Tech giant Google is making radical changes to the leadership of its DeepMind artificial intelligence unit. The restructuring aims to overcome the technological lag of its flagship Gemini model compared to products from competitors such as OpenAI and Anthropic, as well as to strengthen the commercialization of its developments.
Tech giant Google is making radical changes to the leadership of its DeepMind artificial intelligence unit. The restructuring aims to overcome the technological lag of its flagship Gemini model compared to products from competitors such as OpenAI and Anthropic, as well as to strengthen the commercialization of its developments.
Reuters reports on the details of internal processes and personnel reshuffles in the company, citing its own sources among employees.
While last fall’s Gemini update briefly gave Google a head start, new releases from other labs (including Anthropic’s powerful Claude Mythos) have once again put the company in a catch-up position. The situation escalated in August, when the company was forced to delay the release of its new flagship version of Gemini by two months. Internal testing showed that the model still lags behind its competitors, especially in coding tasks.
According to insiders, Google co-founder Sergey Brin is playing a key role in the efforts to accelerate development. Although he does not formally hold a leadership position, Brin uses his informal influence to direct resources to critical areas. The main focus is recursive self-improvement, that is, the ability of AI to improve its own algorithms without human intervention. During a meeting with the team in the spring, he directly urged employees to discard outside projects and focus all efforts on the development of Gemini.
To cut through bureaucracy and streamline operations, the company's leadership has made personnel changes at DeepMind. Longtime lab leader Demis Gassabis has been moved to a less senior role to focus on long-term research. Operational management and ultimate decision-making power have been transferred to his deputy, Koray Kavukcuoglu.
Cavuccoglu has a reputation for being a specialist who better understands the corporation's commercial needs. To manage more effectively, he moved from London to Google's headquarters in California, consolidating power.
The publication's sources note that these changes indicate a gradual loss of autonomy for the DeepMind laboratory, which Google acquired back in 2014. Management has already informed staff that some non-technical teams will be transferred directly to the company's corporate structure, which has raised concerns among researchers about the future of independent scientific projects.
Recall that the other day, Alphabet CEO Sundar Pichai announced large-scale personnel changes in the company's top artificial intelligence management.
