“I’m not a car person,” says Jim Adler, executive advisor to Toyota Research Institute (TRI). “I started my career as an electrical engineer, launching rockets for what’s now Lockheed-Martin.”

But shortly after he joined TRI in May 2016 and began to work on  helping Toyota leverage data obtained from cars to train their artificially intelligent systems, Adler was hooked. “I just loved the mission,” he says. “From that, we transitioned to figuring out what’s next for Toyota. That led us into venture investments and – here we are.”

In July 2017, Adler became the founding managing director of Toyota AI Ventures, a venture capital subsidiary of TRI that has an initial $100 million fund to focus on early stage investments in artificial intelligence, robotics, data, cloud technologies, and autonomous mobility. “What we’re doing is watching the experiments that are being conducted by startups in this marketplace and, quite frankly, learning from the failures and celebrating the successes,” says Adler. “That’s how we learn to be more resistant and adaptable to change.”

In the autonomous mobility space where Toyota AI Ventures has invested in several startups, including a LiDAR company named Blackmore and a self-driving shuttle company named May Mobility. The focus is on helping companies at different stages of their growth develop safe, efficient, effective, autonomous mobility technology, whether self-driving vehicles or microtransit systems. The challenge with autonomous vehicles, Adler says, is calibrating them for specific conditions and environments. An autonomous car, after all, “is just a really, really, big wheeled robot,” he says. “It shouldn’t drive too fast, it shouldn’t drive too slow, it shouldn’t be too aggressive, shouldn’t be too conservative. That driving behavior could be different in Riyadh or Amsterdam.”

Adler says the car of the future “has to be social” and take cues from everyone else on the road. So what is appropriate social behavior for an autonomous vehicle?

“You give a child a certain biased set of data,” he explains, “and, from that data, they learn a set of biased lessons. So it’s important that whatever we teach these machines, just like we teach our children, needs to be representative of the world, so they are calibrated to the world that they will operate in. And that’s a recipe for success.”

Adler might be a relative newcomer to the automotive industry, but he’s no stranger to the world of startups. After stints at Intelius, Metanautix, and VoteHere, a pioneer in cryptographic voting systems for public elections where he raised over $25 million in venture capital, Adler has developed an appreciation for the grit and chutzpah it takes to persevere with a vision and bring your product to the marketplace.

“Every startup ends up getting punched in the face,” he says, “and how they deal with that – the tenacity they show and the ingenuity and creativity they bring to bear to solve problems – really takes a rare depth of character.”

He turns to a famous example to illustrate what it takes.

“Look at Thomas Edison. After trying thousands of filaments for light bulbs, someone said, ‘You have now 1,000 failures,’ and Edison said, ‘No, I found 1,000 things that don’t work – I’ve learned 1,000 lessons.’”

To view the full video, click here.