| Mobile Web

Search results for DeepMind

AI & Enterprise

Google develops Frozen v2 AI inference chip, up to 10 times more efficient than TPU

Google is developing a new server chip that directly integrates the design of its Gemini AI model, The Information reported on July 21, citing two people familiar with internal matters. The chip, called Frozen v2 inside Google, is aimed at addressing an AI computing power shortage that has fueled internal conflict and led Google Cloud to turn down external customer contracts. Employees involved expect Frozen v2 to be 6 to 10 times more efficient than the latest TPU by tokens processed per unit of power. Google plans to deploy it as early as 2028.

AI & Enterprise

Hassabis says AGI coming within years, proposes U.S.-led frontier AI standards body

Google DeepMind CEO Demis Hassabis proposed creating an international standards body ahead of the arrival of artificial general intelligence, or AGI. Writing on X, he said risks grow as AGI nears. He suggested a FINRA-style \"frontier AI standards body\" as a public-private partnership under federal oversight, funded mostly by industry. Models meeting its criteria would be classified as frontier-level, with labs encouraged to follow best practices and submit models for review.

AI & Enterprise

Google DeepMind CEO says AGI could arrive within years, calls for U.S. standards body

Google DeepMind CEO Demis Hassabis said artificial general intelligence could arrive within a few years and urged the United States to create an independent standards body to evaluate advanced AI models before release. He said AGI could have an impact comparable to discoveries such as electricity or fire and warned that safety risks are already emerging, with potential threats extending to biology, nuclear technology and national security. He proposed a FINRA-like model under government oversight with industry funding and independent experts.