IT Home On October 24, according to Business Insider, Google CEO Sundar Pichai believes that there is still a long way to go before AI can improve itself completely autonomously.
Pichai revealed that Google researchers have started using AI agents to assist in model development, but the technology is still in its early stages. “We are still far from the final form of Recurrent Self-Improvement (RSI). To be precise, at every stage of model development, AI researchers have begun working in collaboration with the self-improvement cycle. It can be said that AI agents are helping researchers develop models in a certain degree of automation.”
Pichai expects further progress to be seen next year.
Sergey Brin, co-founder of Google, said earlier this month that the company’s Flash series models and Gemini 4 have demonstrated the ability of AI to participate in their development. Currently, researchers are working with AI models to develop the next generation of models.
According to reports by IT Home, in April this year, Google revealed that 75% of the new internal code was generated by AI. However, Pichai believes that simply counting the proportion of AI-generated code cannot accurately reflect the actual situation. “For us, this number has little meaning, because now every piece of code we process is deeply influenced by AI.”
Pichai also mentioned that Gemini 4 Argon will help people handle more complex programming challenges. The number of times Google restores software to an older version due to programming errors or system failures has also been reduced by one-third. “Therefore, I believe that AI tools not only improve programming efficiency, but also enhance code quality.”