Sergey Brin says he doesn’t think Google engineers use AI for coding as much as they should

Sergey Brin poses for a picture on the red carpet in 2017.

Google cofounder Sergey Brin said engineers at the tech company aren’t using artificial intelligence as often as he thinks they should.

Brin made the comments during a live conversation with entrepreneur David Friedberg at the “All-In” summit in Los Angeles this week. The exclusive three-day event from the creators of the “All-In” podcast, where industry veterans Chamath Palihapitiya, Jason Calacanis, David Sacks, and Friedberg cover all things economy, tech, politics, and poker.

The summit included a rare live appearance by Brin, who discussed the “mind blowing” advancement of AI.

“As a computer scientist, I’ve never seen anything as exciting as all of the progress that’s happened the last few years,” Brin said.

Brin, who said he’s spending nearly every day at Google these days, praised AI’s multi-faceted abilities, specifically with regard to programming.

“Writing code from scratch feels really hard compared to just asking the AI to do it,” Brin said to laughs from the audience.

Brin said he was curious about how good Google’s AI model would be at Sudoku. As an experiment, Brin had the model write code that would automatically generate versions of the puzzle. The test sparked some debate among Google engineers, Brin said.

But when he came back just 30 minutes later, the model had successfully done its job, Brin said.

“They were kind of impressed,” Brin said of the engineers. “Because they don’t honestly use the AI tools for their own coding as much as I think they ought to.”

Google did not immediately respond to a request for comment from B-17.

Brin, later in the conversation, also said that Google had been too “timid” to release its early language models because they made mistakes and said “embarrassing things.”

With AI’s programming capabilities, Brin said, “You need to be willing to have some embarrassments and take some risks, and I think we’ve gotten better at that, and well, you guys have probably seen some of our embarrassments.”

Google’s approach to utilizing AI hasn’t stopped its employees from getting poached as the AI talent war heats up.

B-17 reported in June that OpenAI had hired at least 44 ex-Googlers this year alone, about half of whom were engineers.

Google has taken a proactive approach in trying to stay AI-relevant. As search competition continues to heat up, Google recently updated its AI Overviews, which is a feature that produces AI-generated answers in response to a Google search. The company made the changes weeks after OpenAI revealed a rival search engine prototype.

Engineers’ caution around AI is not entirely unfounded. Coding tools are sparking anxiety in the industry about future job security. Earlier this month, B-17 reported that junior software developers are likely to see the most immediate changes to their workflow, with AI tools potentially changing the way newbies gain developer skills.

At the All-In summit this week, Brin and Friedberg also discussed the future of human-AI interaction and the race to create human-like AI.

The event featured several other big-name speakers, including Tesla CEO Elon Musk.