Google forms dedicated team to enhance code for large models

According to three informed sources, Google has assembled a special task force composed of researchers and engineers to optimize its artificial intelligence code models, aiming to achieve more automation of internal code work and ultimately drive the autonomy of AI research. Two of the sources said that this move was partly inspired by the recent release of a new model by the AI company Anthropic. Researchers at Google DeepMind believe that Anthropic's code tools have surpassed Google's Gemini model in terms of coding capabilities. The ultimate goal of the project is to achieve autonomous iterative upgrades of artificial intelligence, where AI can self-optimize, a direction that Google co-founder Sergey Brin is particularly focused on. Brin has told employees that improving Google's AI coding capabilities is a key step towards this ultimate goal. An informed source revealed that both Brin and Google DeepMind's Chief Technology Officer, Koray Kavukcuoglu, are directly involved in the work of this special task force, highlighting their importance in Google's senior strategic planning.