Can computers and the internet transform the education industry? For decades, teachers, principals, and technologists have looked for ways to use computers to enhance the educational experience, with little to show for it.
But prominent venture capitalist Marc Andreessen argues that that’s about to change. Andreessen made his fortune as co-founder of Netscape two decades ago, and more recently his firm has invested in successful companies like Facebook, Twitter, Airbnb, Slack, and Lyft. And he’s an investor in several education-related ventures — ventures that he hopes will finally bring real change to the education sector.
In an interview conducted in late September, I pressed him on whether software can really transform education — or if there’s something about taking care of children that’s makes it inherently difficult to automate.
This is the second part of our conversation — you can read part one on the future of artificial intelligence here. The transcript has been edited for length and clarity.
Timothy B. Lee
You’ve argued that software has the potential to “eat” almost any industry. But do you think there are industries where this doesn’t apply? One that comes to mind is child care. Even if you could create some kind of child care robot or iPad software or something, it seems like most people want their kid to be interacting with a human being, not a computer.
I would say we've been through phases of what it means to do child care. One hundred years ago child care meant putting the kid to work. Fifty years ago that changed to putting the kid in front of TV. The great concern of that era was going to be that TV was going to rot all of our brains and we were all going to be completely useless.
There's this idealized view of child care, which is like every kid should be in his own cornucopia wonderland of hands-on tutoring and supervision, and then there's real life — busy mom, busy dad, TV set.
So I would argue, for example, child care: Would you rather have your kid be in front a TV set or an iPad? I would argue, you’d much rather have your kid in front of an iPad if you’re faced with that as a practical choice. The interactivity is far superior to a TV.
There’s no question that for day care, nannies, pre-school, and so forth, they’re going to be people powered and that they should be people-powered. I will say though there will be the opportunity through software to upgrade what it means to provide that in-person care.
There are two obvious things that can happen. One is a much better model for labor matching. I'm on the board of this company Honor that's trying to fundamentally reinvent in-home senior care. Among other things, we want to up-level the job and actually pay people more in the process. Somebody could do the exact same thing on the child care side. I think that’s just a matter of time.
The other is that I think there are going to be more and more hybrid child care and education, human and software combinations. An example we're involved with is called AltSchool, which is literally building new technologically powered schools, which has a new approach to having teachers and software working together. It’s a completely new thought-through way to structure a classroom and structure a curriculum and provide a very fluid, adaptable, child-centric and also parent-centric teaching environment.
There are human teachers for sure, but there’s also very rich and sophisticated software that actually touches the kids, as well as software the parents and teachers are able to use to help the kid get the better experience. The classroom of the future is going to be much more software powered than it is today.
Timothy B. Lee
There have been several iterations of people saying technology is going to change education. You’ve had computers in the classroom and smart white boards and flipped classrooms and Massive Open Online Courses (MOOCs). It seems like none of these things has moved the needle a whole lot. A typical kid is still in a classroom with about 30 other kids and a teacher. What do you think it will take for technology to really have a big impact?
Primary education in the US is a monopoly. It's a public sector monopoly with very little competition. Even the charter schools end up under sustained attack for violating the monopoly. You see this most recently in New York with De Blasio trying to shut them down. A government-sponsored monopoly is not easy to move.
With higher education, at least there’s competition, but it’s a government-sponsored cartel. The structure of the cartel is accreditation and student loans, both of which are run by the government. So you have exactly the effects at the higher education level you'd expect to see from a cartel, which is the facade of competition but not really.
The higher education market in the US, the traditional one, can’t adapt to increased demand. Any increase in money is just wealth transfers to the owners of these institutions. Many of which are the government. And therefore you get the student loan crisis, massive inequality, and the fact that there are only so many kids who can get a top-quality college education, and everyone else is confined to lower-quality schools.
New technologies tend to vaporize on impact with those institutions. The last thing a unionized public school wants to do is to fundamentally change how they operate. Of course they don't want to adopt new technology. It’s antithetical to the philosophy.
Timothy B. Lee
You’re an investor in Udacity, which was started in 2011 based on the idea that superstar teachers would be able to teach tens of thousands of people simultaneously. It doesn’t seem like that model has caught in the way optimists hoped. Am I wrong?
Udacity is growing very fast and doing very well, and has locked in on this concept of nanodegrees, which are basically educational partnerships between Udacity and employers. Google is one of their case study employers, but there’s AT&T and many others who are starting to do this now. The employers will sponsor these courses because the employers need more talented people who have these skills.
The big new course they have teaches how to design and code and build a self-driving car. Think about that. In 2005, a self-driving car was a DARPA grand challenge. In 2010 it was 1,500 people at Google doing this dark science thing that nobody understood from the outside. Now you can literally sign up on Udacity and take an online course to learn how to build a self-driving car and go work for Google or Uber or Tesla or any of these other companies.
I think it's pretty significant, number one that that's happening, in terms of how fast it’s happening. But it’s also significant that it’s happening on Udacity and not at Harvard, or for that matter your local community college.
Udacity is a software company. It has software-scale economics. As they refine their model, their ability to scale is present in the model. They're not there yet, but they are scaling quite quickly now.
This goes back to your general question about things like health care and education. We can’t revamp the entire system. Nobody can. But I think more and more, there are gaps in what the current system can accommodate compared to what people actually want. There are opportunities to build on the edges, around the sides, parallel systems. And at the very least introduce choice. In the best-case scenario, it becomes a real challenger to the status quo.
Timothy B. Lee
The basic theory of disruptive innovation is that you start at the low end of the market and move up. Is that what you expect to happen here? Are you imagining that something like Udacity’s nanodegrees could grow and improve and at some point you have kids who get into Harvard and decide to go to Udacity because it’s cheaper? Or do you think it’s going to be that people who can get into a top-tier four-year university are going to do that, and Udacity is an affordable option for everyone else?
There's this study of income correlated to degrees and which institution you went to. The data is very interesting. If you go to college in the US you get a college degree and you get a liberal arts degree, your salary is directly correlated to the institution you went to. If you went to Harvard, you get a high salary. If you went to your local community college, you get a lower salary. There's a straight line relationship by quality of educational institution.
If you get a STEM degree, anything in science, engineering math — also economics — the salaries are almost exactly consistent across quality of institution. It's basically a flat line.
Look, if you can get into Harvard, go to Harvard. This is one of the things I disagree with Clay Christensen. He talks about how Harvard Business School is going to get disrupted. I don't think that's the game. If you can go to Harvard Business School, you should go, because in addition to going to Harvard to get the education, you're also going to Harvard for the connectivity, the reputation, the alumni network, all the educational opportunities, all the employers. McKinsey and Goldman Sachs don't recruit at schools other than the top 10, so if you don’t go to the top 10 you don’t have access to the highest-paid opportunities.
If you can do that that's fine. But add up the total number of seats at Harvard and the Ivy League and the top 10 public universities in the US. How many 18-year-olds a year can they accept? It's a drop in the bucket.
So most kids today are underserved. Either they have to go to a lower-quality institution, or they don't get to go at all because the whole thing is too expensive.
Then think about the rest of the world. You're a sharp 16-year-old in a village in Indonesia or Kenya or Chile. You're every bit as smart as anybody entering the Stanford class of 2022. But unless your parent is a diplomat or a high government official or the head of a drug cartel or something, you're screwed. You’re never getting access to what we would consider a modern higher educational opportunity. And so the vast majority of 18-year-olds every year don't have what we could consider to be any decent options whatsoever.
I think that's the long game. You have so many kids around the world. Even if you had a new Andrew Carnegie or Leland Stanford who decided to go build physical campuses in all these countries, it's a numbers problem. You can't build enough campuses. You can’t train enough teachers. You can't educate the world with the existing system. The long-run opportunity is the real disruption opportunity.
The classic explanation of disruption is not about a better solution for someone who’s already well-served. It’s the underserved market. And it just so happens that most of the kids in the world are a giant underserved market right now.
By the way, it’s not just Udacity. Khan Academy is going to try to go after those kids. YouTube videos are a pretty big part of this. You're seeing a lot of people now zoning in and figuring out the scope of this opportunity. We don’t have it all figured out yet, but I think over the next 10 to 20 years, that whole thing is going to move very quickly.
It’s such a natural bias to think about the US market. But most American tech companies sell more outside the US than they do inside the US. And I think that percentage is going to shift more and more. More and more these are global questions, not US questions.