https://t.me/AI_News_CN
📈主流AI服务状态页通知 | 🆕汇集全网ChatGPT/AI新闻 #AI #ChatGPT
🔙备用群 https://t.me/gpt345
✨BEST AI中转 https://api.oaibest.com 2.8折起 支持OpenAI, Claude code, Gemini,Grok, Deepseek, Midjourney, 文件上传分析
Buy ads: https://telega.io/c/AI_News_CN
📈主流AI服务状态页通知 | 🆕汇集全网ChatGPT/AI新闻 #AI #ChatGPT
🔙备用群 https://t.me/gpt345
✨BEST AI中转 https://api.oaibest.com 2.8折起 支持OpenAI, Claude code, Gemini,Grok, Deepseek, Midjourney, 文件上传分析
Buy ads: https://telega.io/c/AI_News_CN
美国国家安全局(NSA)正在使用Anthropic公司近期宣布但未向公众开放的高级人工智能模型“Mythos Preview”,引发外界关注。 数周前,NSA的上级机构美国国防部(五角大楼)才刚刚将Anthropic列为“供应链风险”,理由是该公司拒绝向五角大楼开放其模型的全部功能以供不受限制地使用。
Anthropic本月早些时候发布“Mythos”,将其定位为专注网络安全任务的前沿模型,但同时强调,这一模型在发起进攻性网络攻击方面能力过于强大,因此不适合公开向大众开放。 出于风险考量,公司目前仅向大约40家机构开放“神话”的访问权限,其中公开点名的只是其中约十来家合作方。
据报道,NSA应当属于未公开名单中的一员,目前主要将“Mythos”用于扫描各类系统和环境中可被利用的安全漏洞。 英国人工智能安全研究机构“AI Security Institute”也已证实获得该模型的访问权限,用于自身的安全评估与研究工作。
值得注意的是,在美国军方不断扩大对Anthropic工具使用范围的同时,五角大楼在法庭上却一再强调,这些工具本身可能危及国家安全。 双方争端的起点是Anthropic拒绝配合国防部,将其旗舰模型Claude用于大规模国内监控以及自主武器系统研发。
在NSA接入“Mythos”的同时,Anthropic与特朗普政府之间一度紧张的关系似乎也有所缓和。 上周五,该公司首席执行官达里奥·阿莫代(Dario Amodei)在白宫与总统办公厅主任苏西·怀尔斯(Susie Wiles)以及财政部长斯科特·贝森特(Scott Bessent)举行了会面。 据消息人士称,白宫方面将这次会面形容为“富有成效”。
相关文章:
联邦机构绕过特朗普的Anthropic禁令 测试其先进AI模型 “Claude Mythos”
via cnBeta.COM - 中文业界资讯站 (author: 稿源:cnBeta.COM)
Users unable to load ChatGPT and Codex
Status: Investigating
We are investigating the issue for the listed services.
Affected components
● Image Generation (Degraded performance)
● GPTs (Degraded performance)
● Connectors/Apps (Degraded performance)
● ChatGPT Atlas (Degraded performance)
● App (Degraded performance)
● Login (Degraded performance)
● Conversations (Degraded performance)
● Compliance API (Degraded performance)
● Agent (Degraded performance)
● Deep Research (Degraded performance)
● File uploads (Degraded performance)
● Voice mode (Degraded performance)
● Search (Degraded performance)
via OpenAI status
Status: Investigating
We are investigating the issue for the listed services.
Affected components
● Image Generation (Degraded performance)
● GPTs (Degraded performance)
● Connectors/Apps (Degraded performance)
● ChatGPT Atlas (Degraded performance)
● App (Degraded performance)
● Login (Degraded performance)
● Conversations (Degraded performance)
● Compliance API (Degraded performance)
● Agent (Degraded performance)
● Deep Research (Degraded performance)
● File uploads (Degraded performance)
● Voice mode (Degraded performance)
● Search (Degraded performance)
via OpenAI status
Elevated errors for uploading files
Apr 20, 13:31 UTC
Investigating - We are currently investigating an issue related to uploading files on Claude.ai and Claude API. Some users will see their uploads failing and they are advised to retry.
via Claude Status - Incident History
Apr 20, 13:31 UTC
Investigating - We are currently investigating an issue related to uploading files on Claude.ai and Claude API. Some users will see their uploads failing and they are advised to retry.
via Claude Status - Incident History
DeepSeek Web/APP 快速模式性能异常(DeepSeek Web/APP Instant Mode Performance Abnormal)
Apr 20, 20:40 CST
Investigating - We are currently investigating this issue.
via DeepSeek Service Status - Incident History
Apr 20, 20:40 CST
Investigating - We are currently investigating this issue.
via DeepSeek Service Status - Incident History