Google promises not to use AI for weapons
Google has removed a significant portion of its AI principles that prevented the company from developing artificial intelligence for weapons and surveillance. The change has been abolished by the “application will not be pursued”, which clearly rejected the AI to harm people.
Non -branded – Knowledgeable
A startup chip maker, a positron that aims to compete with Nvidia NVDA.O, said Tuesday that it collected the production of its US -made artificial intelligence chips. 5 23.5 million.
The Velver Equity Partners, who are known for supporting Elon Musk’s companies, ATTIDS Management, Fallum Ventures and Flexible Reserve, were among the investors who participated in the round.
Reno -based posterone says its chips, which are manufactured in Arizona, maintain the same performance, NVIDIA’s top -off line H100 uses less than one -third of the power of H100 graphical processing units.
The purpose of posterone chips is to use AI models, instead of training AI models, for this period. At this time, the demand for training chips is high, but analysts have predicted that the need for chips can be left behind because more AI applications are deployed.
Generative AI model maker such as Openi, Google Google Dot O and Meta MetaDo have said they will invest in AI infrastructure.
Meta has said it will cost $ 65 billion this year, while Microsoft has said it will cost $ 80 billion. Openi announced last month’s $ 500 billion star gate infrastructure program.
Currently, the market share in NVIDIA chips is about 80 %, but due to growing costs and concerns about relying on a single supplier, consumers like Microsoft, Meta and Openi find internal or external alternatives Can go
Reporting by Anna Tong in San Francisco; Edit by Kate May Berry
