Fixing Higher Education & New Startup Opportunities with Marc and Ben

Steph Smith: Hello, everyone. Welcome back to the A16C podcast. This is your host, Steph Smith. Now, in a recent episode, you got to hear from our co-founders, Mark Andreessen and Ben Horowitz, as they talked about the crisis in higher education. They did so by breaking down the 12 functions of the modern university, going all the way back to when these institutions were founded. But today, you'll get to hear part two of that very conversation as they focus on the present, exploring different solutions for overhauling the modern university, an admittedly difficult task, especially given the many stakeholders with varying incentives.

Steph Smith: Mark and Ben take an entrepreneurial lens to this age-old industry, surfacing opportunities at every step, from recruitment to student experience to the administrative bloat. If you're interested in hearing part three of this series, an extensive Q&A on the topic, you can go ahead and check out the Ben and Mark show wherever you get your podcasts. It's time to fix.

Steph Smith: As a reminder, the content here is for informational purposes only, should not be taken as legal, business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see a16z.com slash disclosures.

Marc Andreessen: We did an episode about a week ago on universities, the prevailing issues in universities, the crisis they're going through. I want to reiterate why we did that one, which I encourage people to listen to first if you haven't heard it before this one. This is part two, but also why we're doing it. We're discussing this rolling crisis that a lot of the universities, particularly the American universities, are going through right now.

Marc Andreessen: It's very important to understand two things from our standpoint. One is the reason we're digging into this topic is really twofold, and we'll talk about both parts today. These are incredibly important institutions for the country and for the people of the country, and by extension for the world. So it's critically important that what happens at American universities goes well. And it's a very big problem, not just for them, but for a lot of the rest of us when they don't go well.

Marc Andreessen: Ben and I talked about this last time, but our personal stories obviously involve us being where we are because of the great experiences we had at universities. The university, both the teaching process, generating graduates, and the research process, is the seedbed of everything in the tech industry in Silicon Valley and everything that we do every day. So these are important topics.

Marc Andreessen: The other thing worth noting is that we're not just doing this to criticize. We're trying to see if we can be constructive. In particular, we're trying to take a look at the issues, not through a moment-in-time, hot-in-the-news perspective, but rather from a structural standpoint. We're analyzing universities as if they're systems and structures with incentives and ways of doing things. Those ways of doing things have built up over a long time, and the nature of large organizations and systems that build up over a long time is that sometimes they accumulate problems and sometimes they need change, improvement, and reform. That's why we think it's a good thing to look at.

Marc Andreessen: The other perspective we have is that there are clearly startup opportunities emerging. We're going to talk quite a bit today about what the existing institutions could do to improve their situations, but we're also going to talk about some of the startup opportunities that are flowing from the crisis in higher education. By the way, those startup opportunities would probably be appearing anyway, because the higher education system just can't reach most kids who need to get educated around the world. So there would probably be startup opportunities even without issues. But it may be that if universities can't fix some of their issues, ultimately there will be opportunities to build new institutions, new companies, new nonprofits, maybe new research entities, and maybe do more of the things that universities have historically done.

Marc Andreessen: We probably need to fix the old ones and then build some new ones, just given how fast the world is evolving.

Ben Horowitz: Yeah, part of the context for all this is that universities historically were only ever built to train a very small percentage of the 18-year-olds in the world each year. I don't have the numbers in front of me, but the number of people who turn 18 per year worldwide is an extraordinarily large number, many tens of millions. The current higher education system was never built and is not built to accommodate that.

Marc Andreessen: It's built for a small number of religious scholars.

Ben Horowitz: Yes, originally, yes. And then later, a small number of secular scholars. In the U.S., the universities have become a broad-based expectation that people go to college, as we talked about last time. But even if you could send every 18-year-old in the U.S. to college, the U.S. is still only 4% of the population. So 96% of kids, actually probably more than that every year, 18-year-old kids are outside the U.S. Most of them are in places that don't have physical colleges or universities. So there's a general scaling problem.

Marc Andreessen: Right, which probably needs to be addressed separately from the existing system anyway. So we're going to dive into all of that. And then we have a lot of great Twitter questions, which we are going to get to at the end. Or if we go super long, we'll do part three.

Ben Horowitz: We'll do part three.

Marc Andreessen: Okay, so the way I think about how to get started is the question we left off at the end of the last podcast. The last discussion, we went through all the issues and then we kind of left angling, "All right, what do you do about them?" So this is the "what do you do" section.

Marc Andreessen: I wanted to start by talking about the way we try to always do this in business: try to start with, "Okay, before you figure out what you're going to do, what exactly are your goals?" And then the very next question from "What are your exact goals?" is "What exactly are your goals for who?" Who are you trying to satisfy? Who are the customers? Last time we listed out the 12 functions of the Modern American University, which you can hear all about on that last episode. What jumps right out of that list is that the universities, the way they run today, certainly have a large number of different constituents.

Marc Andreessen: I made a list. I came up with, I think, 16 or 18 on the list. I'll list them real quickly: Students, faculty, administrators, board of trustees, alumni, donors, downstream employers who hire the graduates, parents who are very involved, immigration officials, sports fans, regulators, politicians, the press, which scrutinizes universities all the time, downstream policymakers who are influenced by the science and policy prescriptions coming out of the universities, and then number 15, society as a whole.

Ben Horowitz: Society as a whole is a tough customer, by the way.

Marc Andreessen: Yeah, it very much is. Universities have designed themselves from inception to have a big impact on broad society. That's one of the goals: they are trying to reform society, advance society. Having put themselves in that position, they're naturally going to get scrutinized by society.

Marc Andreessen: So how do you think about from a leadership standpoint, from a management standpoint, and with your experience on the Columbia board, how do you think about an institution like this? We all know it's hard enough to run a company that just has customers, or customers and employees, or customers, employees, and shareholders. You have two, three, four constituents, it's hard enough. How do you even approach the job of leading an institution that has that many constituents?

Ben Horowitz: Well, I do think companies have a lot of constituents also, but they have better clarity and uniformity on which are the most important constituents. If you look at Andreessen Horowitz, we have investors, we've got entrepreneurs, we've got employees of entrepreneurs, we've got our own employees, we have the press, we have society as a whole. We have all these things to consider. But it's very clear to us, and we try to make clear to everybody in the firm, that if we don't attract the best entrepreneurs in the world, none of the other things matter. So you have all these people who have an interest, but you've got to get the main thing done. Anything that compromises that prime directive has got to go and be subordinated.

Ben Horowitz: In the universities, partly because of their sheer size, you have people who have no concern about the students at all. Large pockets of the university are only focused on one of the other things. That's where it starts to lose its focus and degrade the product for the most important constituent, which I'd argue is students.

Marc Andreessen: Yeah. So if you were, especially with your board experience, placed in charge of one of these tomorrow, how would you rank these? Or maybe that's an unfair question. Maybe the question is, how would you even go about figuring out how to rank them?

Ben Horowitz: I think you have to start with students, and then everything on the list is a little bit in service of students to varying degrees. The faculty are obviously in service of the students. The administrators are in service of the students. The alumni are there to give money to support the students. Depending on your issue in attracting the best and brightest students and giving them an experience, a product, and a career, the things they're looking for, that's optimal. You would prioritize things to get you there.

Ben Horowitz: There are always things like regulators that any business has to mitigate or deal with. That's not going to be a primary thing; you have a small team focused on that. But you can easily lose the thread if you manage to noise levels. The press is probably the most distracting one. If the press calls us a name or says we're not doing our job, all of a sudden a huge focus goes over there. If you're not careful, that'll distract from what you want for your main customer.

Ben Horowitz: That's, to a large degree, what's happened. Which is why I like cost. How is it possible that tuition has gotten so high? What the hell were you optimizing for that let you think your students wanted that? I think there were some market corruptions where the government's providing loans to students, which led to upward pressure. There are other things like professors can get jobs in the private sector, which is a more scalable sector in terms of making money, potentially a lot more than you could pay in academia. That drives salaries up. There are a lot of factors that lead to that.

Ben Horowitz: But at the end of the day, how the hell do you end up with a product that costs $300,000 and gets the average student a job that's worth $50,000? How are they ever going to pay back $300,000 if they only make $50,000 a year? That's insane. That's where I think the product for the student has fallen apart. So maybe that's a good place to start: what would you do to get cost at the university under control?

Ben Horowitz: One of the big things we learned, which is stunning when you hear it, but it got there incrementally, is that at many of the elite institutions, the number of administrators outnumbers the number of students. That's an opportunity for savings. If you just tighten your belt, you'd say, "Maybe we can do a 20 percent reduction there." But if you think about it a different way, it's very analogous to the debate when the Biden administration hired 86,000 new IRS agents. The interesting thing was that anybody in our business would go, "You could have just hired seven good software engineers and they would have done a far better job than 86,000 agents at figuring out who was cheating on their taxes." Because it's a data problem, a forms problem. This is what computers are amazing at.

Ben Horowitz: People talk about what AI is good at. I'll tell you what AI is really good at: filling out forms, looking at forms, getting data out of forms, comparing that data to what the data should be, figuring out anomalies in the data that no human could. Couldn't AI do all these administrative tasks? Couldn't you just get rid of 95% of it? Then you're starting to get costs back where they are.

Marc Andreessen: Well, if I could, just a couple of things on that. One is it begs the question of who your constituency is, because if your top constituency is the administrators, to the extent that the institution is being run for itself, then obviously that's a direct threat.

Ben Horowitz: Yeah, and I think that's always a thing in every organization. One of your constituencies are your employees. But when your employees take precedent over your customers, that's usually the end of the business. This is the pattern. In the private sector, you tend to go bankrupt a lot faster because there aren't government subsidies and tax credits. There are so many goodies that the universities have access to that that's not something that happens in business. But eventually, over time, in the long run, if the employees become more important than the customers, you get to the same end.

Marc Andreessen: One of the quirks and incentives of the whole thing, see if this makes sense, is that one of the parties that is not a constituent, to your point, is shareholders. Companies have shareholders; the universities are non-profits. But the consequence of being a non-profit is a very interesting incentive. The whole point is that you're not for profit, you're trying to break even. If you have the opportunity to have rapidly ramping revenues because of subsidies, you then have every reason in the world to ramp expenses at the same rate.

Ben Horowitz: Exactly.

Marc Andreessen: You don't want to generate massive margins. It's not your purpose to generate margins. You wouldn't be able to do anything with the money anyway. It might even cause problems because it would cause people to scrutinize what's going on. So with a non-profit, the expense statement will scale to meet the available funding on autopilot to make sure the thing goes to break even. This is the counter argument to people who argue that non-profits are somehow better lined up to do good things than for-profits. What if they're actually wired to just grow expenses to the moon on the taxpayer dollar?

Ben Horowitz: I think there's definitely an aspect of that. The university math is a little bit: if we can raise prices, then we will. They benchmark against each other. Student loan money pours in, Harvard raises their tuition, and we can raise our tuition, and it just cascades down the system. That translates into higher salaries, more administrators, more this, more that. But it certainly needn't be that way. I think the idea of lowering tuition is a good idea. If you were to start a university from first principles, why would it cost $60 or $70,000 a year? That's outrageous to educate a student in the most efficient way possible.

Marc Andreessen: Yeah, well, it's one of these things where you can look at what it cost 20 years ago and look at the fact that cost has risen much faster than inflation. And then you can just ask the question, are the results better than they were 20 years ago?

Ben Horowitz: Yeah.

Marc Andreessen: Right, so it's by definition something you see in business also: "What if we just went back to the cost structure we had 20 years ago or five years ago?" You ask the people who are actually running it, by the way. But if you're looking outside in, it is a really key question. In other words, you've proven historically that you could do it at a lower cost structure because you were doing it at a lower cost structure.

Ben Horowitz: Yeah.

Marc Andreessen: The reason I bring it up is you cannot find a university today that is trying to do what you're describing. As you said, they benchmark against each other. They're all on the same track. So you could say that's some sort of inductive proof that it's not possible to do what we're saying because nobody's even trying. But the counter argument is that it is possible to do it, and we know it because they were all doing it 20 years ago.

Ben Horowitz: Yeah, and three times the way. Inflation-adjusted tuition has grown at more than double the rate of inflation, something like triple the rate of inflation on a sustained basis.

Marc Andreessen: Yeah, clearly.

Ben Horowitz: The other thing that's really interesting is, if you're at $60,000 a year or above, that's probably a little more than the cost of having a full-time, very smart instructor for your child or your teenager. You could literally assign every student a really good instructor to teach them all these subjects full-time, no other students, just one student. It would seem like you'd get a potentially better outcome with that method. So once you get to that level of absurdity, it's probably time to take a look at it.

Marc Andreessen: Let's talk about that for a second because there's actually a lot of historical evidence for what you just said. If you go back in time, aristocratic education in prior societies was always one-to-one tutoring. The offspring of royal families were always tutored one-to-one. You have these amazing historical precedents. Alexander the Great is the apotheosis of this because his tutor was Aristotle.

Ben Horowitz: Yeah.

Marc Andreessen: Right. And it's just pretty good. Even the Greek philosophers like Socrates and all those guys, their day job was what they did in the mornings: one-on-one tutoring to kids in Athens. And then in the afternoon, they hung out at the Agora and talked about things. But their job was actually tutoring. That was obviously an amazing civilization. So there's a lot of specific historical precedent.

Marc Andreessen: And then there's this thing in the education research which is really striking. One of the things in education research generally is that it basically fails. All of the attempts to come up with systemic interventions to improve educational outcomes basically fail. This has been the case for many decades.

Ben Horowitz: Pedagogy and things beyond tutors. Is that how you think about it?

Marc Andreessen: Yeah, they just don't. Most things you want to do, whether it's Head Start or laptops in the classroom, any number of things where people try to inject money or new practices into the classroom at any level, the null hypothesis keeps proving out over and over: they just don't change anything. The Gates Foundation, which was very heavily involved in education philanthropy, actually put out a report a couple of years ago where they go through this in detail. It's very discouraging in the sense that it's easy to say a good teacher will do better, but it's much, much harder to say we're going to make a million teachers better. Even the things that work at the micro level just don't scale.

Marc Andreessen: There is one exception. There is one educational intervention technique that reliably generates better outcomes. In fact, it generates what are called two sigma better outcomes. Two standard deviations. It's an intervention that routinely takes kids who would score at the 50th percentile of outcome and moves them to the 99th percentile of outcome. And it's one-on-one tutoring.

Ben Horowitz: Oh, yeah, one-on-one.

Marc Andreessen: This is also true with autistic kids. They've done similar research, and the one thing that works is one-on-one tutoring. There is a professor at UCLA, Ivar Lovas, who kind of proved that out. So it's very consistent across all types of students, interestingly.

Ben Horowitz: Right, right, exactly.

Marc Andreessen: There's this thing called the Bloom Two Sigma Effect. The researcher who did the work on this, his name is Bloom. It's this great white whale of education: we actually know how to make education much better than it is. It's just historically been economically impractical. There's just no way you could afford to have a one-on-one tutor for every student. Until now.

Ben Horowitz: Until now.

Marc Andreessen: We have a suit of long money to go to. Quite possibly, right?

Ben Horowitz: Well, by the way, it's also good. One of the big critiques of academia from people like us is that when you go into academia, you're in this sort of bubble of a world. If you're coming up with new social sciences theories, you're testing them among people like yourselves. But if you go back and say, "Socrates' ideas at least had to stand up to his students in a much more direct way because it's one-on-one." They're going to have questions. If you elevate yourself to "Here I am, king of the class, and I'm going to give you a grade, so you better not say anything nasty about my research," that's a very different kind of thing. So it could be helpful on both sides, although it doesn't quite take the expenses down, but it would hold them steady. Maybe one tutor for every three students or something.

Marc Andreessen: Yeah, but look, the expectation has to be that if current trends continue, tuition will keep rising at 3x the rate of inflation. We'll have a follow-up to this podcast five or ten years from now, and it'll have crossed the million dollar mark per student. At that point, the economics become quite overwhelming in the direction of one-to-one instruction. Sitting here today, it sounds crazy that you would make this switch, but it's starting to sound sane. So it's interesting to have an index to a potential competitive system, at least for the sanity check.

Marc Andreessen: So anyway, with that in mind, we immediately launched into one of the more pie-in-the-sky ideas. But let's go back to the challenge of being put in charge of one of these institutions tomorrow and being responsible for the turnaround or the reform that needs to happen. A lot of smart people at the trustee level, president level, and donor level are trying to reform the existing schools. So let's talk about that.

Marc Andreessen: Ben, let's talk about the "fix the university" kind of plan. You wrote an outline prepping this. So if you want, I can go through it point by point, or you could just launch into it.

Ben Horowitz: Let me get into a thing that will illustrate the customer problem and the systems thinking issue. I hate to get into it because it's controversial, but I'm not going to get into the controversial aspect, which is this whole diversity, equity, inclusion and how these programs are designed. I'll contrast it with the way we designed our program, which we call a talent program, but it's essentially the same sort of thing. We designed ours with the potential employees in mind. I think the system that was designed for the university was designed more with the press in mind. "How can we have a population that reflects America?" was the goal, as opposed to "What's the best product opportunity for students in populations where we're not doing a good job of recruiting them?"

Ben Horowitz: Like very different ideas on how you would start the design. And so if you start the design with, okay, we need 14% black students, whatever percent Jewish students, percent white students, et cetera. And by the way, we've got legacy and this kind of thing. Then that forces you into a methodology that is kind of whatever race or gender based where you're literally having them self-identify on their applications and then trying to funnel them through there.

Ben Horowitz: Well, what's the problem with that? There are a lot of very weird side effects. First of all, if you look at graduation rates or outcomes for diverse students, they are much worse than for your main students. If you're designing for the student, you would never design it that way. That would be an important metric, but that's not the important metric. The important metric is how many you let in. So that's corrupting.

Ben Horowitz: The second corrupting thing is everybody knows about that checkbox. Once the student arrives, they're a little bit of a second class citizen because people are going to judge. You can say, "Well, that's racist to say that." Well, yeah, but you set it up so people don't unsee. I apply, I check my Asian box or whatever. I know there are other boxes. I read the news, so I know how that works. That whole design, based on trying to achieve a goal so that the New York Times says you're not racist as a trustee or faculty, gets you to that outcome.

Ben Horowitz: Now contrast that. How would you design it if you're designing for the customer? I can tell you because we started the firm. It's important to know this is pre-Me Too, pre-George Floyd, pre-when anybody cared about any of this stuff in Silicon Valley. We said, how do we get competitive advantage on talent? We thought, there are certain talent groups that don't get recruited and certain talent groups that are over-recruited in Silicon Valley. Computer science students from Stanford are heavily recruited, MIT students, so forth. So we started with the second tier—the top students at second and third tier universities in computer science. Can we go after those? We put them on our list.

Ben Horowitz: The second one was veterans. Veterans don't make their way to Silicon Valley often because they just don't know they're welcome there. That could be an advantage. They tend to be good at things we need: very loyal, trained in leadership and process development, things we really lacked in technology. So we had a team to recruit veterans. The other two groups, if you just looked at the numbers that were way lower, were blacks and Hispanics. We had teams for that. I was in charge of that part.

Ben Horowitz: By the way, note that because it's a talent program, we didn't hire anybody to run diversity. These colleges have hundreds of people to run their programs. But if it's just talent, then the people in charge of talent run the firm. So early on, probably 2010 or 2011, I'm working on this problem. I'm getting input from people I know who have recruited from those populations, who know the special skills that might exist, how to attract people, and so forth.

Ben Horowitz: This is a very funny story to me. I'm at lunch in Menlo Park at a restaurant called Stacks with Steve Stout, who came from the entertainment industry, music in particular, rap music in particular—dominated by African-Americans. I'm talking to him about it and he says, "Ben, do you know why there are no black people in Silicon Valley?" Just like that. I said, "No, why?" He said, "Look around. There are no black people in Silicon Valley." He was literally the only black person in all of Stacks, which is a pretty big restaurant. So I looked it up. The United States is 14.3% black, but Palo Alto, Menlo Park—you're talking 2%. Right away you go, okay, this isn't even an attractive place to live. People haven't even moved here. Maybe we need to start there.

Ben Horowitz: So what did we do? We did film screenings, gatherings, meetups, and barbecues. We tried to get to know the people we wanted to recruit, to know what would make for a good work environment for them, by just spending regular time. Then we started to go, okay, here's a place in the firm where we need this talent. We already know all the people. We know who's the best. We've invested the time, and we're going to get them. When they come to the firm, our retention and promotion rates are very good because we were always focused on the talent. We never cared about the New York Times because at that time, the New York Times didn't care about it. You get to a very different outcome when you focus on a different customer.

Ben Horowitz: I'll tell the last part of the story just so I can map it back to the university. About five years ago, Henry Louis Gates, the very famous professor of African American history at Harvard, called me up to raise money for the Hip Hop Archive at Harvard. He said, "Ben, I want to create a fellowship called the Horowitz Hip Hop Fellowship at Harvard." I said, "Skip, if you call anything that, everybody's going to hate me because 'Horowitz Hip Hop' sounds really messed up. I have a friend, Nas, who deserves to have a hip hop fellowship named after him. I'll call Nas and see if he wants to do it." We went through that and called it the Nas Hip Hop Fellowship.

Ben Horowitz: They wanted to have a big event at Harvard and invite Nas there, which I did. I got a call leading up to the event from Lisa New, who's married to Larry Summers, who was interestingly fired as president of Harvard for saying something non-diverse. She said, "Ben, I've been reading Nas's lyrics." I said, "You've been reading his lyrics?" She said, "Yeah." I said, "You haven't been listening to his albums?" She said, "No, just reading the lyrics." She said, "He's so good. I can't even believe how good this guy is. I'm talking Ralph Waldo Emerson, Walt Whitman. He's that level good." I thought, wow, that's amazing.

Ben Horowitz: The thing I realized when she said that is Harvard never recruited Nas or anybody like Nas. Black people dominate music in the United States. So why aren't you looking for the talent, doing the right things to recruit them? Why are you looking for the color? I think a lot of the things the universities do—if you're going to take a systems view of it—you've got to start back there and say, how do we find the talent that our regular process doesn't get to? Change our process, change our way of doing things to get to that talent. It's going to be better for us, better for the talent, better for the mission, as opposed to letting some outside force tell us what we should be doing. I think that's really emblematic of a lot of the things that have gone sideways in the university when it comes to diversifying the student body.

Marc Andreessen: Let me steel man the question you get in response to this. I'll use my special skill of being obsessive on all these topics, so I can think about it from all points of view. I'll play super woke here for a second. The whole point of the DEI programs at universities is to try to make sure that every field, like engineering as an example, has equal representation by population. It feels like you're arguing that we should give up on that in favor of an approach that sounds like it involves stereotyping. You're saying we should take away the focus of recruiting black people in the engineering program and increase the focus on taking black people in the music program. And we should do that on the basis of a stereotype that black people are better at music than at engineering. How do you explain that given the moral framework people are starting with?

Ben Horowitz: Well, look, I think the truth of the matter is that everybody's an individual and should be treated as such. But I think anybody with half a brain who's observant can observe some very obvious things about different populations—that they're interested in different things. People have different interests. Forget about talent or this or that. It goes by group and by culture. Every comedian makes jokes about men liking to sit around and watch sports and women liking to watch other things. That's very bad to say these days, obviously. But if you look at job categories, veterinarians are 80% women. Nurses are much higher than that. Those are good paying jobs. Psychologists at the university level are 90 or 95% women now. There's nothing wrong with that. Similarly, coal miners are almost all men. People who work on oil rigs are almost all men. MMA fighters are mostly men, although there are women who do it. There's nothing wrong with that.

Ben Horowitz: In any program where you're trying to get diversity—and the whole point of diversity is diverse interest, diverse talent, or ought to be anyway—if you go into a population with a different culture and very likely different interests, and by the way, everybody's got different genetics too, then you have to be a little more creative than just saying, "We're only going to look at these test scores and these grades and these kinds of courses." Obviously those populations aren't showing up on that. That's why you have the program. You can either go, "We'll just still put them there where they haven't shown an interest or an aptitude yet," or we can broaden our criteria to things that make big contributions to society and are important. People may be able to do better than our friends the white and Asians over here. That's the choice you have. You're not going to get equal distribution just like you don't get equal distribution in almost anything in life. It doesn't even make sense math-wise.

Ben Horowitz: The population is 14%. First of all, 14% in total across age groups, not 18-year-olds. So you're already off. Then we're going international, so you're way off. You're mapping to a number that's a fake number, and now you're not just mapping it to the university—you're trying to map it into every subject. This is almost the American epidemic of innumeracy affecting the whole logic of how these programs work. The problem is it's to the detriment of the people you're trying to recruit.

Ben Horowitz: I'll give you a very good anecdote. You and I just had breakfast with a prominent trustee at one of the most important universities in the world. He made a very offhand comment that he thought was the most obvious thing in the world, and it stopped me in my tracks. He said, "Look, even if we accepted every black applicant, we wouldn't hit 14%." You're so unattractive as the top university in the world to that population that you can't even touch that number. Your problem is you don't even know what you're looking for. You don't know what they're interested in. You don't know how to create an environment that's beneficial where they're going to have great careers coming out of it. You haven't done any of the real work. You're just trying to meet some number that doesn't even make any sense. That's the issue.

Marc Andreessen: I'll take it a step further, just to double down. This is a real issue right now for many people, including the people everybody's trying to help. Your friend Henry Louis Gates, longtime professor at Harvard, was pointed out 20 years ago, all the way back in 2004. There's a big article in the New York Times archives. The two sources were him and Lonnie Guinier, who was one of the top black law professors in the country at the time, also at Harvard. She was almost on the Supreme Court at one point. These two very highly respected black scholars made this case at the time. They said, look what these institutions are actually doing: they're bringing in African and West Indian immigrants to satisfy the African-American quotas. That is exactly what's happening. There are more Nigerians at Harvard than African-Americans.

Ben Horowitz: Right, right.

Marc Andreessen: At an individual level, fine, great. I'm totally in favor of having highly talented Africans. By the way, Nigerians have nothing to do with Somalians genetically. This is the other weird thing about this race theory people are promoting—they're not even related to actual races or genetics. It's just some weird government category.

Ben Horowitz: Government category, over-categorization.

Marc Andreessen: Specifically back to what everybody thinks they're trying to do—if you're trying to help African-Americans, and your answer is we have to bring in lots of literal Africans and West Indians or other population groups to do it, then something has gone wrong in pursuit of the goal of helping African-Americans.

Ben Horowitz: Yeah, I think it's clearly not working. With the ruling on affirmative action, there were so many studies that showed it didn't actually help African-Americans over 50 years. Forget whether you think it's a good idea or a bad idea or fair or not fair. The results were really bad. I think part of the reason the results were bad is the way the programs were designed, and that most of the people were African anyway, not African-American.

Marc Andreessen: Right.

Ben Horowitz: Exactly.

Marc Andreessen: Well, the other benefit to the Africans is they're more likely to pay full freight, right? So from a university financial standpoint.

Ben Horowitz: Yeah, that's right. The other incentive, if you're taking expenses through the roof, you need some people paying full freight. Which turns out to be immigrants.

Marc Andreessen: Okay, good. So that's obviously a big chunk of the reform, fix the university thing. Ben, let's walk through the rest of the Fix the University turnaround plan.

Ben Horowitz: Yeah. One big thing is the credentialing system. This is probably the most important thing to the student: once I pay all that money, once I spend all that time, at the end of it I've got something that's very valuable. There was just a report that half of companies are dropping their bachelor degree requirement, which kind of says the credential no longer means much. If you don't like the SAT part, then you probably need to fix the SAT part. Make it better. But it does prove something to employers. It's very hard to get rid of a measure like that because what are you replacing it with? You're replacing it with grades. We talked about massive grade inflation last time, so that doesn't really work. Recommendations? What are you doing that an employer can't do themselves?

Ben Horowitz: The brilliance of the SAT, by the way, is as an employer, it's actually illegal to do a general aptitude test. If you're looking for an all-around mental athlete, you'd like to have some aptitude measure. If you don't like the aptitude being measured, then enhance that. But to get rid of it is kind of nutso from the student perspective.

Ben Horowitz: The second thing is grade inflation itself. In a way, it's easy to fix: just mandate a grading curve and go back. C is average. F is fail. A is two standard deviations up. B is one standard. Just have it mean something very straightforward, the grade.

Marc Andreessen: Are you talking about zeroing in on the absolute grades with the indicator, or literally grading on a curve with a forced distribution?

Ben Horowitz: Literally grading on a curve.

Marc Andreessen: Okay. Grading on a curve used to be more common in educational settings and also in employee evaluation settings. Companies like Microsoft used to do it famously, and GE. It feels like it's gone very much out of style. The criticism is it sort of forces you to guarantee that you're going to have people who don't make it. What if everybody in the class is actually really good? Then you're singling out people who have to be cut from the bottom because you're forcing the grade on the curve. Isn't that unfair? I haven't seen anybody grade on a curve in many, many years. How would you re-explain that to people in a way they'd think it's a good idea to bring it back?

Ben Horowitz: It's different at companies and in universities. I think it works better in universities. The trouble you run into at companies is you're relative to the next employer. If you hire the top thousand employees in the industry and then you rank them on a curve and fire the bottom 10, those are better than the people on the market. You get into that kind of thing. The companies that were famous for it, Intel and Microsoft, would actually fire the people at the bottom of the curve. That has implications. One could argue Intel and Microsoft did pretty well. They started those programs before they were giant monopolies, and they became giant monopolies. So that actually kind of worked. The retrospective that that stuff is bad is questionable.

Ben Horowitz: In the university, if you grade that way, you don't have to fire anybody. If you feel like the failing point should be lower than two standard deviations below the average, you can do that. But having clarity on who's the top student at Harvard is pretty powerful. And having clarity on what it takes to get through four years and get a bachelor's degree—to your point last time about conscientiousness—it really fulfills that promise. We've gone into this "self-esteem is all that matters" thing, but the result is massive student debt. We just lied to everybody. "You're doing fine, you're doing great. It's all good. Give us your money." Then guess what? You owe $300,000 and you can't get a job. That's a much bigger problem than the self-esteem problem or being three standard deviations below the average and flunking the class. At that point, you probably don't know the material well enough to get a passing grade, let alone what you get now, which is an A minus.

Marc Andreessen: Right. The other response I think would be: every professor, every teacher, every manager, and every employer are fully aware that you have a distribution of talents, capabilities, results, and performance in every group. There is no employer that does what you said a few minutes ago—hiring the thousand best people. Nobody ever does that. Everybody would love to, but nobody ever does. You always have a distribution of talents. It's not just smarts. It's effort. Effort is a big thing. That goes into not only a grade in school but also performance on the job. The other thing you learn in life when you employ people—and every employer knows this—is not everybody is good at everything. The important thing is, what are you great at? Where can we put you in a position for your highest and best use? We use that term all the time. Where can you make the biggest contribution? Let us get you there and not have you do something you're no good at.

Marc Andreessen: I think this whole anti-credential way we've got in universities—we have people come out of school and we don't know what their highest and best use is. They may be the greatest... I'll tell you, Robert Smith, the CEO of Vista, the founder of Vista, is very sharp on this. He's a person I probably rely on most for ideas in this area. One of the things he says is, "We get guys and we have to reclassify them. We look at somebody as an engineer, but when we look at their personality, they'd be way better in sales. We reroute them and their career goes much better." If you're a genius psychologist or social networker and we force you into STEM, that's not good for you. It's not good for anybody. But that's what these ideas do. They force people into things that aren't their talent, their skill, their passion, their interest. They don't enjoy it. I'd be resentful if somebody did that to me. I'm not going to make much money, and I'm going to be a low performer wherever I am. There's no need for it because I'm a super talented person. Why are you doing this to me?

Marc Andreessen: And I think these ideas of people who want to make the world fair, they impose this thing on people where life isn't fair in that not everybody is exactly the same amount of good at everything. But there's so much variety that there's a place for everybody. I do believe that there's a place where anybody can contribute, and we've got to find that for them. That's what the university should be doing—finding people's contribution, not channeling them into something they don't want to do because we need another person of your race doing this.

Ben Horowitz: There's something in the data. The social scientists who studied this refer to it as the Scandinavian paradox. It's very counterintuitive but consistent with what you're saying. The societies that are most egalitarian have the greater dispersion, the greater difference between representation of groups by profession.

Marc Andreessen: Right, way more women go into STEM in Kazakhstan than in Sweden.

Ben Horowitz: Exactly. The truly representative STEM systems at the educational and professional level for science, math, and engineering were the Soviet Union and apparently even still today, Iran.

Marc Andreessen: Yeah, where women have much fewer rights.

Ben Horowitz: Yeah. And everything else that women might want to do is much more dangerous.

Marc Andreessen: Very good point. Being a literature professor in Stalinist Russia was super dangerous, but being a nuclear physicist was a super privileged position because they need you. So if you're a highly capable woman who would love to be a literature professor, in that system you can't do that. You go do the thing you'd rather not do because it's the safe thing to do.

Ben Horowitz: In contrast, if you rank societies by gender egalitarianism, the Scandinavian countries are top of the heap. In Scandinavia today, engineers are like 85% men and nurses are like 85% women. It's a much more unequal outcome.

Marc Andreessen: The explanation is very subtle. If you take out all the societal bias, all the societal-level determinism, every possible restriction on what people can do, and let them fully express themselves, what you're left with is pure choice. The differences in outcomes based on pure choice maximize. They don't minimize. The freer people are, the more dispersion you get in exactly the way that makes the prevailing morality completely freak out.

Ben Horowitz: To your point, you can imagine a university that has this polar opposite view. It would have spectacular programs in music, poetry, psychology, and all these different things, and then equally good engineering or math programs. Everybody in every program is there because it's the thing they really want to do in life and the thing they're best at, without trying to force fit everything to be equal representation across all fields.

Marc Andreessen: That's clearly the right approach. In business, we're a technology venture capital firm and we end up needing all those things. Part of our advantage is that we're a firm that has some poetry to it, that can tell a story, that can do these kinds of things. That's how we built the whole brand. It had nothing to do with math. Then we're a network, so we've got to make friends in Silicon Valley, Washington D.C., Hollywood, Wall Street. This isn't a great job for an engineer. There may be engineers who are good at it, but I'll bet I can find somebody else who's way better at it. The world is very diverse in terms of things that need doing. To force people into a path because you're worried about some buddy covering diversity at the Washington Post or the New York Times—the last thing I want is to be tagged with racism. And by the way, those organizations aren't diverse in that way either. They're telling you how to run your business, but they don't know how to run their own.

Ben Horowitz: If you Google "newsroom diversity crisis," there are absolutely hysterical reports excoriating the same news organizations that criticize everybody else. Of course not everybody wants to be a damn journalist in equal representation to the population.

Marc Andreessen: To go back to credentialing, you want to think harder about the credentialing on the way in—how you're sourcing talent and bringing in lots of different kinds of people. And you also want to think about the credential out, the value of the credential you're generating. They're related because the incoming credential is part of the outgoing credential. You want to think hard about how that translates downstream to the potential employer.

Ben Horowitz: If you want to diversify what you're getting through that credentialing system, widen it. Take more things. Have them do a poetry test, see what their rhyming skills are like. Add music. These are real tangible things. I would love to know how good someone is at human relationships coming in. Nobody's tested that. If you only want to test writing and math and history, you're going to get people who are good at writing and math and history and whose culture puts them that way. If you mix up the population, throw the cultures together, do a Freaky Friday on everybody and then have them take a test, maybe it'll work better. But don't make talent the color of your skin or your gender. That's dumb.

Marc Andreessen: There's a psychologist with a creativity assessment battery test. It has 15 different kinds of creativity—poetry, literature, visual art, music, computer coding, whatever. It has seven degrees of aptitude. Degree one for poetry is "I have written a poem in my notebook that nobody else has ever seen." All the way up to "I've won a national poetry award." He said if you apply this test to any broad-based group in the population, the average result is zero.

Ben Horowitz: Most people have never done any of those things. Most people have no interest in doing those things. Or maybe they haven't been encouraged because those things aren't valued highly enough or they don't think it translates to a future path in life. But for people who have done those things, you could have a completely different kind of criteria. I guarantee it comes out very different across different populations, and it's not the same populations that score high on the SAT.

Marc Andreessen: This is probably my greatest disappointment about the lack of evolution of the universities. We rely on them to help us get to much more of the population and much more of the talent, and to help us map their students to our needs. Instead, they just wanted to pass this political litmus test. It's a blown opportunity. In my friend groups, the interests are just so different. People in my white friend group or my Asian friend group are always surprised at how much I know about popular music. Nobody in my black friend group is even surprised at all, because that interest in music is just higher in the population. Cultures are different. Things that are important in the conversation are different.

Ben Horowitz: Let me give you my other credential thing. With the big Supreme Court case on admissions, Harvard was the target, but they were representative of the entire category. We got a tremendous amount of data from the inside. One thing became very clear: these universities are constantly asked why they don't just admit on the basis of objective criteria, purely academic merit and SAT score. Their response is that there are now too many kids who score 800 on the SAT. If they only recruited on the basis of SAT, it still wouldn't help them because there are too many kids who score 1600. Specifically, there are too many Asians who score 800 on the math SAT and do very well on the verbal SAT. It's no longer an effective testing method to even get to the cream of the crop for people in STEM.

Marc Andreessen: There's no reason why the math SAT has to cap out at 800. The test is designed and calibrated deliberately by professionals. They can make the test arbitrarily difficult. They can make the scale arbitrarily high. You could have a math SAT test with harder and harder questions, and then all of a sudden—

Ben Horowitz: Grade inflation.

Marc Andreessen: Exactly right. This is almost the opposite of grade inflation. It's great capping. Cap at 800. It's like capping and saying my students can only get a B plus in the class because I'm not presenting complex enough material for them to validate getting an A. But you could have a version of the SAT with much, much harder questions as the test goes on. You could have a scale that goes from 800 to 2000. You could identify within the crop of people at the 800 level the 10% that are at a much higher level.

Ben Horowitz: Why is it capped at 800?

Marc Andreessen: Because there's constant pressure on the SAT to equalize itself by demographic group. The overwhelming priority at the company that does the SAT is to reduce group distinctiveness as opposed to unearth talent. They already get tremendous criticism for group dispersion of results. If they let it go further up, you'd logically expect incredibly high Asian representation among the people who have higher than 800 math SAT. In the world where we're looking for special and different, you'd say that's fantastic. In the world where we need everybody to be the same, you'd say that's horrible. There's nothing keeping a university or an employer from developing a test like that for itself.

Ben Horowitz: Diversity is our strength.

Marc Andreessen: Exactly. This is what's so weird about the politics now. We want diversity, but we want everybody to be the same. Make up our minds. If we have diversity, we've got to be able to measure diverse talents and degrees of diverse talent and distinguish those things. If we don't want diversity, why have education at all? Just keep us all dumb as we ever were so we can all be the same. The goal isn't to invent new things or build new stuff or create new ideas or write new movies. The idea is that everybody's the same. The logic gets really wacky. Universities got caught in their own underwear because they weren't willing to have that conversation, which is crazy because the whole idea of the university is free speech and the marketplace of ideas. But those ideas got shut down.

Ben Horowitz: Let's go to another option: starting new competition. Some people are trying to do this. Joe Lonsdale and a bunch of our friends are doing it. Lambda School is a for-profit version. Austin Allred is doing that. There's a nonprofit version, University of Austin, which Joe Lonsdale and his colleagues are doing. And there's another one called Minerva. The obvious pro for doing this is the advantage of starting something new—a clean sheet of paper. You can learn all the lessons from the people who have come before you and do what makes sense for today. This is probably the best time in 100 years to try to do that. You have a lot of donors, parents, and students, including students who are very capable but can't get into top universities right now because of radical changes in admission policies. This is the biggest golden moment in probably 100 years to think about doing this.

Marc Andreessen: There are reasons why this would be very difficult. Number one: existing institutions have very powerful network effects, which is why the big ones are hundreds of years old. Number two: it would take a lot of money for a long time because you need to boot up a network. It's hard to get great students until you have great faculty, and hard to get great faculty until you have great students. You'd have to overpay faculty and probably have a much cheaper student proposition. You'd have upside-down economics for a while, which means you'd need a lot of funding. Number three: you'd be trying to break into a cartel. The accreditation process—maybe you could get accredited, get access to federal student loan funding and federal research funding, or maybe you just get boxed out, which would mean you need even more money to get started.

Ben Horowitz: With your entrepreneurial hat and your venture capitalist hat on, is starting new universities in the same shape and form as the current universities an idea we would encourage or warn people away from?

Marc Andreessen: I would probably warn people away from that idea—the same bundle. Although today is probably the right time to come at that idea. From a venture capital standpoint, it's a very long shot. It's the difference between Tucker DeLorean and Tesla. Taking on an existing incumbent at what they do is tough. Taking them on on something they really don't do, like electric cars, tends to work better. With universities, there's such a huge opening for different lengths of degrees. The four-year degree doesn't make much sense to me. To adopt that as your degree length, given that people who aren't scholars need skills and then new skills and new skills—could you imagine trying to do our jobs based on just what we learned in college? It taught me to program in Pascal and C. Knowledge is evolving very fast. The four-year degree is one thing you might bring into question. If you had a shorter degree where you could get just as high-paying a job—this is something Lambda School does—then all of a sudden you have a value proposition that's starting to look really good. Maybe 10 grand, you lend me 10 grand for a year, and then you pay me back if and only if you get a job. That starts to sound pretty good. There are things that would be much more attractive to students that weren't a full frontal assault on Harvard. From a VC standpoint, there's something very noble about building a new full-out four-year Ivy League of the future type thing. If you believe these schools have lost their way, it's time to build a new thing. The first university was invented so long ago. Why don't we invent one for today?

Ben Horowitz: Let's take it out of the realm of pure venture capital and think about philanthropy. There's a donor strike at some institutions right now. There are very deep-pocketed donors on strike. Let's suppose a group of them get together and put $2 billion, $5 billion, or $10 billion to build from scratch. They do the direct frontal thing. They build the parallel thing. Their logic is: number one, we have the money. Let's say up to the $5 or $10 billion level. Number two, we want to go for a full frontal assault because we don't want people to have to rethink their assumptions. We want to bring the faculty over. We want to bring the students over. We want the parents to be totally comfortable. We want the government to understand how to deal with us. We want to fit into the existing industry structure. We don't want to take the risk of innovating. We just want to be a new and better version of the thing that already exists. How would we advise them given those goals and that level of funding?

Marc Andreessen: Our whole mission in life is we're dream builders, not dream killers. We would for sure encourage them. It got me thinking about what I would advise Joe Lonsdale to do. He should wire University of Austin straight into us and straight into everybody in venture capital who is building new companies and hiring lots of employees. Ask us what we're looking for. Let's do a partnership and recruit straight out of there. That will really enhance the proposition to new students. If I'm a new student and I've got a Harvard degree or I could get a University of Austin degree, why am I going to University of Austin? What if during the recruiting process, they come see us and we say we'd rather have you out of University of Austin than out of Harvard? That would open my eyes. I'd take that seriously. I'd lock in on how to attract the best students. It's partly a function of faculty, but it's partly a function of who's paying and who understands enough about that university to say, "I'm all in, and I can be as big a help to you when you come out as anybody." Artificially create what's like the alumni network, but better, because you're doing it with people who 100% have jobs at the top of the job market. I would start there. Design the system to feed us and feed the students into all the most interesting jobs that line up with the curriculum. If University of Austin had a big focus on creativity, I'd want to wire them into some kind of creative output. What happens after University of Austin? When you go to school, you're looking for your life not to be working at a restaurant. I had been a busboy and I was like, "I do not want to do that my whole life. I'll shoot myself in the head." That's what a lot of people are looking for when they go to college—how can I have a life that has more variety, is interesting, where I'm learning and a lifelong learner? If you can guarantee me that life or give me a better product to get me that life, that's what I want.

Ben Horowitz: You sent me a quote as we were prepping for this. I'll read it back to you: "Today's universities are built on industrial revolution technology that are therefore completely outdated for the information age, both in how they run and the product they offer." How would we advise this hypothetical new institution on what to do on that front? What does that mean in practice?

Marc Andreessen: Industrial revolution technology means you can build big buildings. You can drive there in your car or on a train. You don't have to ride a horse. It's indoor lighting at night. It's not too hot, not too cold. That's the platform. So what do you get?

Marc Andreessen: You get classrooms with instructors. You get a dorm that you can live in, a big cafeteria with a meal plan, all that kind of thing. But, look, in the information age, and a giant library, of course, or multiple libraries if you're an Ivy League school. In the information age, you have AI, you can ask it any question. You have access to the internet, you've got all these other things. And then your school experience right now in the industrial age world is you have instructors and administrators filling out forms, and very little career guidance. I don't know about University of Illinois, but certainly at Columbia and at UCLA, it's very little kind of career guidance to help you find your highest and best use. The university didn't really do that.

Marc Andreessen: I think that in an information age AI university, all that form filling out, a lot of the instruction is taken care of. But what you really need is the university to help you find your purpose and then guide you through your purpose with a team of other students who have a similar purpose, to help you study the right things, prove yourself in the right ways, get the credential, and so forth, using all the best tools to do that. As opposed to waking you up at 8 in the morning to walk to class in your pajamas because you were drinking too much the night before, sitting in a class very bored, not really integrated into the AI and the internet to get the rest of the information.

Marc Andreessen: So I just think there's a whole rethinking of the way your day would go. 45-minute or hour-and-a-half lectures are pretty hard to pay attention to the whole time and retain everything. Smaller chunks of work have been proven out, with a kind of test to see if you retained the information, or some interactive part every 10 minutes. That's a much better format. And then, as you said, a one-to-one tutor. But maybe the machine is the one-to-one tutor in some ways, because you can ask it questions now in English or Chinese or whatever language you speak.

Marc Andreessen: I would definitely make it that and give the professors, the mega tutors, the tools to both identify the capabilities of the students and then help them maximize those abilities, and then map it further into people like us. It could be us. It could be the NBA. It could be Warner Music. It could be whatever part of society works. We take people with all kinds of talent. All kinds of different things are very valuable for us at the firm. In fact, we're 550 people, which probably makes us the biggest venture capital firm in the world. Why? Because we do the most things. Why? Because we have the kind of people who can do lots of different things. And that's a heck of an advantage as an employer.

Ben Horowitz: Right, right. So when the topic of technology and education comes up, a lot of people sort of reflexively assume that you must mean the whole thing moves online, everything's over the internet. That's not what you're saying.

Marc Andreessen: No, no, no, no. I actually think it's good for most students. Maybe not for you, but for most students. There are different personality types. But I think there's something very motivating about being around peers. Here's my cohort of people who are going to be in the world with me. What are they doing and what can I learn from them? You learn as much from your classmates as you do from the university. And that's hard to do online. So I think the college experience, to Joe and University of Austin's credit, is a real thing with real value, particularly for a young person, for most young people. But you have to modernize it. We're not in 1910 anymore.

Ben Horowitz: Yeah, yeah. And so with that in mind, let's go more radical. Let's talk about unbundling. Our old friend Jim Barksdale has this famous line: there's two ways to succeed in life, in business. One is you can bundle, the other is you can unbundle. So let's talk about the unbundling. We had our dozen functions of the major university. I stripped out three, which we can talk about at the end, but that leaves nine. It seems to me there's a case you could be great for unbundling. So let's walk through them as actual potential startup ideas, either for-profit or maybe nonprofit or philanthropic.

Ben Horowitz: First, a credentialing agency. We've talked a lot about credentialing so far, but all the different aspects of it. This concept of credentialing in and credentialing out—both the way you decide who to credential and the actual credential you give them. Is that, in your view, something that could be abstracted out and turned into its own thing?

Marc Andreessen: Oh, I think this may be the best startup idea of everything in education. If somebody had an organization that aptitude and personality tested people in very specific things... In Silicon Valley, everybody gives every engineer some kind of test in their interview—write this piece of code, figure out this algorithm. I think every job has some of that. So if you had a place that could reliably differentiate people's capability in things you needed to hire for, that would be very attractive to employers.

Marc Andreessen: The SAT was invented because it used to be only nobles, the elite, the aristocrats, people from rich families got to go to college. Then you're saying, what if I'm some poor kid from New Lisbon, Wisconsin, how do I show I can go to University of Illinois? Take this test. That's a fucking miracle. That was a specific reform at a specific moment in time.

Ben Horowitz: Yeah, yeah, yeah, exactly.

Marc Andreessen: Among employers, people are worried about bias. To have the ability to show your capability in any dimension and have an employer know about it and have it be valid—we'd be incredibly interested in that. People who didn't have college degrees, who might have gone to a state school that was a little cheaper that they could afford, all of a sudden... That would actually help fix the university system. Now I can go to San Francisco State and get credentialed here, and I'm actually more interesting to Andreessen Horowitz than the person from Harvard with this degrading credential. How about that? And I spent a hell of a lot less money to go to San Francisco State. That would be incredible.

Marc Andreessen: So I think this is such a great startup idea. I've been thinking about it since we started the podcast—how do we fund that one? That's awesome. But you'd have to do a great job on it. It would have to be unfudgeable. Nobody's bringing ChatGPT into the thing with them. Maybe you just have to know what to ask. But something where you knew if they could do that, then they would have that capability. Then as an interviewer, you're just really going into understanding motivation, cultural fit, as opposed to can they do the job, because you know they can.

Ben Horowitz: There's also something in the law. As you pointed out earlier, it's illegal. There's a famous Supreme Court case that made it illegal for companies to do generalized aptitude testing. It's illegal to do a general test and rule someone out of a specific job based on it.

Marc Andreessen: That's right. That basically killed IQ testing at the employment level. That's when the university degree took off, because the SAT score was an implicit IQ test and it laundered through. So employers outsourced the IQ test to the university credential. But as we discussed, 1,700 universities and colleges in the US have stopped using standardized testing as an admission criteria, so the value of that is going to zero as an IQ test. They're doing everything they can to get away from that, but the employer still can't do it.

Marc Andreessen: What's interesting about this as a startup idea is that the thing the Supreme Court said is illegal is an employer can't do this. But here you could have any kind of aptitude testing, IQ or otherwise. You could have a dozen or a hundred different ways of measuring aptitude in whatever domain you want, including creativity and everything else we talked about. It would all be completely voluntary. It's completely legal because it's not tied to employment. So you could do a super version of even what the universities did in the past and actually have it be a fully legal thing.

Marc Andreessen: By the way, it would really help get people into the right jobs. Sometimes people get miscast. Life is like that. You get assigned one thing and you really should be something else. This kind of rigorous assessment might identify that. You can find something you're better at and do better at in your career. We could use more of that, which would be great for us. This is definitely high on the "it would be great for us" criteria.

Ben Horowitz: Yeah. Yeah. Okay. Good. Good. All right. Second is actual educational coursework itself. Of course, there have been attempts—the MOOC online course movement a while ago: Coursera, Udacity, Udemy. And then Khan Academy, which is a different format.

Marc Andreessen: Yeah, exactly.

Ben Horowitz: I'll give you a couple thoughts on this one. Number one, this has already happened in Korea. There are actual teaching superstars in Korea that make big courses. I've had this idea for a long time: if you've got a million kids who are going to take Math 101 freshman year of college, and you get each of them to do a hundred dollars for it, that's a hundred million dollars of revenue. Then hire Steven Spielberg to make Math 101 as a video miniseries. The most compelling courseware of all time. Literally get Steven Spielberg, Christopher Nolan—these guys—to make the most mind-blowing, incredible course lectures you've ever seen, with full 3D effects and graphics and everything perfect.

Ben Horowitz: You could do that, and then couple it with AI tutoring, or in-person tutoring, or matching grad students to undergrads. People have been trying variations on this for quite a while. How would we think about that as an entrepreneurial opportunity today?

Marc Andreessen: Yeah, it's interesting. Of the ones we named, the one that has worked the best is probably Khan Academy, which had the least amount of money going to it and is in its own format. It's these little lessons. When AI started taking off, I had a hard time remembering how to do linear algebra or how hard it was. I did the Khan Academy and thought, wow, I forgot how easy it was. It's much easier than actual algebra. It sounds harder, but it's easier. It's kind of a magical thing.

Marc Andreessen: I think the challenge with the full college experience, unless you get to the Christopher Nolan version, is that it's a very big thing to sign up and commit to without a well-known credential. If I'm really coming from outside the job market, even if I learn it, will anybody believe me? How does that work? I think completely decoupling that one from credentialing or jobs may be tough. But if you could link it into "you take this class, you get a job," then I think that could definitely work. Otherwise, it's a very small market of people who just really want to learn that much about a subject.

Ben Horowitz: Right. So when you talk to professors and university administrators about this kind of thing, what they tell you is that you can't be naive about actual real-world students. They say a lot of students actually don't want to learn, or they're not motivated to. They're not driven to do it. To the extent they actually show up to class and do the work, it's because they're in a specifically structured environment where expectations are set high, and they'll get kicked out if they don't. Their parents are paying for it. They're basically pressured into doing it. When I went to Columbia, which is pretty high-end, I imagine that at regular schools it's even more the case.

Marc Andreessen: I think there might just be two different kinds of students. There are the ones that are super intrinsically motivated, where they're just like—you talk about yourself as an example—you were a busboy, you didn't want to be a busboy. You were a dishwasher, you didn't want to be a dishwasher. You're going to go do this because you know you need to do it for intrinsic reasons. Those students exist. But then there is this other kind of student that arguably more populates the upper ranks of American education, where they actually need to be in a highly structured environment.

Ben Horowitz: I agree. Yeah.

Marc Andreessen: It's almost like the more privileged the student, the more pressure needs to be put on them. If your life is really fun, when you're young, life is incredibly fun. Everything is new. Every movie is amazing. Every experience is incredible. So I'm going to take those years and sit in a classroom listening to somebody drone on about the Iliad? I don't want to do that. Whereas if your life is misery without getting something out of this experience, that's a little more motivating.

Marc Andreessen: I always wonder if people should be more specifically addressing that category of intrinsically motivated students. Just not try to appeal to the people who need more structure and pressure.

Ben Horowitz: Yeah, yeah, yeah. I think that's a very specific market. You'd have to kind of identify them and get them through. But again, if you then link that to the credential, they would see the cause and effect and that would be very clear. This is basically describing Lambda School in a lot of ways. This makes sense.

Ben Horowitz: Okay, third is the Research Bureau. This one freaks people out because anytime you bring up a different way to do research or fund research, basically everybody in the research complex freaks out. The steel man case against any change is that you don't understand the whole point of basic research. The whole point is that it doesn't have an end goal in mind and it can't. How do you know you're doing some research experiment in physics or a new math theorem or decoding the genome? Maybe there's a commercial use case 30 years from now, but you have no idea. That's what makes research different from development. Research and development—development has a specific goal to ship a product and make money. Research is trying to come up with new knowledge.

Ben Horowitz: The modern research university was constructed by Vannevar Bush and his peers 70, 80 years ago to provide an environment in which basic research can happen. That's why professors have tenure—so they're free to do whatever research they want, not risking getting fired if they don't deliver something practically useful. The government funds research because companies won't fund basic research without an end commercial target. You get tremendous pushback out of the gate on this conversation. Having said that, I think there are very interesting things you could explore as ways to do research outside of the university context. Some people are doing this already.

Ben Horowitz: There are a couple of issues with the current research complex. There's a massive replication crisis. Depending on the field, up to 75% of research doesn't replicate. There's a massive incentive problem. How do you get a grant? You publish a research result that seems to validate additional investigation. How do you get that result? You get it legitimately or you do data mining. How do you get tenure? You publish papers that you may or may not have written yourself.

Marc Andreessen: A friend of mine, a very experienced guy who's been in leadership positions across this entire spectrum, whenever we talk about this, I go on about the replication crisis because I think it's such a scandal. He says it's not even the problem. He says the problem is 90% of the research is just useless. It's not helpful. Whether it's right or not is secondary to whether it would even matter if it was right. This is a guy who ran a major government research funder. He's been on the inside seeing how all the sausage is made.

Marc Andreessen: He said in any given field of research—quantum physics, psychology, computer science, whatever—there are five institutions that are on the leading edge. Everybody in each of those fields knows who those five institutions are. Those five institutions generate essentially all of the useful output that actually moves the field forward. It's five institutions, maybe a hundred professors by field, some number of grad students, and the research budget for those people. He said, "Why doesn't the government just fund that?" He said that's 10% of the money. I said, "Then why spend the other 90%?" He said, "Because there are too many mouths to feed. There are too many constituents. There are too many congressmen that have universities in their districts. There are too many people who get tenure who expect this. There are too many incumbent colleges that have research programs, even if they're not productive, they don't cancel them." The system is wired to overfund every category by a factor of 10.

Marc Andreessen: He said the first thing you would do is narrow it down to the 10%. Figure out what is the actual 10% of the useful work to be done and the actual 10% of the people who can do that work. The aggregate dollar amount involved is literally a tenth of what everybody thinks it is to do the actual quality work. Then he made a further argument: you don't always know there's going to be commercial applications for research, but a lot of times there is. If you have some material science breakthrough, the patent is probably going to be super valuable. Universities are in the business of patents and patent licensing, and they get revenue streams from that.

Marc Andreessen: He said you could either have new nonprofit research institutes funded with philanthropic dollars—maybe that's tractable because it's just not that much money for the high-quality work. Or maybe it should be a venture capital model, for profit. You make money in the long run on commercial product development and on patent licensing. You should actually apply a VC mindset to that.

Marc Andreessen: Our friend Patrick Collison is funding philanthropically a program to do independent medical research associated with Stanford called ARC, set up as a separate thing. We know the folks at the Howard Hughes Medical Institute that does specific grant funding to individual young researchers in the biomedical field and has had outstanding results. There are new cuts on this that people have.

Ben Horowitz: I'm on the board of Sean Parker's institute, which is similar. It's called PICI—Parker Institute for Cancer Immunotherapy. They fund researchers to do breakthrough work on cancer. They spin the ideas out into companies using a venture model, and the institute invests in them. They generate patents. He made an incredibly generous donation—$250 million, maybe bigger—to start it. His vision has always been that it will become self-sustaining over time because the things it's doing are so incredibly important. I think that's probably right. They've spun out some very interesting companies already, some of which we've invested in.

Ben Horowitz: Then there's the Chan Zuckerberg Institute. Mark Zuckerberg and his wife Priscilla Chan are trying to cure all disease, which is an incredibly great ambition. Not something you would necessarily do in a company, but it'll probably have a lot of commercial results coming out of it as well. I think we're definitely at a point where philanthropy can do it.

Ben Horowitz: The other thing is, there was research before the current complex that got us some pretty interesting results—say, the theory of relativity. Where was Alan Turing when he did his proof?

Ben Horowitz: That's a good question. And then, of course, Claude Shannon was a master's student. A master's thesis, which is considered like nothing in academia, is probably more important than almost any PhD thesis in the last hundred years. He mapped Boolean logic—the algebra of zeros and ones—onto a circuit. That is the beginning of computers, for those who don't follow that kind of thing. So there's real research that can obviously happen outside of the way the current university system works. That's been very powerful.

Marc Andreessen: So, yeah, I think it would be great if there were a couple hundred of these in different categories. It's certainly something I would love to put more money into. I think that's quite a good idea. Whether they make money or not, I think they're very philanthropically fundable.

Ben Horowitz: Yeah, that's right.

Marc Andreessen: I think there may just be certain areas, like particle physics or whatever, where you're still in government money. But I think a lot of the current complex actually might be fundable. The quality of work happening in it might be fundable separately.

Ben Horowitz: Policy think tanks that are not associated with universities—people are doing that today. Moral instruction is my favorite one to think about breaking out separately. There are many social movements that are not associated with the university. There are many social organizations, activist groups, churches. New religions, old religions, which are maybe coming back a little bit right now. Society has a lot of different ways to organize moral instruction that don't necessarily have to take place in a research university context.

Marc Andreessen: Yeah, and I think the university one is very, very tricky because it is counter to a lot of its other goals, particularly the marketplace of ideas, freedom of speech, freedom of thought. I just think it's not the right people to be doing it. Pastors and priests have come under fire, but in a lot of ways, they're much more the right people to be giving moral instruction because they actually spend time with their congregation. They deal with the actual trials and tribulations of going off the moral path and how to get people back on it. It's hands-on, it's tangible. Whereas in a university, a professor can spout off whatever they want, and then has no tie to that down the road. They don't live in their community, they don't have parents who go to the church and donate. I think in some ways, the university is the worst place to do moral instruction now that it's no longer a religious institution. If there was one thing I would take out of a university, it'd probably be that.

Ben Horowitz: Look, I'm a big believer in intellectual discussion and instruction about ethics and morality. I actually believe in that a lot. I do it at work. One of the things we talk to employees about is: in this situation, what are we going to do? Are we going to do what's transactional? Are we going to do what's long-term? Are we going to do what's right? Or are we going to do what makes us money in the short term? That's real moral instruction, and with real consequence that's going to have real impact. I've got to live with the consequence, they've got to live with the consequence. In a way, it's a better context to do it than a university where you don't have any accountability, any moral accountability. So you're better off doing it almost anywhere than in a university.

Marc Andreessen: Or you could reconstitute the original religious university. You could go back to the future. Go back to the original idea.

Ben Horowitz: Shut Galileo's ass down, do that kind of thing.

Marc Andreessen: No, no, no. Not even that far back. The original Harvard business plan was to instruct moral leaders. Instruct pastors and moral leaders. You can imagine an institution that just does that and doesn't do all the other things that have been added on for the last 400 years.

Ben Horowitz: I think that would be very good. In my lifetime, our society has degenerated no place more than in what's right and what's wrong. There's no agreement on right and wrong anymore. Is stealing right? Well, it's right if you're hungry. There's certainly no right and wrong about marriage or these kinds of things. I think that's caused a real degeneration of society, quality of life, outcomes of the world. Fixing that would be great. But the way we do it now is bananas. It doesn't fit into the current bundle. Having an independent moral university would be fantastic.

Marc Andreessen: There are certainly overtly religious ones. The ones that we've lost are the ones that are kind of moral and ethical without being overtly religious, which is a real challenge in general. Anyway, I'll keep going. Sports league—I think you'd probably argue that the sports function could just be its own thing.

Ben Horowitz: Yeah, I think that the sports league is immoral, just fundamentally, in retrospect. The university-based sports league. We talked about that last time. Big-time college sports has gotten to a point where it's clearly immoral. It's very clearly professional sports where they don't pay the employees, and that generates a colossal amount of money. You've got to fix that. I don't know that you can pay them. Probably the right way to fix it is to spin it out and have it continue to be affiliated with the university, but not run by the university. Run by owners of the various teams, something more akin to the NFL or the NBA, because that's what it is. The athletes need to get paid. It's just crazy. It's really wild that they don't.

Marc Andreessen: South Park did a very hilarious episode on this called the National Crack Baby Association. The funniest part was Cartman went to go see one of the presidents of the universities. He's dressed like an old Southern slave master. He goes, "How do you get away with not paying your slaves?" And the president goes, "Slaves? You mean our student athletes?" And Cartman goes, "Oh, yes, student athletes." You're watching and you're going, "Yep, that's exactly what's going on." Now, not all schools' athletic teams make that kind of money. But the ones that do need to be reformed. In retrospect, this is one of those things where people 20 years from now are going to go, "I can't believe you guys did that." That's going to be bad.

Ben Horowitz: Yeah, that's right.

Marc Andreessen: And then the two other ones—these are serious topics, but also a little bit fun. Adult daycare and dating site. Adult daycare—a lot of people just graduate high school and go get a job. Maybe that's overblown. Or maybe you can imagine new design communities, contexts, maybe even entire buildings where you have a social-cultural matrix that people can plug into.

Ben Horowitz: Yeah. To some degree, the armed services have that function. You get to be 18, what are you going to do with your life? You go to the army, you go to college. It's a community that you step into that's not your family, but you can't stay there. Or the corporate campus. For a lot of kids, even post-college, a lot of corporate campuses are designed to perpetuate adult daycare. I'm just trying to think of the proposition to the parents who are paying for it. That seems hard.

Marc Andreessen: We have a company. We have a company that I would certainly not pitch as adult daycare, but it's intended to provide a much more pleasant, interesting, and enjoyable experience for people new in their lives and careers as adults, from a housing and community standpoint.

Ben Horowitz: Yes, that's a very good idea.

Marc Andreessen: We can't pre-announce it, but there is a new idea in the portfolio that creates a place that is like a college dorm from a living perspective, has a community, and so forth. You pay rent there, but rent somewhere that you would in any apartment. It's a nice bridge from coming out of high school or college into the world, where you're not just living by yourself somewhere lonely.

Ben Horowitz: Yeah.

Marc Andreessen: And then a dating site. I'm absolutely furious that the dating apps took off after I was finished dating. I would so much...

Ben Horowitz: Yeah, this is a hard one for me to comment on. I haven't been on a date in over 35 years.

Marc Andreessen: Exactly. So I would just observe there are new ways for people to date today that are much easier than when we were in college. The dating site part of it might already be solved.

Ben Horowitz: There are certainly tools, although physical proximity is not something you can simulate online. And my understanding is there are a lot of fake photos, or people look better in their photos perhaps.

Marc Andreessen: Well, there's also credentialing. That's one of the reasons why a college campus is a hothouse for dating. It's a lot of young people who are together physically. But another reason is because they all have a shared sense of identity. They all have a shared credential—they're all at X college. They've all been validated, at least to some extent. That's also true later on in life. A lot of people who are college graduates want to only date other college graduates, or only want to date people who went to certain kinds of schools. The credentialing thing reflects itself into this other area of actual real life, which is dating, and then ultimately marriage and offspring. We should probably not underestimate the actual utility of an existing college environment for that, but also think about: could your independent credentialing agency also credential you as a viable marriage prospect or dating prospect?

Ben Horowitz: I love that part of the thing.

Marc Andreessen: Okay, and then we'll close. We're at two hours. I think there's basically the fifth thing that could happen: the existing system could just devolve. It could just unwind. The way it unwinds: the credentialing function shifts to the employers. The education courses shift to a la carte internet options. The research function shifts to the kinds of things we were just talking about. The policy think tank unwinds, shifts to independent think tanks. The moral instructor part loses credibility, it withers over time. The social reformer withers—it's arguably happening already. The immigration agency maybe continues. Maybe that's the ultimate business model: just have high-paying immigrants. Sports leagues break out, go independent, become professional sports. Adult daycare, dating site—people just find other ways to live and date. It may simply be that things just unwind. Sitting here 50 years from now, these institutions still exist in some form, but they look increasingly archaic. It's like driving down the street and seeing something and thinking, "Wow, that thing still exists." They're just much less important because they've isolated themselves socially and economically, walled themselves off from the general progress of society. How would you think about that?

Ben Horowitz: Look, I think it's pretty fragile now in terms of the value proposition. It's so expensive, the cost relative to the value is so precarious. As an example, if the U.S. government said we're not going to guarantee college loans anymore, that would be cataclysmic. And if employers were like, "We're not going to value college degrees anymore—there's no SAT score, no grades that mean anything, no rigor"—there are things that are actually reasonably close to happening that could collapse the system or at least really alter it irrevocably. Universities very much have to think hard about shoring themselves up on the value proposition. I just think it's getting very weak for students, and that's a dangerous place to be in if you're in a university.

Marc Andreessen: Yeah. And the other thing I'll nominate—and this sounds a little bit crazy right now—I don't think we're necessarily that far away from a full-fledged political revolt. The constituency of these places is not a majority of society. It's a small minority in terms of the people who actually benefit from the system today in the voter base. A lot of these places have become so politicized, they inject themselves so directly into national politics. They've declared themselves. You see it in every metric: the distribution of professors' ideology, all the social activism that happens. It's overwhelming, and increasingly public. To be blunt at the two-hour mark: right-wing media, for better or worse, is consumed with story after story of crazy, hostile things that universities are doing. You could imagine a point where half or more of the country just puts its foot down, and its elected representatives put their foot down, and they're just like, "We're just not doing this anymore. We're not paying for it anymore."

Ben Horowitz: Yeah, right. If the country took a sharp—or maybe even a not-so-sharp—right turn, you could imagine an administration and a Congress going, "Why are we funding this?"

Marc Andreessen: Yeah. And it's a small number of programs that pay for the whole thing. Federal student loans, federal research funding, a couple of things in the tax law. It wouldn't take two years to figure out how to kill these things from a legislative standpoint. It would take about two minutes. Maybe that never materializes because these things are so important, they have such existing credibility, they have so much political stroke, their graduates have so much power. Or maybe it's one of those things where there's a tipping point, and at some point people are just like, "I'm not going to tolerate this anymore."

Ben Horowitz: I would register that for anybody still listening.

Marc Andreessen: Good, alright. I think we covered it. As predicted in the beginning, we did not get to the Q&A. Ben, if you're up for it, we will continue collecting questions. If there's popular demand, we will do one more—we'll do part three maybe next week and we'll do Q&A. That'll be six hours of content from us on this topic, and that'll probably be enough for a while. But we have enjoyed talking about it, and hopefully you've enjoyed listening to it.

Ben Horowitz: Yes. Thank you.

Marc Andreessen: Thank you everyone.