Sergey Brin Makes Surprise Appearance at Google I/O 2025, Says Gemini Aims to Be First AGI

When Sergey Brin quietly walked back into Google in 2023, most of the tech world didn’t notice. He wasn’t making press rounds or stepping into boardrooms. He was back in the labs—hands-on, deep in code, tinkering with the AI systems that would eventually power Gemini.

But last week, at Google I/O 2025, Brin made that return official. What was supposed to be a solo fireside chat with DeepMind CEO Demis Hassabis turned into something much more interesting when Brin joined him on stage, unannounced.

Google I/O 2025  

And just like that, one of Google’s co-founders was back in the spotlight—talking AGI, cracking jokes, and laying out his vision for the future of artificial intelligence.

“I Torture People Like Demis”

In typical Brin fashion, his on-stage presence was casual but sharp. “I torture people like Demis, who is pretty amazing,” he quipped, lightening the mood. “He tolerated me crashing this fireside.”

But jokes aside, it’s clear Brin isn’t just there to observe. “I go to the office pretty much every day now,” he said. And his focus? The technical core of Gemini, Google’s next-gen AI model.

He’s not managing teams or pushing roadmaps. He’s in the code. Debugging. Probing. Asking uncomfortable questions. In his words, that’s a “luxury” he now enjoys—because others like Hassabis are handling the rest.

Why He Came Back

Brin stepped down from his executive role at Alphabet in 2019. For years, he was out of the game—officially retired. But the rise of OpenAI, Anthropic, and a new wave of foundational models brought him back in.

“As a computer scientist, it’s a very unique time in history,” Brin said. “Honestly, anybody who’s a computer scientist should not be retired right now. Should be working on AI.”

It wasn’t about corporate urgency. It was personal curiosity. He saw something happening in the world of AI that felt too important to ignore.

The AGI Debate: Sooner or Later?

The biggest moment of the talk? A simple question: When will we achieve AGI?

“Before 2030,” Brin said with confidence.

“Just after,” Hassabis replied, a little more cautiously.

It was a rare bit of direct forecasting from two people who are shaping the future of AI. But what made the moment even more interesting was how aligned they were on the how of AGI—even if they differed on the when.

For Hassabis, AGI is about a system that mirrors the human brain’s ability to reason, invent, and create across a wide range of tasks. He admitted we’re not there yet. Today’s models aren’t consistent or reliable enough to be called truly general. But he also said we might only need “one or two more breakthroughs” to get there.

Brin didn’t just agree—he doubled down. “We fully intend that Gemini will be the very first AGI,” he said.

From Google Glass to XR Glasses—A Second Shot

Brin also took time to reflect on one of his past projects: Google Glass. A decade ago, it fizzled. The product was ahead of its time, but the supply chain, pricing, and hardware challenges held it back.

Now, he’s betting that AI might finally make smart glasses viable.

At I/O 2025, Google showed off a prototype pair of XR-powered glasses, equipped with Gemini AI features. They respond to voice commands, offer real-time translation, and assist with tasks in the real world—all in a more subtle, integrated form than the original Google Glass ever could.

Brin admitted the earlier attempt failed partly because he didn’t understand how hard consumer electronics manufacturing really was. But this time, Google is leaning on partners like Samsung and anchoring the project in Gemini’s AI platform.

Brin Is All In

It’s clear this isn’t a short-term hobby for Sergey Brin. He’s not returning for nostalgia or a vanity project. He’s here because he genuinely believes we’re at a tipping point in tech—and that Google has a shot at leading the next big leap.

Via