Back in April, I talked to Sun Microsystems co-founder and early OpenAI investor Vinod Khosla, asked some important questions, and got some interesting answers.
A lot of this was related to what we’re going to see in terms of general artificial intelligence and specialized applications in the years to come. ChatGPT has taken the world by storm – we seem to be at an inflection point. So what does an early investor have to say about the future of AI?
Well, starting off the whole thing, Khosla talks about getting in on the ground floor. Going over a timeline, he mentions starting Sun in 1982, analyzing the emergence of the Internet in 1996, and starting Juniper later based on a belief in an exponential rise of technology within the next few years.
“Nobody else believed it,” he says. “It was uncertain that’s the time to jump in.”
Take note, also, of some of Khosla’s predictions for the next quarter-century. He predicts that 80% of all jobs will be done by AI! He predicts FDA approval of an “AI-only family doctor” in five years, and talks about a project where a company is trying to come up with a top 10 music hit that’s created by AI without human participation.
“The process is pretty interesting,” he says.
Young entrepreneurs, pay attention: in terms of advice for a YC startup founder or anyone with skin in the game, Khosla suggests it’s better to look to the future, since new things get obsolete quickly.
“The most important thing a founder does is figure out whose advice to take,” he says.
Among other top Vinod Khosla tips: assemble great teams – use long-term analysis – and focus not just on one metric, but on a longer time horizon.
Answering our question about future techniques, he suggests that alternatives to large language models might come out of important research, including research right here at MIT. We’re blushing….
In terms of envisioning the future of AI, Khosla is optimistic.
We’ll be in the era of great abundance,” he says, “of GDP growth – being able to be more creative – music, art, storytelling, all to be done by AI – all the way to oncology, and medicine, and robotics – almost anything you can imagine.”
At the same time, Khosla responds to my next question about inequality this way, agreeing that the outcomes of AI will increase inequality:
“The solutions will be social,” he said, promoting the idea of a universal basic income. “Society will have to address income disparity.”
He also addressed misconceptions about what the future will look like, with a prediction that should be sobering:
“I think people are expecting a wave. What will hit us is a tsunami.”
He makes the important distinction between the last big disruption of the labor in the U.S., a move away from agricultural employment, and the new AI revolution.
“We had multiple generations to adjust,” he said of people leaving the farm to work in other professions over the last century. “(This time) we will have a hard time adjusting.”
When I ask Khosla about five-year predictions, what he says is that smart people won’t be predicting things at all. Touche. When people do make predictions, he adds, they’re more likely to be wrong about AI not being able to do a certain thing. (In other words, AI will be able to do a LOT of things!)
I also wanted to ask Khosla about climate science – and I do, here. He calls it a difficult problem, and suggests that he doesn’t see AI being transformative in this, that papers on AI and climate science are “playing on the edges.”
However, he suggests a small number of people can change climate outcomes, and he cites a blog that he wrote in July 2021 (maybe this one?)
Some of the goals he sees for AI involve discovering new materials like a room temperature superconductor and new lithium ion battery systems.
I’ll leave you to watch the end of the interview where we talk about China versus the United States, but the bottom line that shows through in this exchange is that Khosla is suggesting that America needs to get serious about workforce transformation and social changes, in order to stay at the top of the global economy, and to remain an important geopolitical superpower in this new era of abundance and disruption. In any case, you can get a lot of insights from this talk, like our event audience did – people were talking about this at the event, and after.