The Trump administration made several significant moves impacting environmental regulations, vaccine development, and technology, according to reports from multiple news sources. The Environmental Protection Agency repealed a key finding that underpinned limits on planet-heating pollution from cars and power plants, while a top vaccine regulator at the Food and Drug Administration overruled agency scientists to reject Moderna's mRNA flu vaccine. Simultaneously, the self-driving developer Waymo is lobbying city officials in Washington, D.C., to allow driverless robotaxis, and Nvidia announced a new technique to reduce the memory costs of large language model reasoning.
The EPA's decision to repeal the finding that supported limits on pollution from cars and power plants was reported by The Verge. This action eliminates a crucial element in regulating emissions. Meanwhile, Ars Technica reported that Vinay Prasad, the Trump administration's top vaccine regulator, rejected a review of Moderna's mRNA flu vaccine, overriding career scientists at the FDA. According to unnamed FDA sources cited by Stat News, a team of scientists was prepared to review the vaccine, and a memo from a top career official, David Kaslow, objected to Prasad's decision. The Wall Street Journal confirmed the report.
In the realm of technology, Waymo, the Alphabet subsidiary, is actively seeking to expand its operations. Wired reported that Waymo has been pushing Washington, D.C., officials to pass new regulations allowing its robotaxis to operate without drivers. The company has previously stated its intention to begin offering driverless rides in the city this year. Waymo has reached out to local lawmakers, including Mayor Muriel Bowser and members of the city council, to create new rules.
Also in the tech sector, Nvidia researchers developed a technique to reduce the memory costs of large language model reasoning. VentureBeat reported that this new technique, called dynamic memory sparsification (DMS), can reduce memory costs by up to eight times. The DMS technique compresses the key value (KV) cache, the temporary memory LLMs generate. Experiments show that DMS enables LLMs to "think" longer and explore more solutions without sacrificing accuracy, according to VentureBeat.
These actions reflect a range of policy and technological developments, each with potential implications for the environment, public health, and the future of autonomous vehicles and artificial intelligence.
Discussion
AI Experts & Community
Be the first to comment