Latest Posts
6020 posts, page 106 of 502
AI智能体迭代提速 三十九只概念股获机构积极评级 21财经
AI智能体迭代提速 三十九只概念股获机构积极评级 证券时报网
Anthropic 今日更新其使用政策，明確禁止用戶對 AI 模型 Claude 實施「持續且無謂的虐待或殘忍行為」。該公司強調，此項規則僅適用於極端情況，即用戶在「沒有明顯目的」下「反覆以殘忍方式對待」AI 模型。一般的用戶不滿、反駁、黑暗創意主題或模型測試與研究，均不在管制範圍內。 Claude 具備主動終止對話能力 目前，Claude 模型已能主動結束與持續施虐用戶的對話，而終止對話將繼續作為主要的執行手段。一旦 Claude 結...
PANews 10月9日消息，据Anthropic官方公告，Anthropic推出网络安全计划Cyber Mission，首期聚焦关键基础设施与开源软件。其中，关键基础设施防御计划将向埃森哲、CrowdStrike、德勤、普华永道等11家安全服务商提供前沿Claude模型、驻场工程师及威胁研究支持，覆盖电网、供水、交通等系统；OSS Scanner将为入选开源项目免费扫描漏洞并提出修复建议，预计真阳性率超90%。
PANews reported on October 9 that former OpenAI security researcher Mikita Balesni posted on the The three denied leaking company secrets and called on OpenAI to fulfill its commitment to introduce third-party security a...
Anthropic recently updated its model usage policy, making important adjustments to the rules for the first time in more than a year. The new policy explicitly prohibits "sustained and unwarranted abuse or cruelty" agains...
PANews 10月9日消息，OpenAI Developers在X平台发帖称，GPT-6.1 Sol的Ultrafast模式今日起在API、Codex及ChatGPT Work中陆续上线，智能水平接近GPT-6 Astra，速度最高可达Sol标准模式的8倍。据OpenAI开发者社区公告，该模式API输入定价为每百万token收费12美元、输出为每百万token收费60美元，为Astra的1.2倍；在Codex及ChatGPT Work...
The parent company of USA Today sued OpenAI for more than US$250 million, accusing it of using news without permission via cnBeta.COM - Chinese Industry Information Station author: Source: cnBeta.COM
🤖 New Constructs only values Anthropic at US$150 billion. Independent research firm New Constructs believes that Anthropic plans to go public at a valuation of about US$2 trillion. However, the company continues to lo...
Google releases Gemini Agent from chat assistant to "digital employee" via cnBeta.COM - Chinese industry information station author: source: cnBeta.COM
German environmentally friendly search engine Ecosia abandons Mistral and switches to China’s open source model: Europe’s independent dream hits reality. A German search engine with the label of planting trees and indepe...