Google co-founder Sergey Brin (Centre)
Speaking during a roundtable discussion focused on the development of Google’s new Gemini 4 model Argon, Brin explained that Google is approaching a critical technical threshold known in computer science as recursive self-improvement (RSI). Brin pointed to the rapid succession of Google’s Flash models preceding Gemini 4 as concrete evidence that automated engineering loops are already taking hold inside the tech giant’s labs. And if you look at even before Gemini 4, when we had Flash 3.6, 3.7, 3.8, we’re starting to get onto that RSI treadmill,” he added.
Because melvin and I were talking about the next version of the model,” Brin said, actually, I just popped in here randomly. Google co-founder Sergey Brin has revealed that the company’s artificial intelligence (AI) development is accelerating rapidly, with current models now directly being developed by both humans and AI.
The Google co-founder noted that the company is “starting to get onto that RSI treadmill,” adding that internal teams are moving so fast that by the time the public views the video discussing Argon’s rollout, internal engineering tracks will already be a couple of model iterations ahead. “AI agents now operate as high-leverage assistants that author a majority of production code and bug fixes, though code changes remain subject to human review before submission,” the safety report stated. He explained that current internal AI models are collaborating with company researchers to design and train their own successors, essentially signaling that there will be no slowing down in the race for advanced systems. Under an RSI framework, an AI model assists in engineering (or writing) superior algorithms, codebases and architectures, creating an automated flywheel effect – a concept of several small, consistent efforts combining over time to build self-sustaining momentum – that dramatically shortens the lifecycle required to ship subsequent generations. “Our researchers and the model work hand in hand to make the next version of the model. “I’m very excited about the cadence that we have nowadays.
“And probably by the time folks are viewing this (Gemini 4 Argon) internally, we’re already using a couple models ahead.
Brin said that Google teams have already been evaluating an advanced, unreleased successor to Argon. Brin has become deeply engaged in Google’s Gemini initiatives, using his status as co-founder to redirect engineering talent and computing infrastructure toward projects with the highest impact, including an intense focus on RSI. So it’s just very exciting to see the progress and the cadence. And I expect it’ll only continue to accelerate. His comments came even as a large group of tech leadership is urging the AI companies to slow down AI development, highlighting the ‘doomerism’ associated with the technology. You use AI every day. Now get your AI Quotient. Take the AIQ test.