Vinod Khosla is now more popular than ever. The Sun Microsystems co-founder became a outstanding investor — first at Kleiner Perkins and over the past 20 years at his enterprise capital firm Khosla ventures — has at all times been sought after by founders because of his sound advice and his company’s history, including bets on Stripe, Square, Affirm and DoorDash. But we’re risking $50 million OpenAI in 2019 – when it was unclear whether the team would achieve success on this scale – they put Khosla Ventures and Khosla himself in the highlight.
He’s having a great time. I met with Khosla in Toronto last week Collision conference and before our appearance on stage, he told me that he has been making public appearances several times a week these days – on stage, in podcasts, or in TV interviews. Asked if he was exhausted from his schedule – he flew to Toronto just hours before our meeting, for example – he shrugged.
There are actually things he prefers to speak about, and the art of constructing deals is not one of them. “Honestly, the investor side is much less interesting to me,” he said when I asked him about something I recently heard, which is that he hasn’t taken a dollar in management fees since founding Khosla Ventures, though it’s currently she has $18 billion in assets under management. (He confirmed this, but said it only applied to himself, not a corporate-wide policy.)
He’s much more passionate about the startup opportunities he sees in a landscape that is changing every day with advances in artificial intelligence, so we talked about some areas of that white space. We also talked about what worries him most about the effects of artificial intelligence; FTC Chair Lina Khan; and why, in his opinion, “Europeans have thrown themselves out of leadership in any field of technology.”
First, we talked about Apple’s shiny latest deal with OpenAI, which allows Apple to integrate ChatGPT with Siri and its generative AI tools. Apple may strike similar deals with other AI models, including Meta, but obviously as an OpenAI investor, Khosla is bullish on this deal, which is the only one Apple has announced publicly yet.
Khosla called it a “validation” of OpenAI; announcing its pact with OpenAI at a celebrity developer conference, Apple also “expressed, I think, confidence about [OpenAI CEO] Alone [Altman] to steer [developments in AI] in the next five or 10 years,” Khosla said. “When a company like Apple commits to technology, it usually doesn’t change it the next year.”
But we wondered whether this was good or bad news for Khosla? As we noted at TechCrunch, many startups are more likely to go out of business resulting from some of Apple’s latest features, and it seemed likely that Khosla’s portfolio firms weren’t entirely immune. I used to be particularly interested in Rabbit, whose artificial intelligence-powered device is intended to be a form of AI assistant for users and is supported by Khosla Ventures.
Asked whether Apple could make the device obsolete, Khosla suggested it is more flexible than people imagine and may very well be used by businesses reminiscent of hospitals, including emergency rooms. He put it in a growing range of things that may “watch what you do, see what you do and respond automatically.”
In fact, Khosla suggested that his team actively avoided anything that would turn out to be “roadkill” as large language models like OpenAI proceed to advance. And he identified at least one company that is not in his portfolio: Grammaticallywriting assistant startup that was valued at $13 billion by its backers not too way back.
“If you are dealing with, say, grammar, it’s really a minor challenge in comparison with today’s model and Grammarly cannot sustain; this could never have been an app. It shows the need for this capability, but it can be a part of Word or Google Docs. It’s quite obvious. When we seek advice from YC firms or other firms,” Khosla continued, “I can normally say, ‘Half of those firms shall be obsolete by the time the YC batch runs out.’
Khosla sees a lot of opportunity in industries where expertise shall be almost free, although it isn’t clear to me how these firms will sustainably make money (even after asking him). Think about tutoring or even oncology.
Said Khosla: “Open AI or Google won’t build a chip designer [to have on your smartphone]. OpenAI and Google won’t build a civil engineer. They are not going to create a primary care physician or a mental health therapist,” he said. “So there are so many areas [founders to mine]. But they should look at where the models are going next yr and five years from now and say, “We want to realize this potential.”
We also talked about regulations. I noticed that Khosla had previously said that closed large language models like OpenAI needs to be protected, though there needs to be a regulatory framework around them. I wondered if this meant Khosla would endlessly eschew other “open source” AI.
Not at all, he said, noting that he’s a “huge fan” of open source. He said Sun was one of the first firms to “leap into open source” and open source its file system. He also noted that Khosla Ventures was the earliest investor in GitLab, whose software invites people to work on code together.
However, he suggested that open source in the context of huge language models is a completely different matter. “The biggest risk we face with AI is China” and “the powerful Chinese AI” that competes with the “liberal values” of the United States, he said, adding that “we need to make sure China stays behind us.” . Otherwise, he warned, China will provide the remainder of the world with “free doctors and free oncologists” and in the process “export both the economic power of artificial intelligence and its political philosophy. “
On stage, I discussed to Khosla my recent meeting with FTC Chair Lina Khan, who does not imagine in the national champions model as a reason to coddle firms like Google and OpenAI as they proceed to develop artificial intelligence.
Khan consistently hears from executives and investors who say that government intervention will lead the U.S. down a dangerous path. However, during my conversation, she argued that the United States has time and time again chosen the “competitive path,” which “has ultimately driven and catalysed many of those breakthrough innovations and much of the extraordinary growth that our country has enjoyed, which has allowed us to sustain advantage in the international arena.
If you look at other countries that have as an alternative chosen this model of national champions,” Khan added at the time, “they have been left behind.”
However, I had barely mentioned Khan when Khosla became dismissive, calling her an “irrational human being” and accusing her of not understanding the business.
“She shouldn’t be in this role,” Khosla said. “In every country and economic system, it is good to have antitrust laws. But antitrust [that’s] Over-enforcement or over-execution is bad economic policy. The one thing the United States has over its European rivals is a much more rational business environment. This is why Europeans are no longer the leader in any field of technology; they have simply regulated themselves beyond artificial intelligence, all social media, and all internet startups.”
Of course, if some antitrust enforcement is good, but too much is not good, the query is where to attract the line. At this point, before we parted, I discussed the “abundance” that Altman predicts shall be created by artificial intelligence. At one of TechCrunch’s StrictlyVC events last yr, Altman said the “good argument” for artificial intelligence is “so incredibly good that you sound like a really crazy person when you start talking about it.”
Khosla said he believes the same, but I have long wondered how society will enjoy all these advantages if regulators don’t get more involved in the development of those firms. After all, I told Khosla on stage, we have already seen massive aggregation of wealth and power tied to an increasingly smaller group of firms and people. When is enough?
In this case, Khosla said the issue concerned him greatly. “I think in 25 years, when I’ll hopefully still be working. . . the need for work will mostly disappear.” Still, while AI should deliver “great abundance, great GDP growth, great productivity – all the things that economists measure,” he said, he worries “more than anything else” about “growing income disparities.” As we [ensure the] fair distribution of the advantages of artificial intelligence?
He has a feeling where the breaking point could be. “If [U.S] GDP growth will increase from the current 2% – currently lower than 1% in Europe – to 4%, 5%, 6%; we’ll have enough abundance to share the wealth and advantages.”
Whether and how this happens are, after all, even greater questions, and for all his brilliance, Khosla, a self-described techno optimist, didn’t have the answer.
Instead, after his GDP remark, he thanked the audience for their time, stood up, then walked off the stage and towards at least a dozen young founders who had gathered backstage, each of them hoping to get him to hearken to him. so long as they might.