On ‘Dark Talent’, MOOCs, Universities, and Startups: An Interview with Our First Professor-In-Residence

Vijay Pande


Editor’s Note: Professor Vijay Pande of Stanford University is joining us as the first Andreessen Horowitz Distinguished Visiting Professor of Computer Science, just as we’re bringing together a great group of computer science professors from around the U.S. for talks and discussion for our second annual Academic Roundtable event that bridges academia and industry.

pande1Pande is currently a Professor of Chemistry, Computer Science, and Structural Biology at Stanford University, where he is also the Director of its Program in Biophysics and runs an award-winning lab responsible for several advances in the field. A hallmark of his team’s approach is developing novel software-based approaches to push the limits of existing methods. One example is the Folding@home project, which seeks to understand protein (mis)folding through the use of distributed computing. His lab’s work has contributed to our understanding of Alzheimer’s, Huntington’s, and various kinds of cancer.

But why do we even have a professor-in-residence role here at a16z? Many of the most successful companies that a16z has been associated with have their foundations in academia — from routers to data science to the web browser itself. Yet the process of getting from “bench to bedside” (to borrow a term from academic biomedicine) is hardly a process at all, even at Stanford. We aim to change that.

In this role, Pande will be a16z’s dedicated liaison to academia. With one foot in both worlds, he is a natural first point of contact for any professor or graduate student in a CS or STEM discipline considering commercializing their research. Our message here is clear: If you have an interest in moving your research from academia to the market, Andreessen Horowitz — and Vijay — are interested in hearing from you!

We sat down with Pande to find out more about his views on how universities can support entrepreneurs, the role universities play in innovation ecosystems, MOOCs as the ‘newspapers’ of academia, and just what makes Stanford so special when it comes to startups…

a16z: What do you think Stanford does differently when it comes to the startup mindset?

Vijay: When we started Folding@home in 2000, there were aspects of what we were doing that would have gotten a lot of “no’s” in other universities. But I didn’t get any pushback that this project was weird or strange … the only question I got was about whether it was going to work.

The environment here is constantly looking to see what we can do — not just innovate for the sake of innovation, but to really make an impact in terms of some results.

a16z: There was a NYT article this summer about how Harvard is now considered the ‘Stanford of the East’ — no longer the other way around — because Stanford is “riding a wave of interest” in technology. Which is especially interesting given the record-breaking number of students that recently enrolled in Harvard’s introductory computer science course. What do you think makes Stanford different (at least for now)?

Vijay: Stanford has grown to prominence based on its engineering and medicine and more applied areas, whereas Harvard grew to prominence based on its liberal arts. The former may be a bit more concrete and easier to see than innovation in the latter. So there’s a psychological effect that we see all this innovation around us because it’s more publicly visible.

There’s also a permissive environment in terms of allowing (versus stifling) faculty to do unusual things, to push the boundaries. Other schools are much more rigid about how things are done, much more top-down.

I think top-down is like a benevolent dictator: It sounds like a great idea, but it’s hard to find the right person for the job. The benefit of a bottom-up approach is that crazy ideas — ideas that sound crazy, but are actually the real solution — can bubble up much more easily. And that’s something that’s very much a part of the Stanford spirit. (Part of that spirit also means that there’s not as much excitement for people who do incremental work. You really want to hit it out of the park.)

But we don’t think we’re #1. I’ve heard Hennessy [the university president] say “We should always think we’re no better than #2.” I think there’s a truth to that. It’s when you feel like you’re #1 and that nobody can touch you, that you forget you always have to prove yourself.

a16z: How does Stanford’s connection to Silicon Valley play out?

Vijay: People go back and forth between Stanford and the Valley all the time; it’s common for faculty members to be involved in startups. While it may not be as common for a faculty member to go to a VC firm for their sabbatical [like I’m doing], I think it is very much in the spirit of learning new things and connecting to interesting problems.

The sabbatical system is common in academia, but at Stanford, one option that’s always available is that the faculty has one day a week where they’re allowed to do non-Stanford work. A common model is a professor’s working with some students who graduate or leave to start a company, so the students are day-to-day, the professor’s there one day a week.

When a university’s faculty is involved with startups, they’re thinking about innovation all the time — because in a startup, you have to innovate or die. That type of spirit continues after the faculty member comes back to Stanford and starts thinking about their work there.

It also helps that when you walk around campus there are physical artifacts, like the building [Jerry Yang and Akiko Yamazaki Environment and Energy Building] donated by Jerry Yang, or the building [Jen-Hsun Huang School of Engineering Center] donated by Jen-Hsun Huang. I think their existence proves what students from here can do and keeps it top of mind.

Finally, besides informally “getting it” by what’s in the air or from hearing about friends and colleagues doing interesting things, Stanford’s really strong in terms of entrepreneurial activities. Students aren’t thinking about entrepreneurship after the fact; they’re thinking about it as the work is going on, they’re thinking in real time about what would work.

For example, I’m the faculty manager of the Stanford Bitcoin Group, and that’s a great example of very motivated, smart, entrepreneurial undergrads.

a16z: You mentioned these kinds of ideas bubbling up from the bottom. But surely there are some top-down incentives or other things the university does that helps create this environment?

Vijay: Stanford is really good about I.P. [intellectual property]. I think the university wants to see something come of it much more than they want to make money from it.

To borrow a phrase from computer science, you can take a “greedy algorithm” approach to I.P. — which is to try to squeeze the most money you can in every deal. Or, you could take a big-picture approach, which is, if we’re a good citizen in the Valley, and we do the right thing, good things will happen for us. For example, Stanford would love to see a new drug actually help people; they’re not going to hold it back due to money or details. At Stanford, the Office of Technological Licensing is more like a partner than an adversary. In fact, in their questionnaire (which I fill out once a year), they actually ask: do we feel more like partners or more like adversaries?

The reverse extreme is something like the end of The Raiders of the Lost Ark, where all the I.P. is locked up in the Ark of the Covenant and put in a warehouse. It’s got all this great stuff inside, but nobody knows what it is or what they could do with it. The licensing offices holding onto this box are looking out for the university as opposed to looking out for the idea.

a16z: So intellectual property is a big factor. What else?

Vijay: Funding. The traditional funding model in academia is for professors to write grants, and based on your grant, you get funded or not. Note, I say “based on your grant” rather than necessarily based on quality of your work — because you can have people with great ideas and bad grants, or bad ideas with great grants.

What Stanford does, somewhat uniquely I think, is fund graduate students directly, through the Stanford Graduate Fellowship program and Dean’s funds in the Medical School. Not just in tech but in English and so on; the fellowships are everywhere, even in the medical school.

And the interesting thing about funding graduate students directly is you essentially make them “free” to faculty members — but then, they get to choose which faculty they work for. (The graduate students I get are often better than the post-doc candidates from other places that try to get jobs in my lab!). So you’ve got motivated people who can vote with their feet to work on what they think is exciting and important.

This is key because I often find that graduate students are better judges of what’s exciting and what’s going to be successful than faculty members are. If you think about it, grad students have much more vested in this because their whole career is riding on the decision, whereas when it comes to evaluating which grants to fund, another professor might read it in 10 minutes and jot some notes down. It’s a different way of selecting ideas.

This is a model of innovation where you essentially take really bright people and empower them to do whatever they want. The deal is we’ll give you all the resources you need — we just expect you to do something great.

a16z: But what happens when graduate students leave? Don’t you lose all that work?

Vijay: The interesting thing about cycling through — this is kind of a dark analogy — is it’s almost like how people live and die. With death comes change. My research can now change because new people came.

So we’re not really wedded to anything we did before because those people have long gone, and there’s nobody left in the lab to defend or not defend it. That’s intrinsic to and I think one of the strengths of academia. The weakness, obviously, is that new grad students need to be trained, but if you get strong people that’s not so much the issue.

In most corporate jobs, it would be a negative thing to have to rebuild your research team every few years. This is more like building a sports team.

a16z: Does Stanford have any formal mechanisms for building entrepreneurship into its programs?

Vijay: There’s this industry stereotype of academia that academics are sort of head-in-the-clouds in their ivory towers. And there’s something to be said for that because academics typically prize ideas over products.

But bringing in people from industry to talk about the realities of what it takes to get stuff done has been a huge educational boon for our students, who hear it here before they get on the job market. This also facilitates us making impact in the spaces we care about. So it’s not just people in academia talking about making products, but really blunt, dead-on advice from people in the industry.

The building I’m in is the Clark Center, which is the headquarters for the interdisciplinary Bio-X program: biological everything. Another example, in the medical school, is the SPARK program. All of the translational work that my group has done was really made possible by the fact that SPARK brings in people from the pharmaceutical industry to give us advice for how to design drugs. I think the Engineering School and Design School have similar types of ties with industry.

This matters because a lot of innovation is not scheduled. It’s who you run into and whose office you can walk to and see if they’re there.

a16z: What can other universities do to create innovation clusters in their regions?

Vijay: My friends at peer institutions ask me this all the time; how can we transform ourselves this way?

In the medical school decades ago, they convinced several pioneers to come join the Biochemistry department. That was years ago, and it really kicked things off for Stanford’s reputation. By nucleating a few key departments that were at the very top of their game, the neighboring departments got stronger (just like property values going up in neighborhoods) — to the point where there’s now an expectation all of them should be #1 in their respective fields.

This reflects very thoughtful long-term planning because you have to be very careful how to do it. I’ve spent all this time talking about avoiding top-down stuff, but when you do top-down stuff, you want to make sure that you’re doing it right; and I think, again, it’s about picking people rather than telling those people what to do.

a16z: Does all this industry focus create a tension between applied and basic science?

Vijay: I do have colleagues who lament this a little bit because at other places, basic science is more cherished. It’s kind of like a parent’s love; you love all your children, but if you put more attention into one, the other one will get a little jealous. But it’s hard for a parent to put attention equally everywhere.

Stanford emphasizes impact — which often feels like it has chosen to be more applied — and which as mentioned before is more obvious for people to see and recognize opportunities. Perhaps the greatest issue with funding basic research is that its impact is typically measured decades later but can be truly transformative.

Without the fundamental basic research, we would not have any of this now. I’ve heard Marc [Andreessen] mention Bill Janeway’s thesis that the fields in which venture capital in the Valley has had an impact are mature ones that have long had funding from the U.S. federal government.

With all that basic research laid out, it’s very quick for people to nail things; there’s this huge stack of stuff they can just take for granted and build from. When new companies come online they don’t have to think about compilers or operating systems.

a16z: So let’s talk about MOOCs. I know you and [a16z General Partner] Balaji Srinivasan are teaching one that’s coming online later this year. What’s your view on these massively open online courses?

Vijay: If you’re lucky enough to have one of the world leaders on X topic teaching you something, and you could be one-on-one with them and get answers to your questions directly, that’s the ideal. The only problem is that’s extremely expensive.

Last time I looked it up, only a fraction — something like 5% — of U.S. students go to one of the top 50 universities. That means 95% of students don’t necessarily have access to the best lecturers, the types of things that a Stanford or Harvard student would maybe have. For those students, a MOOC, especially taught by someone who is both a leader in the field and has really good pedagogy, would be irreplaceable. An asset they wouldn’t be able to access any other way.

This is where the scalability of MOOCs also becomes interesting. At Stanford, I teach 200-300 students, which is considered to be large. But if you multiply that by the 95%, that’s an impact Stanford alone could never touch.

And this is in the United States alone. When you multiply that further by people not in the United States, that’s where there could be some of the greatest impact.

a16z: That’s one of the things people always talk about with MOOCs – access for students outside of the United States. But that’s so abstract. What does this mean, really?

Vijay: So 15 years ago, I went with my dad to India. And we went to the village where he grew up and where a lot of my distant cousins still are. It’s a village where there are thatch huts and dirt floors and only one building has running water, and it’s where I would be now if my father hadn’t gone to college. I walked into one of the huts, and there’s a little teenage boy going over a textbook; it was Halliday & Resnick, which is the same physics textbook that Princeton used when I was there. And in that moment it became evident just how easily we could have switched places.

The impact of MOOCs is in reaching this “dark talent”. People talk about “dark fiber” or “dark energy”, but there is a huge market of “dark talent” [to use a term Balaji coined] in places like rural India, China, and all over the world. A billion programmers. Being able to reach that dark talent, and giving them tools to create, could be hugely transformative.

That’s one of the reasons Balaji is so interested in bitcoin for the MOOC. According to him, bitcoin + computer + MOOC allows people to really do amazing things, with minimal resources, and make a real business. Or even just make an iPhone app. We mock app innovation, but you could make $100K a year on a good app, and for someone living in a hut in India, that’s dramatically life-changing. It also uplifts those around them, whether through money or education or quality of life or helping others.

a16z: By ‘computer’ [+ bitcoin + MOOC], do you mean mobile phones?

Vijay: I think it would be hard to program on a smartphone (at least right now) — you probably need a keyboard or something like that. But you don’t need much.

I mentioned how at Stanford the deal is that we will give you all the resources you need, and we just expect you to do something great? In my mind, that’s the deal MOOCs make with students. Even if you aren’t getting all the resources you need from a single source, all the resources are out there.

That’s what’s really unique about this time in I.T. right now: Because of the cloud, you don’t have to construct this huge backend. You don’t need to buy any infrastructure. The ability to jumpstart from nothing then focuses more on the quality of your ideas and your ability to do good work.

I can’t think of a time like this before, ever. In the industrial revolution, nobody got to build their own machines out of nothing…

a16z: Is there anything that worries you about MOOCs and universities in the context of the future of education?

Vijay: I worry about what the future of the university is going to be when MOOCs are driven by students who want their education to further their career only (versus more general intellectual development). I took a lot of philosophy in college because I found it interesting, and it affects the way I think about things now.

So my fear is that 20 years from now we will be talking with kids that just took a bunch of MOOCs and will have no cultural context for anything. Just like with basic research, where outcomes are unclear for many years, the benefits of the arts and humanities may never be transparent or immediate.

While I’m worried about this loss, it kind of feels inevitable in the current scheme of how the world works. It’s like with newspapers: You need the moneymaking parts to subsidize the other parts. In a sense, universities do that too. As university dynamics get more driven by the way of the internet, it may become like iTunes: You don’t buy the whole album, you buy a song; you don’t buy the whole university experience, you buy a couple of courses.

But this cherry-picking of things could also drive things in an unexpected way; people may pick the things that they really, really need, and then maybe a little bit of fun or surprising stuff too. We don’t know. All I know is that we can’t have the old world back.

a16z: If MOOCs allow education a la carte, why have a Stanford when all its courses can be online?

Vijay: The one thing that can’t be replicated is human interaction. I think traditional universities have an advantage in terms of the social and networking components of education.

But there are efforts to build in more social components to MOOCs. Like if there are 100,000 people in the class, odds are there are at least a few people living near you (and if you’re in a major city, there are likely a lot of people living near you). Is there some way to meet up? I think all of that can be aided by social media-like technology.

The difference, I think, with a place like Stanford is that you are surrounded by a concentration of people who are at the top of their game, and that pushes you to do things in certain ways. I think that’s something top universities have that will be harder to do in a MOOC, but you never know. Over time, if there’s some way for social tools to get the best people to interact with each other, and for people to know who they are, that could work too.

One way that this is already starting to happen — at least in coding and programming — is that you can quickly tell who’s good or bad by some of their public work. Like by looking at their commits on GitHub and what impact they’re having in various projects. It’s not just someone writing you a letter of recommendation saying you’re good; it’s all right there.

Eventually we’re going to see MOOCs get close. I like the newspaper analogy applied to universities here: There will always be a New York Times and Wall Street Journal; I can’t imagine that the internet’s going to kill those. But they will have to change their models.

The world will be different than we can imagine now though… I have two daughters, ages 10 and 6. Six years from now, if the 10-year-old had a growing, profitable company and told me “Dad, I don’t want to go to college” — I think she might be able to convince me. She could learn on the fly from a MOOC rather than sort of “front-loading” all that knowledge; she’d be learning how to learn and then learning more later. There could be a very different model.