人工智能安全领域近期因开源模型的潜在风险引发激烈辩论。Anthropic发布报告指出,中国公司Z.ai推出的新型强大模型因其允许任何人免费下载和修改的特性,构成了重大的网络攻击威胁 1。这一警告加剧了业界关于前沿AI模型是否应保持开放性的讨论 1。与此同时,Anthropic已将其首次公开募股(IPO)的目标估值从2万亿美元下调至1.4万亿美元 1。
除了AI安全问题,科技界还面临其他重大地缘政治与法律挑战。涉及英伟达的间谍丑闻持续发酵,据报该事件关联价值75亿美元的芯片交易,Alice Huang据称获取了24万颗受限芯片 1。美国参议院少数党领袖Chuck Schumer指责总统Donald Trump忽视了此前关于Nvidia首席执行官Jensen Huang与中国间谍存在所谓关系的警告 1。此外,MIT Technology Review的简报还提及乌克兰无人机袭击导致Yandex云客户数据损失3%,以及COP31峰会仅有约十几个国家准备签署为35%能源使用提供电力的旗舰承诺等动态 1。
Anthropic has issued a warning regarding significant cybersecurity threats posed by a powerful new model released by Chinese company Z.ai, which is available for free download and modification by anyone. This assessment intensifies the ongoing debate over the openness of frontier AI models and their potential misuse for cyberattacks 1.
The tech landscape is also marked by serious allegations involving major hardware firms and shifting valuations in the AI sector. Reports indicate that a Nvidia spy scandal involves $7.5 billion worth of chips, with Alice Huang allegedly obtaining 240,000 restricted units, prompting Senate Minority Leader Chuck Schumer to accuse President Donald Trump of ignoring warnings about ties between Nvidia CEO Jensen Huang and Chinese spies 1. Concurrently, Anthropic revised its initial public offering valuation downward from a target of $2 trillion to $1.4 trillion 1.
In other global developments, a Ukrainian drone strike destroyed 3% of Yandex cloud customer data, while efforts at COP31 face limited support as only about a dozen nations are prepared to sign the flagship commitment to electrify 35% of energy use 1.
评论
还没有评论,欢迎留下第一条。