Sundar Pichai talks about the challenges of innovation at a big company and what excites him this year

Alphabet CEO Sundar Pichai took the stage at an event at Stanford University’s Graduate School of Business on Wednesday and offered some insight into how he views running one of the world’s most valuable tech companies.

It’s a noteworthy appearance because Pichai has been through some tough stuff lately. Google is widely seen as a late starter in generating artificial intelligence, lagging behind Microsoft-funded OpenAI. Still, the company under Pichai has been focused on artificial intelligence for much of the past decade, with Google researchers writing a formative paper on the transformer model that really kicked off the generative AI revolution. Recently, Alphabet’s Gemini LL.M. has been heavily criticized for producing wildly inaccurate images of historical situations, such as depicting the United States’ Founding Fathers as black or Native American rather than white British, suggesting certain types of bias. overcorrected.

The interviewer was Jonathan Levin, dean of the Stanford Graduate School of Business, who was not a hostile investigator—eventually, he revealed that the two men’s sons had played together in a middle school band— Pichai is good at answering difficult questions by framing them as further questions about how he thinks rather than directly answering them. But there was something interesting during the conversation.

At one point, Levine asked Pichai what he was trying to do to keep a company with 200,000 employees innovative against all the startups trying to disrupt its business. This is clearly something that worries Pichai.

“Honestly, this is an issue that has kept me up at night for years,” he began. “One of the inherent characteristics of technology is that you can always develop amazing things with small teams outside. History has proven this. Scale doesn’t always give you that – regulators may disagree, but I’ve always felt, at least when running a company, you’re always susceptible to whoever is in the garage with a better idea. So I think, as a company, how do you move fast? How do you have a culture of risk-taking? How do you incentivize that? ? These are things where you actually have to put a lot of effort into it. I think at least the larger organizations tend to default. One of the most counterintuitive things I’ve seen is that the more successful something is, the more risk-averse people are. It’s so It’s counterintuitive. You often find that small companies almost make the decision to bet the company, but the bigger you get, and that’s true for big universities, and that’s true for big companies, you stand to lose more, or you Think you have more to lose. So you find you don’t take as many ambitious risk-taking moves. So you have to do it consciously. You have to push the team to do it.”

He didn’t offer any specific strategies that have proven successful at Google, instead pointing out how difficult it is to create the right incentives.

“One example I often think about is how to reward effort, risk-taking and good execution, not always results. It’s easy to think that results should be rewarded. But then people started playing with it, right? People take a conservative approach and you will get good results. “

He recalls an earlier time when Google was more willing to take odd risks, pointing specifically to the company’s ill-fated Google Glass; it didn’t succeed, but it was one of the first devices to try out augmented reality.

“WRecently, we returned to an idea from the early days of Google Labs. So we’re building something that makes it easier to launch something without always having to worry about, you know, the full branding and weight of building a Google product. How can you roll something out in a simple, lightweight way? How do you make it easier for people to prototype internally and get it out to people? “

Later, Levine asked Pichai what development he was most excited about this year.

First, he cited the multimodality of Google’s latest LLM—that is, its ability to handle different types of input at the same time, such as video and text.

“All of our AI models are already using Gemini 1.5 Pro now; it’s a million contextual windows and it’s multi-modal. The ability to process a lot of information in any type of way on the input and deliver it on the output , I think it’s exciting in a way that we haven’t quite processed yet.”

Second, he highlighted the ability to connect disparate discrete answers together to provide smarter workflows. “Today you use the LL.M. simply to seek information, but Linking them together in a way that can handle workflow can be very powerful. It might make Stanford Hospital’s billing system easier,” he joked.

You can watch the entire interview, as well as a previous interview with Federal Reserve Chairman Jerome Powell, on YouTube. Levin and Pichai start around 1 hour and 18 minutes.