Anthropic-Pentagon Contract Talks Stalled Over AI Safeguards
Anthropic PBC's negotiations with the Pentagon to extend a contract for its Claude AI tool are stalled due to the company's insistence on implementing additional safeguards against misuse, specifically to prevent mass surveillance and the development of autonomous weapons. The Pentagon, however, seeks the flexibility to use Claude without these restrictions, leading to a public dispute over the technology's applications. This disagreement has raised concerns about potential repercussions for defense contractor Palantir, which provides the infrastructure for the military's use of Claude.
Meta Signs Multiyear Deal to Buy Nvidia Processors
Meta Platforms Inc. has entered a multiyear agreement to purchase millions of processors from Nvidia Corp., enhancing their collaboration in the artificial intelligence sector. This deal will see Meta, which contributes approximately 9% to Nvidia's revenue, utilize more of Nvidia's AI processors and networking equipment, including the deployment of Nvidia's Grace central processing units in standalone systems.
Researchers Identify Security Flaws in RISC-V CPUs
Researchers at the CISPA Helmholtz Center for Information Security have published a paper titled 'RISCover,' which addresses the automatic discovery of user-exploitable architectural security vulnerabilities in closed-source RISC-V CPUs. The study highlights the challenges posed by the proprietary nature of most commercial CPUs, emphasizing the need for improved security measures in this emerging architecture.
