Silicon Valley's 'smartest guy' on deep learning and sustainability
Steve Jurvetson has been referred to as “the smartest guy in the room,” “the smartest person in Silicon Valley” and a “brainiac,” among other laudatory monikers attesting to his prodigious intellect.
Hyperbole, perhaps, but not by much. The Internet is chock full of videos of lectures by and interviews with the venture capitalist, a partner at Draper Fisher Jurvetson. They span such topics as rockets and space, Moore’s Law, machine learning, synthetic biology, technological innovation, the rich-poor gap and “the democratization of matter.”
That begins to reflect the breadth of Jurvetson’s interests, and also his investments. Over the years, they have included companies that became transformational, from Hotmail (the Web as a platform) to Tesla (automaker as energy company). Today, they include pioneering companies in space, biotech, optical communications, genomics, deep learning and brain health.
At VERGE 2015, Jurvetson will sit down with New York Times senior writer and author John Markoff, who covers technology from Silicon Valley, to discuss how deep learning — essentially, a branch of machine learning that paves the way for artificial intelligence — can be applied to addressing the world’s biggest challenges.
In the run-up to VERGE, I talked with Jurvetson about how he connects deep learning to sustainability. The conversation has been edited for clarity and length.
Joel Makower: How does deep learning begin to address the climate challenge? Where do you see the opportunities?
Steve Jurvetson: Generally, deep learning lets you build a learning machine. You could call it machine learning or neural networks, there’s a whole gamut of different technologies. The whole approach to building things is very different from traditional engineering.
What it allows us to do is, in a sense, evolve or grow solutions to problems. And this approach is very powerful and can be used in many ways. Right now, Google and others use it for speech recognition and vision systems. But it's still a pretty domain-independent thing.
So how could it apply to climate change? You have complex data sets, such as from measuring remote-sensing data from satellites and ground sensors, or from the Internet of Things and the sensors that are all over the planet in various devices. Think of all the mobile phones that have temperature sensors, and cars that have sensors. We’re not really using that data for anything, but we could. Putting all the data on a screen wouldn't necessarily give you insights. But you could imagine some kind of a learning harness being applied to try to figure out patterns in those data sets.
Makower: Deep learning is partly about adapting. In a world with a changing climate, there's going to be a lot of things that will need to adapt — us, of course, but also our cities, buildings, food systems and energy grids. Can deep learning help with that?
Jurvetson: Here's the problem. Much of our infrastructure is built in a different engineering modality. It does exactly what it's supposed to do. And if it doesn't, it's a bug. I worry about the lack of robustness in these engineered artifacts. So if you think of the grid, if you think about a lot of public infrastructure, it's really brittle to outlier, black-swan-like events it's just never been tested against.
Where you're co-evolving with the technology, where you're iterating and growing these kinds of solutions in what feels like organic ways, that's ultimately less prone to catastrophic collapse. There'd be less of a monoculture. But it's also not necessarily easily integrated with the traditional engineering approach.
Makower: Can we retrofit our existing infrastructure to be less brittle?
Jurvetson: What we may be able to do is take our existing plants and equipment and think of an overlay network that almost runs in parallel, that helps with outlier cases. And maybe one day we can swap it out. I could imagine some transitions like that.
If you look at retrofitting an existing city, it can be kind of daunting, right? The amount you can do is very limited compared to, say, a new city in China, where you might say, “Okay, let's not have any parking lots and build everything around on-demand mobility.” So, robotic Uber vehicles in the future, but human-driven fleets of cars in the present, and no car ownership. Imagine designing a city around that. It's a very, very different city than the ones we have today. And so much more efficient.
You can see why that's not so easy if you start with existing cities, like Los Angeles, which is 40 percent parking lots. Still, there lies an opportunity for recovered real estate value.
When it comes to large systems spanning cities or nations, like the smart grid, it's companies like EnerNOC doing demand response, that are interesting to me. But frankly as an overlay; they don't really touch the grid directly. What they offer is a service to utilities that's exogenous to the grid. It doesn't have any touchpoints in the grid as we know it. What it does is touches the endpoints, and can shed demand or generate supply at those nodes.
There is something to be said for innovations that aren't trying to change a juggernaut, but sort of operate in parallel with that juggernaut. To improve it, if it can. Or just eclipse it altogether with something better on the side.
For example, distributed home generation instead of fixing the utility. It's so much easier to see how a SolarCity could become a huge distributed utility much more quickly than any existing utility could fix itself. It's just a night-and-day difference in timing.
Makower: So when you look at companies like SolarCity, and you look at sustainability as an overlay to venture capital, do you see opportunities in how technology can address some of the big sustainability challenges — food, water, housing, climate?
Jurvetson: Oh, yeah. And that's a great question. Investors who are just looking to make a quick buck and are arbitrage-seeking opportunists will get dismayed when a sector falls out of favor and with the whims of popularity. So, that's going to be a tough place to be, a place of great anxiety, and a lot of flip-flop behavior on what's hot and what's not.
Now, what if you focus on what are the biggest challenges of the planet that need fixing? In the past you may have just wrung your hands and thought it was impossible to fix. And all around, you can see examples — Tesla and SpaceX and SolarCity, just to pick three associated with a single entrepreneur. You can see how change that exceeds most governments' capacity to produce change is all around us.
Makower: So you don’t buy into the notion that the cleantech hype was misplaced?
Jurvetson: True, there has been a big funding cycle and a big failure cycle in the sector. But my enthusiasm is no different. I look for those true entrepreneurs — true in a sense of driven by that massive passion to make a difference and living their primary passion, meaning the profits are not Job One; they're a byproduct of Job One.
By solving the big hairy problems, there will be profits to be had. It's not just a philanthropy project. It's meant to be a business.
Tesla is a perfect example of this, by the way. To make all vehicles electric, an entire transition of an industry, you can't have a bad business that no one wants to copy, right? If you want to persuade other automotive companies to go electric, you can't do so with a worse business model than they have. You can't suck and lead.
So those kinds of companies are exciting. I would include SpaceX and Planet Labs from the satellite side, as well as Tesla and some others that are less well known. They’re the most successful companies in so many different ways — in terms of revenue growth, pace of success, loyalty of customers, loyalty of employees, loyalty of partners. It's just so many benefits, some of which are hard to quantify.
So with all that preamble, I’m as excited as ever. I will say that's not necessarily the case for all venture investors. I don't know what the mix is, but I'm hoping there's more and more folks who say what I'm saying. As opposed to the opposite end of the spectrum, which would be, "I'm looking for the best way to make a quick buck for my investors at any given moment." Our ecosystem is full of folks like that, too.
But that's not really what I'm all about, or what I want to see a career based upon. I don't think that's the foundation for long-term success. I think it's opportunistic at best.
Sometimes opportunism pays, but I'm not sure it pays over the course of decades. It may only pay over the course of a few years. I'm looking more at the decade-long game in terms of strategy. And I think this kind of focus area has that kind of staying power.