Chris Dixon on competing with Internet giants for budding AI and VR talent

VC Chris Dixon of Andreessen Horowitz thinks it’s a lot harder to predict financial cycles than it is to see a new computing platform coming down the pike. As he noted in a recent post, new cycles tend to begin every 10 to 15 years; assuming the 2007 introduction of the iPhone kicked off the last wave, we’re fast heading toward the Next New Thing.

Or things, technically, according to Dixon, who we caught up with yesterday. Among the trends that Dixon is watching closely, he says, are virtual reality, augmented reality, IoT, wearables, drones and cars. (Missing from the list: bitcoin, which has long held Dixon’s fascination but that he refers to as a “long-term project.”)

Not that it’ll be easy to make money off these newer technologies. In fact, Dixon suggests it could be ridiculously challenging, given how quickly Facebook, Google, and Amazon are bringing aboard related talent. Here’s more from yesterday’s interview, edited for length:

TC: People think of you as the person at Andreessen Horowitz who invests in weird stuff. 

CD: We obviously invest in a wide range of things. My own area of interest has been in drones and VR and AI and maybe more speculative categories. Some startups, the question is more about ‘Will this startup win versus other solutions,’ where, in speculative categories, the question is whether it’s going to work at all.

You can kind of jokingly call it weird, but if you look at where Amazon, Facebook, and Google are investing — I think Google’s VR team is significantly bigger than Facebook’s; Microsoft has 1,500 people working on HoloLens; and from what I can tell from its hiring and acquisitions, Apple is [investing a lot of money] — probably the biggest area [of interest and investment] is AI. Large tech companies are investing very heavily in this stuff [whereas] there’s much less investment by VCs.

TC: Because VCs don’t understand the tech or else the opportunity?

CD: No, because it’s hard to figure out where the start-up opportunities are and because [some of this tech] requires so much money. With virtual reality, you have to build a complex platform then line up content partners. Or self-driving cars — I’d assume that Google has spent many billions of dollars on it already, including mapping. The venture world isn’t set up for that. It works around the model of seed rounds, A rounds, $20 million B rounds, not for massive projects. It’s kind of a puzzle if you’re in VC how to make those investments.

TC: You’ve remarked  before how quickly teams are getting snapped up, which must compound the issue.

Wit.ai [a Y Combinator startup that built voice-activated interfaces] that Facebook bought and now powers its Messenger platform was only [in our portfolio] for a few months when Facebook bought it.  It sounds paradoxical, but our model depends on companies staying independent for a period of time, and because large companies have been so aggressive, it’s harder for us.

When it comes to machine learning, you’re competing with offers from Facebook and Google and Amazon and [their offers] are considerably higher in terms of cash compensation. They pay a lot for people with that expertise, and startups will never [be able to match it]. So you have to really convince people that what you’re building is important.

TC: VCs can’t wait out this next computing cycle obviously. So how do they nurture lower-capital models?

CD: With self-driving cars, for example, there are numerous approaches that startups have taken to overcome that capital advantage. Some startups are starting on private campuses — colleges, corporate campuses, military bases — where you can go at a slower speed so it’s less technically challenging because you have enough time to stop and the regulations are controlled by the campuses.

Some companies, like [our new portfolio company] Comma.ai and the public Israeli company Mobileye are starting with driver assist, which is semi autonomous. If you can get your advanced cruise control on the highway and get your system out there, most of the products will get better because they’re pulling in sensor and optical [and other] data and their systems are getting better.

Others are just raising a bunch of money. I think Cruise [the self-driving tech company acquired last month by General Motors] was on the path to probably go and raise more money.

TC: Your new investment in Dispatch, which makes last-mile delivery robots, is interesting. How big an opportunity is that, do you think?

CD: It’s early and we don’t exactly know, but if you talk to people who make deliveries to college campuses, they get lost looking for dorm rooms. So I think in the near term, we’ll see partial solutions where the delivery person might drop off mail or food at a central depot where [he or she] doesn’t have to worry about parking, and the little bots will do the last bit of it. If you did that on college campuses, corporate campuses and military bases alone, it’s a big opportunity.

TC: What about drone delivery? 

CD: I think [that opportunity mostly centers on] photos and capturing imagery for inspections and other things. I think drone delivery would be really expensive. The batteries are the biggest expense; they can be used [just] 200 times. The math is probably $5 per delivery. It probably makes sense if you really want something right now, like medical supplies. So there are a bunch of interesting use cases. But I think the ultimate solution will be a combination of things: drones for emergency medical supplies, automated trucks for longer deliveries, ground drones for local deliveries. I think like today, it’ll be a patchwork.

TC: How many boards are you on and how many companies do you see a week?

CD: I’m on 10 boards and [former TechCrunch writer] Kyle Russell, who works closely with me — we probably see 20 companies a week?

TC: Given that you’re looking at a lot of nascent tech companies in VR and AI that don’t necessarily know what they’ll be when they grow up, how do you avoid conflicts?

CD: The best we can do is take what they’re working on and try to project it out and make a prediction. But in the cases where [companies become] really successful, it’s hard to know. No one would have predicted 20 years ago that Google would be building cars. We just try to be respectful of the entrepreneurs. We want them to feel like we’re 100 percent supporting them. But it’s hard because companies change directions a lot.

TC: You were a very visible proponent of bitcoin in recent years. Are you still?

CD: It’s a little frustrating. [There are] disagreements in the bitcoin developer community about the path forward. I  think using cryptography to develop a new currency is a long term project. A lot of these things — VR, bitcoin, autonomous cars — they’re hard and they take a while. I’m on the board of [bitcoin wallet provider] Coinbase and they’re doing well. They planned accordingly; they expected this would take a long time.