Looking back at 40 years, Khosla says that he has not been able to find any major innovations that have come from large companies
and they are mainly being created by the startup firms.
For instance, he says automobile giants General Motors and Volkswagen couldn’t design an electric car and large plane makers Boeing and Airbus were not able to create a venture like Elon Musk’s space transportation services company SpaceX. Khosla also sees the area of AI-led drug discovery being dominated by startups compared to large pharma companies.
“I can tell you that in most cases, the innovation came from somebody who had never worked in that area,” says Khosla.
“Like Elon Musk (co-founder and CEO of Tesla) hadn't worked in cars and the automaker's sort of laughed at the Silicon Valley startup with no experience in automotive. He made lots of mistakes, but he fixed them quickly. And he figured out a better way they would have figured out by applying conventional wisdom.”
After graduating with a bachelor’s degree in electrical engineering from IIT-Delhi, Khosla failed to start a soymilk company to service the many people in India who did not have refrigerators. Instead, he came to the US to further his academic studies.
Khosla, who co-founded Sun Microsystems in 1982 and helped incubate the idea and business plan for Juniper Networks, has invested in companies
focusing on areas ranging from solar, biofuel to energy storage and even wearable device makers to increase energy and lower stress. One such venture is Impossible Foods Inc., a company that develops plant-based substitutes for meat and dairy products.
“It makes vegetarian burger...It tastes the same as meat. Most people cannot distinguish between this and (the real meat) burger, except that it doesn't come from an animal,” says Khosla.
“Plant proteins are the best way to save the planet. It's healthier than meat proteins for humans (as they) come with cholesterol and other negative things.”
Khosla also sees AI replacing doctors such as the role of radiologists. He says looking at the population of India and number of doctors here, one won’t be able to train enough doctors in the next 30 years to achieve the doctor-patient ratio like present in the West. Also, the doctors have limitations, for instance, keeping up with the last 5,000 articles published on prostate cancer or remembering a certain type of treatment for every new mutation that regularly keeps emerging.
“These are not things for human beings to do...A normal blood test should result in 30,000 data points. We don't do 30,000 or even 3000 or 300 data points because humans can’t look at it,” says Khosla.
“So my point is really whether it is food, buildings, construction, transportation, education and health care, they are all open to radical innovation with technology. That is optimistic me and that is why I have too many problems for the remaining part of my life left.”