The key to making effective flying cars is eliminating the need for a human pilot, according to Nathan Myhrvold, the co-founder and chief executive of Intellectual Ventures.

“The problem with flying cars is that we’re not all good enough to fly them,” he said at a session at Bloomberg’s 2016 Technology Conference in San Francisco. “The leading cause of plane crashes is pilot error.”

Flying cars have been a longtime fantasy of the nerd set. The topic has gathered steam in the last week, since Ashlee Vance and Brad Stone reported in Bloomberg Businessweek that Google co-founder Larry Page has been secretly investing in flying car startups. The vehicles these companies are making are based on the idea that they will be able to fly themselves.

Making a self-flying car is easier than making a self-driving car, according to Myhrvold. “Even a little quadcopter does great,” he said. Earlier this year, DJI, one of the leading manufacturers of drones for hobbyists, introduced a quadcopter that uses two optical sensors and an on-board computer to avoid obstacles while flying at speeds up to 22 miles per hour. While many planes also have some auto-pilot modes, Myhrvold said that the conservatism of the airline industry and the Federal Aviation Administration has held the technology back.

Myhrvold, who was previously chief technology officer at Microsoft and held other positions at the company for more than a decade, also spoke about the process of invention practices at Intellectual Ventures. During brainstorming sessions, he, Bill Gates and others get together to try and come up with inventions on broad topics like “climate” or “energy.” As the days wear on, the ideas get increasingly loopy, he said. But that’s the point. No one at the sessions has a fully realized idea. Instead, they each come with a piece of the overall puzzle, and the main objective is to think about things in new ways.

Myhrvold predicts that the breakthroughs in the next five years that will have the most profound long-term impacts will come in the field of synthetic biology, a discipline that combines biology, computer programming and genetic engineering. The real-world impacts of these discoveries are unlikely to come until farther in the future, he added.

After making his fortune at Microsoft, Myhrvold has become something of a notorious figure in tech circles. He founded Intellectual Ventures in 2000 with the idea of amassing patents and asking companies using the technologies to pay to license them. Those that refused would be taken to court. This turned Myhrvold into a symbol of one of Silicon Valley’s most hated archetypes: the patent troll.

Even those who aren’t inclined to compare Intellectual Ventures to the mafia or a parasite have begun to question whether its original business model even works. It has become decidedly more difficult to make money suing people for violating patents, largely due to policy changes pushed by the kinds of companies Intellectual Ventures was asking for money from. The company has laid off a significant portion of its workforce in recent years and increasingly emphasized bringing its own inventions to market.

Myhrvold has also pursued a range of eclectic personal projects. He recently challenged NASA’s methods for tracking asteroids, setting off a minor controversy. In the last few years he has published both a scientific paper arguing that leading paleontologists made serious errors in their research about how dinosaurs grow and a five-volume, 2,400 page cookbook on the “art and science of cooking.” The sequel, which will focus on making bread, is due out in March.

“I do a crazy set of stuff,” he said.