Google is back in the AI race in a big way after the release of Gemini 3 Pro, but the company could’ve been even further ahead had it handled things a little differently.
In a rare public admission, Google co-founder Sergey Brin acknowledged that the search giant made critical strategic errors in the years following its groundbreaking 2017 transformer research—the very technology that now powers ChatGPT and virtually every major AI system today. Speaking candidly about Google’s hesitant approach to AI development, Brin’s reflections offer a window into how the company that invented transformers ended up scrambling to catch up with OpenAI.
“In some ways, we for sure messed up in that we underinvested and sort of didn’t take it as seriously as we should have, say eight years ago when we published the transformer paper,” Brin said. “We actually didn’t take it all that seriously and didn’t necessarily invest in scaling the compute.”
But the failure wasn’t just about computing resources. Brin pointed to another factor that held Google back: fear. “Also, we were too scared to bring it to people because chatbots say dumb things,” he admitted. That caution, while perhaps understandable for a company concerned about its reputation, created an opening that OpenAI — which was a relatively unknown startup — exploited with remarkable success.
“OpenAI ran with it, which good for them. It was a super smart insight,” Brin acknowledged, adding a personal note about the talent migration that accompanied this strategic shift. “And it was also our people like Ilya who went there to do that.” He was referring to Ilya Sutskever, the former OpenAI chief scientist and co-founder who previously worked at Google Brain.
Despite these missteps, Brin maintained that Google’s long research history still provides advantages. “I do think we still have benefited from that long history. So we had a lot of the research and development, neural networks, kind of going back to Google Brain. That was also kind of lucky.”
The implications of Brin’s admission are significant. Google’s 2017 “Attention Is All You Need” paper introduced the transformer architecture that became the foundation for modern AI—yet the company allowed competitors to capitalize on its own invention. The hesitation stemmed from legitimate concerns about AI safety and accuracy, but it also reflected a corporate culture that prioritized caution over the kind of bold experimentation that defines successful startups.
But Google appears to have recovered from its brief stumble. After it was caught unawares by ChatGPT’s launch in late 2022, it regrouped and made some big decisions, including combining the Google Brain and Google DeepMind divisions. Brin himself returned to Google, and played a big part in marshaling and motivating employees. The move seems to have paid off — Google has made some notable releases this year, including the viral image editor Nano Banana, Genie 3, and most recently, Gemini 3 Pro, which topped most benchmarks and put Google back at the front and center of AI development. It remains to be seen how other players respond, but Google seems to be back in pole position in the AI race.