The Good Fight
Larry Summers on What Went Wrong on Campus

Larry Summers on What Went Wrong on Campus

Yascha Mounk and Larry Summers also discuss the promise and perils of artificial intelligence.

Larry Summers is an economist, the Charles W. Eliot University Professor and director of the Mossavar-Rahmani Center for Business and Government at Harvard Kennedy School, and a member of the board of directors of OpenAI. Summers is the former President of Harvard University, the former Secretary of the Treasury under Bill Clinton, and was a director of the National Economic Council under Barack Obama. 

In this week’s conversation, Yascha Mounk and Larry Summers discuss how universities can re-commit to pursuing truth and protecting academic freedom; how current economic indicators contrast with how many people actually experience the economy; and how Biden can improve his odds for re-election.

The transcript and conversation have been condensed and lightly edited for clarity.

Yascha Mounk: The last few months have been rather eventful at Harvard University. Tell us your view of what has happened and why it matters.

Larry Summers: It's been a very difficult time. I think what universities do is as important as the work of any other institution in our society, in terms of training young people and preparing them for careers of leadership, and in terms of developing new ideas that set the tone for the cultural, the political, the policy debates that go forward.

Paul Samuelson famously said that if he would be allowed to write the economics textbooks, he didn't care who would get to perform as the finance ministers going forward. So I think what happens in universities is immensely important. And I think there is a widespread sense—and it is, I think, unfortunately, with considerable validity—that many of our leading universities have lost their way; that values that one associated as central to universities—excellence, truth, integrity, opportunity—have come to seem like secondary values relative to the pursuit of certain concepts of social justice, the veneration of certain concepts of identity, the primacy of feeling over analysis, and the elevation of subjective perspective. And that has led to clashes within universities and, more importantly, an enormous estrangement between universities and the broader society.

When the president of Harvard is a figure on a Saturday Night Live skit, when three presidents of universities combine to produce the most watched congressional hearing film clip in history, when applications to Harvard fall in a several-month period by more they've ever fallen before, when alumni are widely repudiating their alma mater, when they're the subject of as many legal investigations as the Boeing company, you have a real crisis in higher education. And I think it's been a long time coming because of those changes in values that I was describing.

Mounk: Tell us a little bit more about the nature of the conflict here. What is the conception of the university that has historically guided it, and how is it that those values have changed over the last ten years?

Summers: I think the values that animated me to spend my life in universities were values of excellence in thought, in pursuit of truth. We're never going to find some ultimate perfect truth, but through argument, analysis, discussion, and study we can get closer to truth. And a world that is better understood is a world that is made better. And I think, increasingly, all you have to do is read the rhetoric of commencement speeches. It's no longer what we talk about. We talk about how we should have analysis, we should have discussion, but the result of that is that we will each have more respect for each other's point of view, as if all points of view are equally good and there's a kind of arbitrariness to a conception of truth. That's a kind of return to pre-Enlightenment values and I think very much a step backward. I thought of the goal of the way universities manage themselves as being the creation of an ever larger circle of opportunity in support of as much merit and as much excellence as possible.

I spoke in my inaugural address about how, a century before, Harvard had been a place where New England gentlemen taught other New England gentlemen. And today it was so much better because it reached to every corner of the nation, every subgroup within the population, every part of the world. It did that as a vehicle for providing opportunity and excellence for those who could make the greatest contribution. But again, we've moved away from that to an idea of identity essentialism, the supposition that somehow the conditions of your birth determine your views on intellectual questions, whether it's interpretations of quantum theory or Shakespeare. And so that, instead, our purpose is not to bring together the greatest minds, but is back to some idea around multiplicity of perspective with perspective being identified with identity. We used to venerate and celebrate excellence. Now, at Harvard, and Harvard is not atypical of leading universities, 70 to 75% of the grades are in A-range. Why should the institutions that are most celebrating of excellence have only one grade for everyone in the top half of the class, but nine different grades that are applied to students in the lower half of the class? That is a step away from celebrating and venerating excellence. 

We celebrate particular ideas in ways that are very problematic, and we are reluctant to come to judgment: What started all the controversy at Harvard, and it has many different strands, was on October 7, when 34 student groups at Harvard, speaking as a coalition of Harvard students, condemned Israel as being responsible for the Hamas attacks. Those reports of the 34 student groups were reported in places where literally billions of people read them. And based on some inexplicable theory, the Harvard administration and the Harvard corporation (the Trustees of the University) could not find it within themselves to disassociate the university from those comments. I have no doubt that if similar comments had been made of a racist variety, there would have been no delay in the strongest possible disassociation of the university. But because Israel demonization is the fashion in certain parts of the social justice-proclaiming left, there was a reluctance to reach any kind of judgment, even about the most morally problematic statements.

It is not that the university was slow to comment on George Floyd. It is not that the university was slow to comment when some within it wanted to host a “black mass.” It is not that the university has been slow when social scientists have wanted to speculate about group differences. So I think that this combination—the veneration of a particular concept of social justice, the act of disrespect for excellence, the celebration of identity rather than the pursuit of opportunity, and the rejection of truth—have made these institutions problematic in the impact they have on those who pass through them, in whatever influence they have on the broader society and estranged from the broader society. And I think for any kind of private institution, it has to find a social contract in which it can operate with the broader society. And the fact that the ways in which great universities have acted have so enabled the Elise Stefaniks, the Bill Ackmans, and the Christopher Rufos, speaks to the danger with which they have been governed. 

I come from a left of center tradition. And I'm not far left of center, but surely left of center. And I've always been acutely aware, in thinking of universities, that Ronald Reagan got his political start by condemning and running against what was happening at Berkeley in the mid-1960s. And that the tradition of then-Governor Brown—who had inaugurated this wonderful idea of free college education for anyone who had a B-average in a California high school—got completely blown away in a tide of fury about “welfare Cadillacs.” But what brought that tide to prominence was a general revulsion at what had been going on at Berkeley that Ronald Reagan rode to his political career. And so it seems to me that universities that fail to govern themselves effectively are at immense peril to themselves and to the broader progressive values that they hold.

Mounk: How dangerous do you think this moment is, not just to the reputation of universities but to their actual ability to function as co-institutions of the United States? I'm a little torn on this. On the one hand, you can make the case that even the most affluent and insulated universities like Harvard need federal funding for the research that they undertake, and to finance a lot of the student loans that its undergraduates take out. On the other hand, Harvard has an endowment of, what, $50 billion? And it does continue to have real support in the population. Where would you place yourself on the worry scale about sort of the worst-case scenarios here?

Summers: I think one would find for any Ivy League school that the federal government was ten times as large a donor, at least, as any other donor. And I think it's fair to say that the universities have thumbed their nose at what is by far their largest donor. And they're certainly not prepared to take that casual and cavalier attitude towards much smaller individual donors because of what they think the consequences would be. I think it's fine to stand strongly against a set of people who in many ways are riding this horse, but wish the process of thought and wish academic freedom ill. The problem is not that Harvard has worked itself into a war with Elise Stefanik. The problem is that it got itself condemned from the White House press briefing room of the Biden administration, that it finds itself subject to investigation from the Department of Education of the Biden administration, that the attacks on it are coming in a bipartisan way.

I think one of the aspects of how this has happened is that while on the one hand we think of intellectual communities as being the most broad-minded of communities, on the other hand they are actually among the most narrow, insular and inward-looking in the way they evaluate themselves and in the way they think of the necessary decision making. There's an old story about when Pat Moynihan had decided to leave the UN and called the Dean of Harvard to say he would be returning. He said he'd let the president know and the Dean of Harvard assumed he was referring to the President of Harvard rather than the President of the United States. And that bespeaks a kind of attitude that I think is very problematic. 

Now, Yascha, I am a great believer, however, in the power of self-denying prophecy. I think that's what has actually propelled the United States over many years, and I suspect it may propel universities, that it is precisely the capacity to become very alarmed when off-track that leads to correction and leads to the most dire prophecies not being realized. So I suspect that some universities will in some way find their way back.

Mounk: I love the idea of a self-denying prophecy. What might this correction look like? You identified earlier, as part of a key of the problem at many of these universities, is the hypocrisy, right? The hypocrisy of standing in front of Congress and saying, “We really care about free speech,” when in many situations and controversies in the last years the free speech rights of faculty members and others at the university were not respected when it didn't come to anti-Semitism, but other kinds of forms of bigotry or supposed bigotry. The hypocrisy of very loudly condemning certain kinds of political statements and events but not condemning other political statements and events.

Now, one response is to say the problem is the Jews haven't been sufficiently protected: All of these other groups have come to have these group protections, we have become incredibly sensitive about anything that any member of that group might claim to be discriminatory or insensitive (whether or not that's true in a particular circumstance) and where we went wrong is not recognizing the Jews are also a vulnerable group that needs to be protected against speech that they might find offensive and so on. Another way out of this is to say, no, we actually have to go back to having a genuinely robust conception of academic freedom in which we tolerate that some people will say unpleasant things, in which the university stops speaking for all of its community members. Even when some of the community members may be saying unpleasant things.

I obviously tend towards the second of those solutions. And I think you do too. I don't think the solution here is to create one more group that has special protected status. It is to refocus the mission of the university on the pursuit of truth and the protection of academic freedom. But how do we do that? How do we make sure that the correction doesn't go precisely in the wrong direction?

Summers: I think you and I are very much in agreement. I don't think any reasonable person can fail to recognize a massive double standard between the response to other forms of prejudice and the response to anti-Semitism. And yes, you could have debates about when anti-Zionism or the demonization of Israel is and is not anti-Semitism. But on any reasonable conception of what's going on, there has been a double standard. And I think those of us who are concerned about the double standard come to a view about how we want it remedied. And I think for the most part, the right way of remedying it is with a de-emphasis rather than a re-emphasis on identity.

Everyone needs to be enabled to feel safe. That doesn't mean that they have a right to avoid being triggered by speech they don't like, or to be spared exposure to ideas they find noxious. That doesn't mean they have a right to bean-counting exercises where the share of members of their group is evaluated against a share of its population. It does mean that they're entitled to the maintenance of an open and tolerant community where no one is allowed to shut down any set of ideas, that they have the right to be protected from discrimination, and that they have the right for there not to be indoctrination. I think in many ways what would be most problematic would be an indoctrination arms race in which a larger and larger fraction of an education is consumed by a recitation of the grievances of various groups. 

Mounk: Let’s change gears a little bit. You recently joined the board of OpenAI after very tumultuous events at the organization. Tell us about what you see as the promise and the peril of artificial intelligence. Evidently, this is a giant technological revolution whose impact we're only just beginning to feel. 

It seems to have the potential to vastly increase economic productivity and perhaps even human creativity, and, at the same time, it is a potential danger. It's a potential economic risk because it may, as people always fear when there's new technologies, lead to mass unemployment. It is, I think, a threat to our self-conception as humans because it is at least possible that, within the next years, artificial intelligence will be better at creating songs than musicians, better at creating pictures than painters, better at writing novels than novelists or erudite works of academic scholarship than academics. And finally, of course, there's the risk of AI going rogue and perhaps being the beginning of the end of humanity. How do we think about this tremendously important development and the impact it's going to have on our society?

Summers: I don't think anybody should speak with complete confidence. But I think there's very substantial reason to believe that this is a major event in history. One of my academic colleagues, when I was asking him about its significance, said the most perceptive thing that I have heard about it. He said, “It's the first major new way of knowing for mankind since the invention of the scientific method.” And that is something that is very powerful and I think is very existential. And there is a capacity not just to execute, but to create, that I think really is going to be profound in its implications. And I think you saw that in the degree of virality—which was not anticipated by its authors, even—with ChatGPT. And I think that progress is very, very likely to continue. 

I think one thing that is probably underappreciated at this point is the magnitude of the inputs that will be necessary for this process to take place. It is imaginable that data centers will be substantial, as a use of electricity—a decade from now on a planetary scale. And if one thinks about what that's gonna mean for the demand for “fabs,” the demand for power production, the demand for all of that, I think it is something that is truly, truly important. I think it's important to remember that, over time, we change our conceptions of what the human role in things was. I remember there was discussion during the period when I was coming out of high school of how if people no longer used slide rules to do calculations, they wouldn't have the fundamental intuition for what was going on within the calculation that using a slide rule provided if they used a calculator instead. I remember when there was discussion of how the printing press would diminish the role of telling and of memory by enabling storage in this other way on pieces of papyrus.

Mounk: Well, a lot of developments have pro tanto losses, right? I'm sure that there's a certain kind of intellectual endeavor and a certain kind of getting to know a text that comes from copying a book and that we do lose something there. That doesn't mean that, all in all, we have a loss.

Summers: We may have lost that, but what we have gained is not unrelated and is much, much larger. I think it's the essence of Thomas Kuhn, that there were questions that Ptolemaic astronomy addressed that Copernican astronomy does not, but Copernican astronomy supplanted Ptolemaic astronomy and it was for the best. It was kind of taken for granted, around the time it happened to me, that when you learn to drive, if you were a responsible person, you had to understand a certain amount about what went under the hood of your car, because you never knew when there'd be a problem and when you'd have to fix something. Somehow, nobody has any expectation of any kind like that around learning to drive a car today. And I think in the same way there are a variety of things that will be no longer necessary to understand to have coding performed on one's behalf. I have heard it said, and I don't really know, that a nuanced conversation between two world leaders or two sophisticated business leaders on a very complex deal, where neither of them speaks a word of the other language, will be able to take place with machine translation that will capture every nuance better than a human could, within a matter of a year or two; that Kissinger could have had his celebrated dialogues with Zhou Enlai better with a machine as the translator than any human translator, within a couple of years. 

I think this is going to change what we learn ourselves and what we delegate. I think it is likely to change the skills that are valued from the skillful execution of the basically-understood cognitive, towards tasks that involve more profound sparks of creativity, that involve more iconoclasm outside of an existing corpus, and at the same time put more of a premium on EQ relative to IQ. And so I think the set of individuals who are in a position to be most productive and contribute most is likely to change. I think, and this is something we very much wrestle with at OpenAI: That with technology so potent, there are profound questions of governance for the broad society that one wants, on the one hand, to provide a set of checks and balances and tensions to preserve against disastrous discontinuity, and one wants at the same time to allow things to move forward. 

I suspect that this is going to engage people from every discipline, from philosophers to quantum scientists. And I think that maybe this is a point to sort of bring that conversation together: I wish I could feel that our great universities were in a stronger position to be contemplating all of this and providing some of the wisdom and some of the thought that will be necessary to help us as a society manage all of this. Because I think it's quite possible that we're involved in another transition like the post-1870 transition.

I'm very much influenced by the writings of my former student, Brad DeLong, and many others who basically make the point that from the time of Pericles to the time of London in 1850, standards of living on average, economic change on average, was probably less than one- or two-hundredths of one percent each year. And then, all of a sudden, we now live in a world where standards of living have doubled four times in China since I left graduate school—four doublings compared to hundredths of a percent. And that's the transition that took place in the aftermath of the Industrial Revolution. It's why people live twice as long as they used to, why child death is now a rarity, why literacy is generally possessed around the planet. But look at the first half of the 20th century and the transition was also associated with all kinds of catastrophe. And we may well have another very major transition. And we want to make sure we get the fruits of that transition, but we sure hope that transition happens in a better way and I know that's something that the people who are at the cutting edge of this research are very much engaged with. 

I think it's a mistake to despair. If you had asked anybody associated with the Manhattan Project or anybody watching, what was the chance as assessed on Labor Day of 1945 that no nuclear weapon would ever be used in anger over the subsequent 70 to 80 years, I think people would have regarded that as extraordinarily unlikely. So it's not that we lack the capacity to manage these things but we can't assume that it's something that happens automatically.

Mounk: I broadly agree with that view, but let me push you on a couple of points. The first is that surely the invention of a printing press was on the whole a good thing for humanity, but it did lead in part to a lot of religious sectarianism and strife and religious wars and other things in the next centuries, right? So when you take a long enough historical view, surely the invention of a printing press was a very positive thing. But it certainly had very significant negative consequences over the course of the following century. 

You invoke the invention of the nuclear bomb. With respect, I would say that the jury is still out on that. It is indeed a great achievement of humanity and of political governance that we have not had a nuclear war in the last 80 years. But we also have to not have a nuclear war in the next 2000 or 20,000 or 200,000 years. And I'm not sure that 80 years of having been able to manage this is a sure indication that we will continue to manage it for the next 2,000 years. And relatedly to this, there is of course some amount of nuclear proliferation. And more broadly, my hunch is that it has proven to be very difficult in history to limit the spread of a human-invented technology. At the beginning, it is usually preserved to a very small number, but, over time, more and more people come to be able to avail themselves of that technology. And broadly speaking, with perhaps very limited exceptions, our state of knowledge about the world is that which is conformable with human ingenuity and whatever institutional constraints we have that make us not as good at developing medical drugs as we should be and so on. But humanity's collective capacity to say, “That technology seems potentially dangerous, let's go no farther,” has proven to be very, very limited. 

So isn't the question here simply what a fully developed AI looks like and whether that will submit itself to human will or whether it does in technological terms tend towards becoming self-serving, and if it's the latter, do we really think that not just for the next ten years, perhaps not just for the rest of our lifetime, but for the rest of humanity, we're going to be managing to contain that destructive tendency? I guess I'm a little skeptical about that.

Summers: I certainly did not intend any complacency. You referenced the post-printing press experience, I referenced the post-Industrial Revolution experience, as examples of problematic transition. But I think it is maybe a little too easy to succumb to extreme pessimism.

I do think it's an extraordinary achievement that it's not just that nuclear weapons have not been used for the last 78 years, it's that there hasn't been war between major powers for the last 78 years. And yes, 78 years is the bat of an eye. But there's probably been no moment in the last 3,000 years when 78 years went by without a war between major powers. So yes, we need to do everything we can with as much imagination as possible to contain the possible risks here. And I don't mean at all to be heard as minimizing them. I would say that (and this is a comment that always comes up in these discussions) if those with the greatest orientation to responsibility and concern fail to move forward, others will not fear to tread. And so I think that it is very important that American institutions maintain leadership in this sphere to the maximum extent that is possible.

I think it is very important that as much reflection goes into how this will interact with the broader society as goes into the development of the particular technologies. So I don't want to be heard at all as exuding complacency. But at the same time, I do think that the capacity to augment the kinds of cognitive strengths that we already have is something that is very, very valuable and can be extraordinarily constructive.

I don't find the view of my economist colleague, Daron Acemoglu, that sees the history of progress as a history of woe (that's somewhat unfair to his view, but I think you can understand why I said it) to be a compelling reading of history, but this is going to be immensely important. And I am 69 years old. So a large part of what is going to tell the story is going to be the work of generations like yours, not mine.

Mounk: I want to broach one last topic, which is the economy. First of all, there was widespread expectation that the United States would go into a recession. That appears not to have happened. What do you think is the economic outlook for the coming months and years, and why is it that this expected recession has, at least so far, not materialized?

Summers: I was much less confident than most others that we would head into recession because I felt that it was very difficult to assess just how contractionary monetary policy was. My own suspicion was that because, for a variety of reasons, bigger budget deficits, greater renewable energy demand, for example, the neutral interest rate had increased. I was much less sure that the increase in interest rates was going to be quite as contractionary as many other people supposed, and I think that's turned out to be the case. So I would be quite surprised if the economy went into recession over the first two-thirds of this year. I'm a bit less certain that inflation is quite as “down for the count” as many people suppose.

There are still some populist forces. I think there are still quite tight labor markets. I think there are still risks of various geopolitical shocks. So I see inflation as a bit less down for the count than the consensus does, though inflation has certainly come down faster and apparently more durably than I would have expected. In the absence of an economic downturn, I think that may speak to the credibility which the Fed was able to regain by acting very strongly. But as I look at the world, I don't see the macro-financial outlook as being the principal risk that keeps me up at night. I see populist politics and all that can follow from revanchist populist nationalism in this country and others as the greatest threat to our ability to move forward in ways that are prosperous. I see problems on a global scale—climate change, pandemic, disorder—as risks going forward. I think those are larger risks at this point than cyclical economic fluctuations.

Mounk: Why is it that most economic indicators seem to be very positive at the moment? When you look at the kind of metrics that people usually examine to see whether or not the economy is doing well—employment, rising wages and so on—there seems to be a story of the economy going well, and yet Americans and voters continue to disagree with that reading, and continue to say that actually the economy is not going nearly as well as economists and some opinion writers seem to believe. Is there an obvious explanation for that apparent discrepancy?

Summers: There's nothing obvious, nothing that is a complete explanation. My best guess is that the single most important explanation for that phenomenon is that for reasons that you can explain in theory, economists do not treat the cost of money as being part of the cost of living. But people do.

They don't look at the price of a new car; they look at the monthly payment when they buy a new car. They don't look just at the price of a house; they look at the monthly payment that goes with purchasing that house. And if you look at calculations of inflation that include the cost of money, they show a much, much less favorable picture. And so I think it's the failure to recognize that in a technical sort of statistical sense, what explains some of the anomaly. 

Mounk: That's interesting. I had a slightly different version of that theory. And I wonder what you make of it. 

Economists compare inflation usually to over a period of a year. And so today, the inflation rate, excluding the cost of money, is not that high. To buy a loaf of bread at the supermarket today is not that much more expensive than it was a year ago. But most people don't compare it over the course of a year, right? They have a sense of prices used to be X and that used to be some sort of inchoate time period that is ten to three years ago, right? And so each time we go to the supermarket, they say, look at how much more expensive this loaf of bread is than it used to be five years ago. And so even as the inflation rate has come down in economic terms, they still feel the inflation because they're not comparing it to a year ago.

Summers: That was the second explanation that I was going to give. And I think that points towards the fact that ultimately people do sort of understand that we're not going to see some one dollar gasoline in the United States again, that we're not going to see cars that have four-digit price tags in the United States. So I think ultimately people do judge prices relative to some norm and over time those norms do adjust. But my guess is that you will see some catching up of perceptions with statistics. How rapid that will be, I think, is hard to know. 

I think there are also issues of referred pain around a whole range of social phenomena, some kind of anomie. Some people blame it on social media, some people blame it on other things. And so I suspect that when other things in people's lives aren't working ideally, they often blame the economy. You know, I learned something, Yascha, from a kind of trivial experience I have as a college professor. Some semesters I lecture better, some semesters I lecture worse. Students evaluate me and sometimes it's better and sometimes it's worse. My reading list changes very little from semester to semester. But in the semesters when they like me as a lecturer, they think I've got a great reading list. And in the semesters when they don't like me as a lecturer, they think I have a poor reading list. And it's not really because the reading list changes. It's because there's a kind of tendency to generalize about everything together. And I think that's some part of what's going on as well.

Mounk: So people are not happy with the incumbent president and so therefore they blame the economy. That naturally leads me to what was going in any case to be my last question, which is that you've served in distinguished positions under the last two democratic presidents, Bill Clinton and Barack Obama. They were both two-term presidents, managing to win reelection. Though you were never on the political campaign side, you got an inside view of how they were able to win and engineer those re-election victories. Now we have another president who we also know well, Joe Biden, who is trying to get re-elected. Those re-election efforts are going somewhere between medium and bad. Certainly, he seems at this point to be trailing Donald Trump in many national polls and many swing state polls. I think, on the prediction markets, Trump and Biden are roughly equal. I think as we're recording this, Trump seems to be a little bit favored. A few days ago, Biden was a little bit favored. 

What is it that Joe Biden can do at this point—and what is it more broadly that the Democratic Party can do—to improve the odds for November? And is that going to be mostly something on the economic front or mostly something on the broadly-speaking cultural front?

Summers: I tend to find political experts' opinions on economic questions to not be very sound and thoughtful. And I'm not sure why I should suppose that my opinions on political questions will be particularly sound and thoughtful. So I answer the question with humility. But my instinct is that political parties prevail and incumbent presidents prevail by returning to a broad American center. And I am hopeful that Joe Biden, whose roots are with an American middle class, will find a broad expressive American voice in the months ahead that will place less emphasis on responding to each particular identity element in the Democratic bouillabaisse and instead speak to the hopes, the obligations, the expectations of all Americans in a universalist kind of language. I think that he has styled himself over many years as “middle class Joe,” and that's something that goes deep within him.

My hope and my best guess is that we will see that come out and that as it comes out and as the clamor of the various activist groups within the party comes to seem less dominant, he will emerge as a unifier and as a successful candidate. But again, I answer economics questions with confidence and political questions with trepidation.

Please do listen and spread the word about The Good Fight.

If you have not yet signed up for our podcast, please do so now by following this link on your phone.


Podcast production by Brendan Ruberry and Jack Shields.

Connect with us! Spotify | Apple | Google

Twitter: @Yascha_Mounk & @JoinPersuasion

YouTube: Yascha Mounk, Persuasion

LinkedIn: Persuasion Community

The Good Fight
The podcast that searches for the ideas, policies and strategies that can beat authoritarian populism.