huge. I think all of the available evidence (and now I’m completely on the same page as Tim was) suggests that the pendency of the 1969 case shaped this. Bill Lowe, who did the chart that I showed you, he says that. Paul Allen says that in his autobiography. This is a more recent book that I know you reviewed for the New York Times. There’s a quote there like that. I don’t know what to make of that, in the sense that… Lowe even says that the 1956 case continued to matter. The final judgment that was pending there.
I think the PC is opened up in a way that it would not have been, absent the pendency of this antitrust case. When I start to work my way back to Chad’s question, and GPTs. If the PC opens up faster, and if the PC rolls out faster, and antitrust matters for that, how big a deal is that for the economy? I don’t know; that’s not my area. But like the last panel, I hope that it matters.
Then there’s a different way to run at this.
After the case is dismissed, how does IBM behave when that threat is gone and they change their behavior? IBM buys a chunk of Intel in December of 1982. Intel’s sort of in trouble at this point. Why is Intel in trouble? Why Japan won the chip race. Intel had effectively invented silicon memory, and Japan has just completely overrun that industry. A great concern that that’s what’s going to move, will go to logic.
These are geopolitical conflicts. Across the course of my book, I would say there are four. There’s obviously the World War II era. There’s the Cold War era. I doubt we get integrated circuits at the timing we got those (that’s not an antitrust issue, it is the government playing a role) if it weren’t for the case that NASA and the military was willing to basically pay anything for integrated circuits. You get this conflict with Japan. And obviously in the current era with China.
IBM, freed of the pendency of the 1969 case, makes an investment in Intel.
There is something that I pulled out of a slide that I’ve only seen once. James Cortada has a history on IBM. He worked there for 30 years and also has a PHD in history. His claim is, in 1986, Microsoft offered to sell a chunk of Microsoft to IBM, and IBM said no. I don’t know what to make of that.
Then there’s more. The more is that, as the world of clone PCs emerges, and IBM loses control over the platform that they created—and maybe felt they had to lose control over because of the pendency of the antitrust case— freed from that, they try to lock down the system. When they roll out the PS/2, it’s got a microchannel bus inside it. Heavily patented. That’s how they hope they’re going to change the industry and shift it to that.
They also try to do that with software. They’ve got something called OS/2. This slide is a press release with Microsoft and IBM in 1989. You had to walk away from that press release thinking the future, the next, was OS/2 and not this program that gets mentioned at the bottom of page two: Windows.
None of that played out. IBM was never able to get it back once the genie was out. I think we owe that to antitrust.
I want to talk, then, about a different angle on these questions, and what I’ll call the long reach of the 1956 AT&T final judgment. We took an incredibly capable company off the table in not letting them compete in computers. Would they have been good at that? Subsequent history suggests, no. We can talk about that. But IBM faced less competition because of the way we quarantined AT&T.
The Unix point that was made earlier? Absolutely. Unix is a huge, important story. There’s a lot to say about it. Not today, but it’s absolutely an important story. The fact that Unix is awkwardly situated in the 1956 final judgment clearly matters.
What if we had the internet in 1980? Well, that’s sort of an odd thing to say (though the French had Minitel, if you know that, and we look back at Minitel and sort of giggle, but I think to the French it was a wonder in its era). AT&T had a device that they were calling the teleprocessor. They had a clear killer app for it: electronic Yellow Pages. They wanted to put that into the home. They got blocked. Newspapers didn’t want them in, but they also face the barrier of the 1956 final judgment. Again, a situation where we’ve taken someone off the table. We could have maybe moved forward faster. We didn’t. So, it’s not merely observing what did happen. It’s also about observing what didn’t happen.
I don’t think we get the 1982 consent decree, a consensual deal, absent quarantine from 1956. AT&T wanted to go into the computer business. They saw computers and communications coming together. They desperately wanted to go in to compete in that market. They didn’t seem to care very much about the local operating companies. Next, the future, was this combination. To get free of that, they wanted to go in. What they didn’t know, of course, is that they would get free of it and then go into that market and lose billions of dollars, multiple times. They turned out to be a terrible competitor in this market. But you don’t know. How do you tell?
All right. One more and then I’ll stop.
The 1982 consent to modified final judgment is amazingly silent on wireless. It’s just not in the vision. But the consequence of breaking up the company into AT&T and then the seven RBOCs—Regional Bell Operating Companies—means that, at the point where we got to the kind of spectrum auctions that Bob was talking about in the last panel, we had more competitors present to compete in those auctions. Again, how does that matter? I don’t know.
I look at Chad and I keep saying, “Have I persuaded you Chad yet?” No? Okay, you don’t have to say anything. But those are the kinds of things I think we need to talk about. I could do another half an hour on other subjects, like the ones we’re going to talk about this afternoon and others. I think there are tons of topics to talk about in this space. It’s a fantastic space, obviously. But with that I’ll stop.
Tim Wu
Well, first let’s thank Randy. I want to mention that you went 26.9 seconds over 30 minutes. That was pretty impressive! Very impressive display for that.
Let me ask you a few questions, and then we’ll open it up for further discussion. I think you’ve been working on this book for a period of time.
Randy Picker
The joy of having a job that I can do that. We are at 320,000 words. Yes, exactly, now Chad wants to make faces.
Tim Wu
So I shouldn’t have called it a book. The volume that you’ve been working on for a period of time, I wonder, going into it you may have had some priors or whatever. I wonder if reading the history… History always makes you think that you can’t say anything about anything, sometimes, because everything is very complicated. A lot of people. But I wonder if you can sense any way in which your views or priors you had were different. Any change you experienced during the research itself.
Randy Picker
That’s a really good question. I think I had a conception that if you had a patent, that really meant something. I don’t sort of think that anymore. A patent’s an opportunity to eventually bring a lawsuit to maybe get paid. Obviously, there are situations where the ITC will block things at the border.
I don’t know if it’s cultural. So, try this. I didn’t say this, but let’s talk about this. Did the 1956 settlements create a kind of culture where people said, “All right. I know you’ve got patents. We’re going to build some stuff. We’ll see what matters and doesn’t matter. No one cares if you build products that don’t go anywhere, so let’s go find out who’s going to win. And then we’ll settle up afterwards, and maybe you’ll have patents.” There’s obviously a huge era.
There’s a patent fight that emerges between Apple and Samsung in 2012. That patent fight, I think, is hugely important. Not because of the patents, but because of the fact that Samsung was a supplier to Apple, and all of a sudden Apple’s getting its chips built by Samsung. And they go, “Maybe we should find another company to build those chips.” And that company’s TSMC. The fact that all of those fabs are along the west coast of Taiwan, and 250 kilometers away from China. There’s a decision that gets made that, maybe because… yeah, okay.
Tim Wu
No, no, that’s great. I’m going to sort of continue that along that line. Some of these ventures…
Randy Picker
Sorry. I should say, Samsung has fabs in Austin. The whole geopolitical issue here looks very different if all of the Apple chips are getting built in Austin.
Tim Wu
Following up on this culture question. Are there at times, reading the history carefully, ‘50s, ‘60s, ‘70s, where you find firms behaving in a way quite different than we might expect now? Do you think there’s been cultural changes? Was there stuff that was surprised you when you thought about that? You just mentioned culture.
Randy Picker
I did, yeah. I don’t know. I know that Brad Smith, who is the current Microsoft president, I think he thinks that the Microsoft case (we’re going to hear in the panel next), that that case changed the culture of Microsoft.
Bill Gates says this isn’t about culture. Bill Gates says, “Because of that, we perform poorly in trying to build cell phones.” I’m not sure I buy that, but that’s what he says.
Tim Wu
Although he certainly, certainly felt it.
I wonder if also, reading this history… Remedies was a discussion, a big topic in the last panel. You spoke at some length about the quarantine remedy. Has that at all left you thinking, maybe we should think about quarantines more often? You know, if you do it right? I was going to link it to… Well, maybe I’ll just ask you that. Has it made you think more? I don’t think quarantines are something that were particularly in fashion for, let’s say, 1990 through 2016.
I just wonder if you have been led to a different view of quarantine as a remedy.
Randy Picker
The AT&T situation is specific because of the fact that you’ve got these regulated and unregulated entities fused together. And so, all of a sudden, as you know, when… Well, I assume you say in the books. There’s obviously this concern that they’re effectively subsidizing the unregulated part with the regulated part.
I think the vision of the quarantine there is possibly sensible in those industries, so I tried to highlight what I saw as the costs of that here.
If you said, “Oh, should we quarantine Apple from something?” Well, they’re not a regulated industry, so I think the quarantine idea is less obviously valuable there.
Tim Wu
Well, maybe I’ll push you and say, if you look back and you turn the clock back, even a little. We talk about AI as an opening and developing industry. Is there a case we should have quarantined Microsoft? You know, if we had the powers of equity that I was referring to earlier?
By the way, you notice in the 1956 decree there’s not, well, I know there’s settlements. But there’s a lot of questions…
Randy Picker
They’re settlements, exactly. I heard Carl’s question, but they’re just settlements.
Tim Wu
Well, if they thought they were totally…
Randy Picker
Of course, and there’s obviously [indecipherable] subsequently and all that, but…
Tim Wu
Sorry, that’s just a little aside that I think maybe no one knew what the hell we’re talking about.
But on this quarantine, you know, in a different version of history. Should we have been more aggressive in trying to quarantine Microsoft and Google from AI, if AI is next? Would that have made some sense to someone operating in 1956? Saying, you know, let’s try to keep these big, quasi-, they’re not quite regulated monopolies, but these big guys from trying to, own the next thing? Or is that one step apart?
Randy Picker
I gave chapter 19 in the book two weeks ago today at the law school. Part of that chapter is a litany of Microsoft failures. It’s just extraordinary. What? This is a company we should be scared about? Their position on the desktop has remained largely unchanged. That would be true of IBM with regard to mainframes. But they do a terrible job in their actual market. Like, so Vista’s terrible. Windows 8 was imaginative and a failure. They completely fail in the smartphone market. They’ve got their own software there by Nokia.
Billions, billions, billions, all down the drain. There’s this sort of the Steve Ballmer era that is a dark era for Microsoft. It’s only when Nadella takes over that all of a sudden things start to change. If you look at their market cap, it just is basically flat until 2015. So, would you have said we should have? Who would have thought Microsoft mattered even?
Tim Wu
It was a good period for the Seattle Seahawks, so maybe that’s related somehow?
Randy Picker
Well, and he obviously buys the Clippers, so it all worked out nicely for Steve.
Tim Wu
There’s a way in which Microsoft in that period does almost feel like a trust fund person who has this guaranteed income and can do anything.
Randy Picker
But trying. My vision of the trust (I shouldn’t say this) of trust fund people is, they don’t do anything. Microsoft is trying desperately. And failing.
Tim Wu
Well, let me double down on the point. I agree that maybe Microsoft’s not scary, but let me talk about Google then.
If you talk about quarantines, a big case. If DOJ said, “Listen, Google. You’re going to lose, but we’ll offer you a settlement, which is: You stay out of AI. You agree to divest all your AI stuff. And we’ll let you go for all these cases.”
Do you think that would be a good one for the country?
Randy Picker
So, a couple things in that.
The most natural settlement of the Google case at some level, because the heart of that case, as people almost certainly know, is this deal between Google and Apple so that Google is the default on iOS. As we’ll hear tomorrow morning, Europe has now got the DMA. The DMA means that, when you turn on a new device, you face these three choice screens.
The natural “solution,” as it were, into the Apple case is to say, “Okay, Google. You can’t pay Apple any money, and we’re going to set up a choice screen for search.” Yes, exactly. Do you think that matters in the slightest?
Tim Wu
I would say the deal is…
Randy Picker
No. The answer is no, right? I’m playing, but that has to be no.
Tim Wu
I mean, I don’t like remedies where the ultimate result is consumers or people click on more stuff for a little bit of so slam on…
Randy Picker
Even Europe had limits, right?
Tim Wu
You’re kind of not answering my question, which is to say…
Randy Picker
I am, I think. Well, go ahead.
Tim Wu
That doesn’t sound a good remedy to me either. I’m asking, would it be a better remedy…
Randy Picker
I’ll say no. I’ll say no. I’m happy answer your question. I’ll say no.
Tim Wu
…you know, “We trade all of our cases. Also the Ad Tech case. We’ll trade all these cases. You can have the markets of the past.” I’m drawing on Fiona’s stuff a little bit here. “Grand deal. You get all the markets of the past. You can continue your search advertising monopoly. But divest all your AI stuff, create another competitor. And you stay back there and let the next…”
Randy Picker
Great. So tell me what you’re trying to accomplish there. What I would worry about is that we’re taking a terrific firm off the table in creating the next thing.
Tim Wu
We’re trying to…
Randy Picker
What we should worry about is, this is I think part of the discussion, is whether they have something in their incumbent position which lets them—this is your adjacent market or next—distort that next competition. Right? That we should worry about, but I’m not hearing the linkage between, necessarily, Google’s current position and that distortion.
Tim Wu
Well, I’m trying to recreate the 1956 consent decree which you prized. And I guess what I…
Randy Picker
But they had a vision there that was driven by the fact that we had regulated and unregulated businesses—we would have, had they gone into computers—fused together. And the government didn’t know how to regulate that.
Tim Wu
Maybe replace the word “regulated” with “We have a dependable monopoly and we can do all this other stuff.”
I mean, it might have been a favor to Microsoft if, frankly, they had been quarantined out of everything. You know what I mean?
Randy Picker
Yeah. “Stop us before we buy Nokia!”
Tim Wu
Or something like that.
Randy Picker
Yeah, I get it. I completely get it.
Tim Wu
And had been quarantined out of everything. It’s a vision that says… Similarly they, in 1956, AT&T was a very impressive company. We didn’t let them get into what was obviously a promising market. In a way, they weren’t so different today from Google. So that you’re going to take that loss. And I’m arguing maybe that would be a good start right there.
You think they’re good?
Randy Picker
I think we can talk about how they’re doing. That’s a different discussion, and obviously all the image problems that Gemini’s had. But push that to one side.
Tim Wu
By the way, we’re going to leave time for audience questions. We’re going to go 10 minutes over.
Randy Picker
You’re watching the clock?
Tim Wu
Yeah, I’m watching the clock.
Randy Picker
Okay.
Tim Wu
So, one… although I suddenly forgot my other question.
Randy Picker
That was my move.
Tim Wu
Do you find, when you studied the, you know, you spoke highly of the antitrust, the Justice Department in the 1950s, which seemed to have much more of an industrial policy. Like, they realized where things were going, and they were thinking…
Randy Picker
I think that’s what they did, boldly. Boldly.
Tim Wu
And were bold. Do you… It’s kind of a pointed question.
What do you think was different about that Justice Department? Were they, in kind of contemporary language, thinking more about ecosystems as opposed to markets? What was different about the way you think they were thinking back then?
Randy Picker
Okay. I want to take a step forward, in the sense that, Giovanna has one in her piece that I think people don’t pay enough attention. She wrote a piece in ProMarket—you should go to ProMarket—mentioning the FTC settlement with Intel regarding the GPU. That’s 2010, I guess. I have a chunk of this in a chapter.
I think the FTC gets a lot of credit for recognizing—it goes exactly to what we were just talking about—that Intel was in a posture where they could affect competition in this new adjacent market and distort it. The FTC jumps in—this is Bill Kovacic, I think—in a very smart and sophisticated way and controls that.
Kudos to that. I don’t think that was obvious.
Tim Wu
In a way, I’m asking what is your advice for enforcers and the brain trust at Justice and FTC? Not specific, but how should they be thinking?
Randy Picker
Bring the cases. Settle the cases. Focus on the next.
Tim Wu
Okay. Let’s open to audience questions.
Go ahead Chad, since you were mentioned.
Randy Picker
And you’re at risk. If you ask another question now, that could be the next five years!
Audience member 1 (Chad Syverson)
Thank you. The hard thing for really answering that kind of question is, what’s the counterfactual?
You know more about the market that you just walked through than anybody. Even there it’s hard to do, “What’s the equilibrium that would happen in there?”
Is it useful to do polar cases as counterfactuals instead?
Forget about what the most likely outcome is. Just, what’s the worst case—conditioning on technology and the politics—what was the most anti-competitive possible outcome, and what was the most competitive possible outcome? Not likely to happen, but just could have happened. Just bound, “This is as good as it could have been. This is as bad as it could have been. It’s going to be in the middle somewhere.”
Is that useful? And do you have any sense of what either of those scenarios look like?
Randy Picker
Can you say something about that?
Tim Wu
Oh, no, I think it was a question for you.
Randy Picker
Okay, I’m happy to, but I mean you’re… Okay. .
I think, again, there are situations… We briefly were going to break up Microsoft, right? What would that have done? What would have been the consequences of that? That’s a pretty interesting question, right? I don’t know if that quite goes to polar bounding, but I think it’s in that spirit. And we chose not to do that, and what we did was…
Well, I don’t want to say anything; we’re going to hear from the people next, right? So, I’m really curious to see what they’re going to say.
Tim Wu
I’ll repeat my point. I think, first, that could be very helpful. I also think a form of counterfactual is other countries and other industries where we didn’t do anything, and what they look like. All this is complicated. These are all imperfect methods.
Randy Picker
Yes, I agree, but that’s… And you’re one to go look at these other industries, God love you. I would too, if I had the time, and so I don’t mean it that way at all. It’s more that the kinds of industries that we’ve been talking about, so again, Giovanna was talking about the IBM case in Europe, right? So these are worldwide industries.
Tim Wu
That’s true.
Okay. Matt, go ahead.
Audience member 2 (Matt Stoller)
Thank you. And, 330,000 words?
Randy Picker
Only 320. And one more chapter to write.
Audience member 2 (Matt Stoller)
My heart goes out to you for the editing process.
A couple of questions about how executives thought in the 1950s, 1960s, maybe 1970s.
Did AT&T and IBM buy a lot of companies in that time? Were they serial acquirers? Because Big Tech companies today are just constantly buying. That’s actually pervasive in corporate America today.
And then the other thing is, how often in your encounters in the archives were these executives talking about their stock price?
Randy Picker
I don’t know the answer to the latter. I assume much less, but I don’t know that.
On the first question, which is the purchases, and I assume you can talk about AT&T?
Tim Wu
Yeah.
Randy Picker
I don’t think IBM did very much of that. When they bought Lotus in 1995 (if that date’s right), that was seen as a huge deal, and it was a hostile takeover. In a later era where IBM is struggling—I talked about the Steve Ballmer years at [Microsoft], it’s probably the Ginni Rometty years where the company just doesn’t make progress—they start buying things as they desperately look for a strategy.
But core, classic IBM? I don’t think that was part of what they did.
Tim Wu
Yeah, and I think it’s kind of like, when you read these histories in the middle of the1980s when IBM has a dominance PC and there’s all these new challenges, there’s all these moves that would seem incredibly obvious in the 2010s period. Which is like, you should buy all these guys. You know? They seem so obvious. There’s all this talent out there. Why don’t you buy all these guys? This guy Bill Gates looks dangerous. You buy it. Why don’t you own him?
Bill Gates, I said…
Randy Picker
I heard that. I was paying attention.
Tim Wu
I talked to him once, and he thought, “We were sure they were going to crush us or buy us.” But they just don’t do it.
Now partially, maybe that’s an antitrust paranoia. But also, maybe this idea that you buy your way out of every problem was a consequence of non-enforcement of merger law for a period of time. Or relative non-enforcement against tech. I think that might have had something to do with it, such that Mark Zuckerberg openly talks about buying Instagram because, “Well, they could become a dangerous competitor. We got to get rid of them.”
Randy Picker
I teach those emails, yeah.
Tim Wu
Matt, you wanted to say something?
Audience member 2 (Matt Stoller)
Can I just rephrase my second question, because I think…
Randy Picker
The stock one? I just, I don’t, I’m not aware.
Audience member 2 (Matt Stoller)
I’ll rephrase it so you don’t have to avoid it. You can avoid it a different way. [laughter]
Randy Picker
[laughs] It’s fine.
Audience member 2 (Matt Stoller)
So, I was really struck. AT&T had those military contracts, and we’re doing this to help the Korean War. What were the executives focused on? Today they’re focused on market capitalization, increasing that.
What was the goal? Were they just like, “Oh. Let’s help them win the war!”? Was it increasing revenues? Increasing profits? Better, like… What was the lodestar for them?
Randy Picker
I don’t know. I heard what Richard was saying about the AT&T archives. I’ve done three archive trips. I’m not a historian, just so we’re clear. I’m a lawyer and an economist, but I have done three archive trips. And I have to say, going to archives is absolutely addicting. I mean, really. That would be like, you know, spend your summer vacation. It’s just so fantastic. It’s like time travel.
But I have not done that for AT&T.
Richard, do you know the answer to that? No, okay.
Tim Wu
What were they thinking?
Randy Picker
It’s a good question.
Tim Wu
I don’t think Randy’s avoiding question. I think he genuinely is admitting he doesn’t know the answer.
Aaron?
Audience member 3 (Aaron Honsowetz)
Something I’ve heard is recurring theme, both in your talks and Eric’s talks, is the emphasis on the ability to have access to use the technology by their businesses. I’m really curious how much, when we talk about antitrust and technology and technological development, are we really talking about antitrust bailing out the US patent system?
To give you a side note. You’d mentioned patent fights in the 19th century. There’s a really big patent fight over telegraph technology, to the point where companies are forced to go out business because their patents are ruled ineligible (I can tell you more about that later).
The key thing I want talk about though is, when we talk about antitrust technology, are we really talking about the antitrust system being used to bail out the US patent system without necessarily being very efficient doing its job?
Randy Picker
I keep saying things I don’t know. I’m not a patent expert either. I guess I’m skeptical about that. I think the patent system, Tim asked up front what was I surprised by. I thought a patent meant I owned it, you can’t use it. No. They all use them. I’ve not seen a discussion, for example, of the AT&T era and saying, “There were lots of junky patents that that AT&T had.” Bell Labs is seen, I think completely correctly, as a wonder. Really fundamental inventions.
This era that we live in today, where—I’ve got a figure like this at some point—where Google starts buying patents to sort of have a defensive trove. They say that their devices, the smartphone, implicates 250,000 patents.
How is that possible? That doesn’t sound to me as an outsider like a well-functioning patent system.
Tim Wu
It does strike me that if someone said, “I know the solution to all of these cases. We’ll force Google and Facebook to license all their patents,” people would be like, “What’s that going to do?”
You have a question, Filippo?
Randy Picker
He has a question. I don’t know when we’re done.
Audience member 4 (Filippo Lancieri)
How do you think about this broader competitive world? You’re in a world in the 1950s and it’s like an equilibrium effect, of course, because there’s just a lot more companies, in a way. A lot more military suppliers, probably. A lot of smaller companies. And a lot of what you describe is just people jumping firms and starting a new firm, which is something that you could do in the absence of non-competes in California, which is becoming less and less common as companies go to other places.
How do you think about this broader competitive ecosystem enabling the type of innovation in the absence of antitrust intervention, versus today where markets are much more concentrated and there’s just less places to go to if you want, or more ways to block you from moving around?
Randy Picker
Well, I do want to say, I say this in the text, I think that the creation of this common pool of patents effectively facilitates employees exiting firms. They know that they can go to this place and use this technology. When Texas Instruments hires Gordon Teal, he knows how to pull the germanium to make the chips. He knew that he wasn’t going to face some sort of claim from AT&T with regard to that.
What you obviously are doing there is, you are decentralizing the decision makers for the technology. The idea that AT&T had the silicon transistor and—not for a long time, 6 months, maybe 4 months—were sitting on it, and all of a sudden TI’s got it. How long would they have sat on it? I don’t know. The fact that Teal was able to exit clearly facilitates that.
Audience member 4 (Filippo Lancieri)
Can I ask another question connected to that, connected to what Richard said? How does the antitrust settlement solidify a belief that the companies can grow? So they actually can hire people. They get more money. Because now you basically took the company away from the market in a way that, before you were just being afraid of being squashed.
Randy Picker
It’s interesting. So, Giovanna, again, mentions in this piece, Gordon Moore. And Gordon Moore, one of the people who starts at Shockly Semiconductor.
Shockley was a terrible manager, so there’s a group that exits. The traitorous eight, as they’re known in the field, and they form Fairchild. And Gordon Moore says, “Well, the 1956 consent to final judgment, we knew we could rely on that.” This goes a little bit to Chad and counterfactuals. What no one asks him is, “Did you know that if you’d started two years ago, AT&T was giving out all these licenses.” He doesn’t pay any attention to that. Why would he? So I think that’s what’s hard to figure out.
Tim Wu
Laura, go ahead. Laura, I think, is one of the few computer scientists in the room, I’ll add.
Randy Picker
Oh, this scares me. Go ahead.
Audience member 5 (Laura Edelson)
Don’t worry, I’m not going to press you on anything too crazy.
Something that has been driving me nuts all day is, everyone has been, absolutely correctly, talking about the wonder that was Bell Labs. As a computer scientist, like, boy do I feel that. It was.
The thing that I am painfully aware of is, virtually all post-copper signal transmission technology was invented in Bell Labs. Coaxial transmission, the technology for fiber optic, microwave technology, cellular technology. All of it. Satellites. Do you know the idea for satellites came from a Bell Labs engineer, who then never did anything with it? They didn’t build any of this stuff. They built a demo of coax transmission, and then never built production lines. They built a demonstration microwave line, and then they never built any. MCI built the first one. The first cellular network was built in Japan.
One of the things that occurs to me, the reason we got all this. We got all of this stuff in a burst once the antitrust cases against AT&T, both the private and the DOJ case, began stacking up. That’s when we got all this stuff that had been invented, in some cases, many decades before.
Randy Picker
That’s a Tim book question, I think.
Audience member 5 (Laura Edelson)
I’m just wondering. In this framework you’re talking about, you know, settle and move on to the next. Is that really going to be a deterrent that is going to get all of this technology off the shelf?
Randy Picker
You want to run first, and I’ll run second? Go ahead.
Tim Wu
Yeah, that’s what I think.
The only thing I’m worried about, Randy, I agree with like two-thirds of what you say in your thing. But with settlement, the danger of the wimpy settlement is so obvious to me. You have this 1956, which was not a wimpy settlement. I also worked at the FTC in the 2010s and, boy, we settled some stuff for, like, a penny on the dollar and let all this stuff kind of go. Everybody gets to declare victory and talk about how great they did and pat you on the back and have drinks and so on, and then go on to jobs with the guys you settle with.
That risk of wimpy settlement strikes me as so serious. I don’t know, structurally, how we do anything it. But I hear you about, they’re faster, but…
Randy Picker
In a world in which the technology just goes and goes and goes.
Tim Wu
But I agree with this cul-de-sac. Another one, the answering machine was in working form in like 1934 or something, and obviously reached consumers in 1980.
Richard seems like he’s dying. I know we’re out of time, but…
Randy Picker
If we’re out of time, we should stop.
Tim Wu
Richard, why don’t you just say one thing, as long as it’s not too long.
Audience member 6 (Richard John)
I was just going to say, the patent issue was enormously important [inaudible] the 1880s, one of the few historians to work with Schumpeter. Chandler was always startled by that it because it didn’t fit Chandler’s story, but he taught that book.
It’s very important in the independent Bell struggles, and it’s a daily question in the press. Of course, the dialup ends up with the independents and not with the Bell.
It remains important even after Bell Labs in 1925. The question about magnetic recording and the answering machine. One of arguments that’s come up recently is that Bell was aware that an awful lot of telephone calls were about illicit matters, and they didn’t want to push something…
Tim Wu
Yeah, obscenity on the recordings was one of the…
Audience member 6 (Richard)
Or gambling, yeah.
And then the radio act of 1927 is basically a patent settlement.
So I wouldn’t want anyone to come out of here thinking that patents were not absolutely central to the business strategy of high-tech firms, going back to a landline telegraph. The telegraph industry was organized around patents, in fact. It goes back to 1838.
That’s all I wanted to say.
Tim Wu
Thank you, that’s helpful. Why don’t you join us again?
Randy, I hope you keep the whole book at its full length. I think it’ll be fantastic.
[applause]
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
