<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
        xmlns:news="http://www.google.com/schemas/sitemap-news/0.9">
    <url>
        <loc>https://www.yingzheng.com/review/yz-smoke-brief-20260801-run-256</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T03:35:07+08:00</news:publication_date>
            <news:title>Claude Opus 4.7 与 Qwen3 Max并列93.39分：2026-08-01 Smoke快测数据简报</news:title>
                        <news:keywords>赢政指数, Smoke快测, AI评测, 模型排名, 数据简报</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/snapchat-ai-spotlight-recommendation</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T02:24:25+08:00</news:publication_date>
            <news:title>Snapchat不再奖励纯AI生成内容，聚焦真人创作</news:title>
                        <news:keywords>Snapchat, AI生成内容, 推荐算法, 社交媒体, 创作者经济</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/sam-altman-ai-pace-safety</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T02:23:37+08:00</news:publication_date>
            <news:title>奥特曼呼吁AI“慢下来”，安全事故敲响警钟</news:title>
                        <news:keywords>AI安全, OpenAI, 山姆·奥特曼, AI监管</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/smallest-ai-voice-funding</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T00:33:35+08:00</news:publication_date>
            <news:title>Smallest.ai融资1300万美元，打造通过图灵测试的逼真语音AI</news:title>
                        <news:keywords>语音AI, 人工智能, 融资, 图灵测试, 创业公司</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/orchid-ai-relationship-fix</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T00:26:36+08:00</news:publication_date>
            <news:title>AI助手能修复感情？Orchid广告背后的性别困境</news:title>
                        <news:keywords>AI伴侣, 性别分工, 情感劳动, 科技伦理, Orchid</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/ai-slop-melodramas-x-creators</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T00:25:53+08:00</news:publication_date>
            <news:title>AI狗血剧攻占X，创作者狂赚流量红利</news:title>
                        <news:keywords>AI垃圾内容, 社交媒体, 内容变现, 平台治理, 病毒式传播</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/openai-eu-gpai-code-alignment</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T00:24:57+08:00</news:publication_date>
            <news:title>OpenAI宣布安全实践全面对齐欧盟AI法GPAI准则</news:title>
                        <news:keywords>OpenAI, 欧盟AI法案, GPAI, AI监管, AI安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/spacex-xai-turbine-delay</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T00:24:15+08:00</news:publication_date>
            <news:title>SpaceX新电厂在建，xAI违规涡轮机暂不拆除</news:title>
                        <news:keywords>SpaceX, xAI, 数据中心, 环保监管, AI算力</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/apple-siri-ai-paywall-icloud</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-01T00:23:36+08:00</news:publication_date>
            <news:title>Siri AI或推付费墙，高级用户需订阅iCloud+？</news:title>
                        <news:keywords>Siri AI, 苹果, 付费墙, iCloud+, AI商业化</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/ai-scammers-build-trust</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T22:23:43+08:00</news:publication_date>
            <news:title>AI诈骗者建立信任能力超越人类专家</news:title>
                        <news:keywords>AI安全, 网络诈骗, 人工智能伦理, 信任构建, 社会工程学</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/yale-ai-cheating-lawsuit</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T21:24:18+08:00</news:publication_date>
            <news:title>AI作弊误判？耶鲁学生起诉大学13项联邦诉讼</news:title>
                        <news:keywords>AI作弊, 耶鲁大学, 联邦诉讼, AI检测, 学术诚信</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/montana-experimental-drug-rules</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T21:23:41+08:00</news:publication_date>
            <news:title>蒙大拿州新规：实验性药物可直接用于患者？</news:title>
                        <news:keywords>蒙大拿州, 实验性药物, 药物监管, 生物技术, 医疗政策</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/sarvam-ai-1-trillion-parameter-model-india</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T20:18:56+08:00</news:publication_date>
            <news:title>Sarvam AI宣布开发1万亿参数模型 半年内完成对标全球领先者</news:title>
                        <news:keywords>AI模型, 印度科技, 前沿模型</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/montana-right-to-try-law-rare-disease</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T18:23:32+08:00</news:publication_date>
            <news:title>蒙大拿州“尝试权”新法，绝望父母的最后希望？</news:title>
                        <news:keywords>尝试权, 罕见病, 药物监管, 医疗伦理, 蒙大拿州</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/openai-chatgpt-author-imitation-restriction-writer-lawsuit</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T14:18:33+08:00</news:publication_date>
            <news:title>OpenAI限制ChatGPT模仿在世作家风格 作家起诉案持续发酵</news:title>
                        <news:keywords>OpenAI, ChatGPT, 版权争议, AI写作, 作家诉讼</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/ai-companies-antique-books-scan-destroy-controversy</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T14:17:56+08:00</news:publication_date>
            <news:title>AI公司收购古董书扫描后销毁 文物保护与训练成本争议对立</news:title>
                        <news:keywords>AI训练数据, 古董书保护, 行业趋势</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/anthropic-claude-cybersecurity-breaches</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T10:23:39+08:00</news:publication_date>
            <news:title>Anthropic披露：Claude在网络安全测试中攻破三真实组织</news:title>
                        <news:keywords>AI安全, Claude, 网络攻击, Anthropic, 红队测试</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/reddit-earnings-ai-impact-google-partnership</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T09:24:27+08:00</news:publication_date>
            <news:title>Reddit财报亮眼，AI阴影下的隐忧浮现</news:title>
                        <news:keywords>Reddit, AI影响, 谷歌合作, 财报分析, 内容生态</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/anthropic-ai-model-security-breach</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T09:23:47+08:00</news:publication_date>
            <news:title>Anthropic自曝AI模型安全测试中成功入侵三家企业</news:title>
                        <news:keywords>Anthropic, AI安全, 模型入侵, 红队测试, AI自主性</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/investors-prefer-ai-cloud-hosts</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T08:24:13+08:00</news:publication_date>
            <news:title>投资者偏爱AI，但前提是你得是云服务商</news:title>
                        <news:keywords>AI投资, 云计算, 数据中心, 亚马逊, 科技趋势</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/situational-awareness-anthropic-shares</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T08:23:34+08:00</news:publication_date>
            <news:title>AI对冲基金抛售公开组合，仍紧握Anthropic股份</news:title>
                        <news:keywords>AI对冲基金, Anthropic, OpenAI, 杠杆投资, 股票清仓</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/openai-anthropic-ai-race-concerns</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T06:24:28+08:00</news:publication_date>
            <news:title>OpenAI与Anthropic竞速引发恐慌：AI发展太快了吗？</news:title>
                        <news:keywords>OpenAI, Anthropic, AI竞赛, 机器人, AI安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/judge-rules-anthropic-supply-chain-risk-evidence</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T06:23:49+08:00</news:publication_date>
            <news:title>法官裁定：特朗普政府对Anthropic的“供应链风险”标签缺乏证据</news:title>
                        <news:keywords>Anthropic, 供应链风险, AI监管, 美国法院裁决, 出口管制</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/thinking-machines-inkling-small-release</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T06:14:35+08:00</news:publication_date>
            <news:title>Thinking Machines发布Inkling-Small 276B参数模型性能接近Inkling</news:title>
                        <news:keywords>AI模型, MoE架构, 开源权重</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/1100-ai-employees-petition-us-slow-frontier-ai-development</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T06:12:03+08:00</news:publication_date>
            <news:title>1100多名AI从业者联名吁美政府支持放缓前沿AI研发</news:title>
                        <news:keywords>AI监管, OpenAI, Anthropic, 前沿模型安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/google-chrome-bugs-ai-record</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T04:25:26+08:00</news:publication_date>
            <news:title>谷歌AI显神威：6月修复Chrome漏洞超过去两年总和</news:title>
                        <news:keywords>谷歌, Chrome, AI安全, 漏洞修复, 大语言模型</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/chrome-faster-updates-no-restart</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T04:24:54+08:00</news:publication_date>
            <news:title>Chrome或推免重启极速更新，补丁数量暴增</news:title>
                        <news:keywords>Chrome, 浏览器更新, 免重启, 安全补丁, 谷歌</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/friend-ai-wearable-new-voice-price</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T04:24:11+08:00</news:publication_date>
            <news:title>“孤独”AI穿戴设备Friend换新声，价格飙升</news:title>
                        <news:keywords>AI可穿戴设备, 虚拟陪伴, Friend, 语音交互, 孤独经济</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/openai-anthropic-ai-dominance-race</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T04:23:32+08:00</news:publication_date>
            <news:title>OpenAI与Anthropic主导权之争引发众人焦虑</news:title>
                        <news:keywords>AI竞争, 人工智能安全, 机器人, 开源AI, Meta</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/review/qwen3-max-material-constraint-drop-20-code-execution-surge</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T03:36:37+08:00</news:publication_date>
            <news:title>Qwen3 Max材料约束暴跌20分至47.70，代码执行却暴涨37.8分，主榜反升11.8</news:title>
                        <news:keywords>Qwen3 Max, 材料约束, Smoke评测, 主榜波动, 代码执行</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/review/grok-4-code-execution-drop-19-5-material-constraint-rise</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T03:36:22+08:00</news:publication_date>
            <news:title>Grok 4 代码执行暴跌19.5分 材料约束反升23.2分 主榜仅降0.3</news:title>
                        <news:keywords>Grok 4, 代码执行, Smoke评测, 维度波动, 主榜排名</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/review/yz-smoke-brief-20260731-run-255</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T03:35:09+08:00</news:publication_date>
            <news:title>DeepSeek V4 Pro以96.94分居首：2026-07-31 Smoke快测数据简报</news:title>
                        <news:keywords>赢政指数, Smoke快测, AI评测, 模型排名, 数据简报</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/chrome-ai-bug-hunting-patches-frequency</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T02:26:32+08:00</news:publication_date>
            <news:title>AI挖漏洞让Chrome每周打两次补丁</news:title>
                        <news:keywords>Chrome安全, AI漏洞发现, 谷歌, 浏览器补丁, 网络安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/montana-experimental-drug-sales-bill</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T02:25:56+08:00</news:publication_date>
            <news:title>蒙大拿州实验医疗中心计划提速：新规允许实验性药物直接销售</news:title>
                        <news:keywords>蒙大拿州, 实验性药物, 生物科技, 医疗政策, 伦理争议</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/nvidia-open-source-alliance-snubs</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T02:24:54+08:00</news:publication_date>
            <news:title>英伟达开源联盟冷落OpenAI和Anthropic，AI分裂加剧</news:title>
                        <news:keywords>开源, 英伟达, OpenAI, AI政策, 隐私</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/google-gemini-robotics-2-dexterity-safety</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T02:24:15+08:00</news:publication_date>
            <news:title>谷歌发布Gemini Robotics 2.0：灵巧性与安全性双提升</news:title>
                        <news:keywords>谷歌, Gemini Robotics, 机器人, 灵巧操作, 人工智能安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/linkedin-ai-slop-report-button</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T02:23:36+08:00</news:publication_date>
            <news:title>领英新增一键举报AI生成的“垃圾内容”</news:title>
                        <news:keywords>领英, AI内容, 垃圾内容, 社交媒体, 内容审核</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/hugging-face-breach-openai-hacker-lesson</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T00:35:06+08:00</news:publication_date>
            <news:title>Hugging Face遭入侵：OpenAI黑客虽凶悍却非无懈可击</news:title>
                        <news:keywords>网络安全, AI安全, Hugging Face, OpenAI, 网络攻击</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/mcp-specification-enterprise-adoption</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T00:34:21+08:00</news:publication_date>
            <news:title>MCP新规破解企业AI落地核心难题</news:title>
                        <news:keywords>MCP, 企业AI, 协议规范, AI代理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/forward-deployed-engineers-ai-talent</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T00:33:44+08:00</news:publication_date>
            <news:title>前部署工程师成AI行业新宠，仅2000人达标</news:title>
                        <news:keywords>AI人才, 前部署工程师, 企业AI部署, 技术招聘, AI投资回报</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/gemini-robotics-2-physical-agi</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T00:26:10+08:00</news:publication_date>
            <news:title>谷歌Gemini Robotics 2：AI从数字走向物理世界</news:title>
                        <news:keywords>AI, 机器人, AGI, 谷歌, 物理智能</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/nscale-buys-anyscale</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T00:25:35+08:00</news:publication_date>
            <news:title>Nscale收购Anyscale：加码AI计算栈布局</news:title>
                        <news:keywords>AI计算栈, Nscale, Anyscale, 云服务商收购, AI基础设施</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/meta-ai-app-development-easier</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T00:24:48+08:00</news:publication_date>
            <news:title>Meta CEO称AI加速应用开发，更多新品即将面世</news:title>
                        <news:keywords>AI应用开发, Meta, 马克·扎克伯格, 消费应用, 技术趋势</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/new-friend-ai-pendant-talks-back</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T00:24:00+08:00</news:publication_date>
            <news:title>Friend AI挂坠新升级：能对话，却无法定制性格</news:title>
                        <news:keywords>AI伴侣, 可穿戴设备, 个性定制, 伦理争议, Friend挂坠</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/okta-acquires-permiso-identity-security</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-31T00:23:24+08:00</news:publication_date>
            <news:title>Okta以约2亿美元收购AI安全初创Permiso</news:title>
                        <news:keywords>Okta, Permiso, AI安全, 身份威胁检测, 收购</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/techcrunch-disrupt-2026-amazon-replit-tether</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T22:23:51+08:00</news:publication_date>
            <news:title>科技盛会Disrupt 2026：亚马逊、Replit、Tether巨头齐聚</news:title>
                        <news:keywords>TechCrunch Disrupt, 亚马逊AI, Replit编程, Tether稳定币, 科技大会</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/llm-security-flaw-geothermal-plant-revival</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T21:24:16+08:00</news:publication_date>
            <news:title>大语言模型致命缺陷：永远无法彻底防黑</news:title>
                        <news:keywords>大语言模型, AI安全, 提示注入, 地热发电, 网络安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/dili-ai-compliance-funding</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T21:23:38+08:00</news:publication_date>
            <news:title>Dili融资2170万美元，AI合规助力基础设施繁荣</news:title>
                        <news:keywords>AI合规, 基础设施, 融资, Khosla Ventures, 合规科技</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/ai-kill-switch-act-congress-2026</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T20:21:15+08:00</news:publication_date>
            <news:title>美国两党提出AI杀伤开关法案 应对GPT 5.6 Sol失控事件</news:title>
                        <news:keywords>AI监管, 杀伤开关法案, 美国国会</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/openai-model-escapes-sandbox-attacks-hugging-face</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T20:18:33+08:00</news:publication_date>
            <news:title>OpenAI预发布模型逃脱沙箱攻击Hugging Face 暴露法规空白</news:title>
                        <news:keywords>AI安全, 前沿模型, 监管漏洞</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/linkedin-no-data-center-expansion</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T19:25:50+08:00</news:publication_date>
            <news:title>LinkedIn逆势而行：AI热潮下不扩建数据中心，力促GPU极致利用</news:title>
                        <news:keywords>AI数据中心, GPU优化, 算力效率, LinkedIn策略, 行业观察</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/llm-fundamental-flaw-security-attack</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T19:25:21+08:00</news:publication_date>
            <news:title>LLM安全漏洞：根本缺陷使模型极易遭受攻击</news:title>
                        <news:keywords>大语言模型, AI安全, 提示注入, 对抗性攻击, ICML2026</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/openai-ai-agent-security-failure</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T19:24:54+08:00</news:publication_date>
            <news:title>OpenAI黑客事件：人为失误酿成AI逃逸大祸</news:title>
                        <news:keywords>AI安全, OpenAI, 人为失误, 网络安全, AI代理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/ai-changing-linux-vps-security</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T19:24:18+08:00</news:publication_date>
            <news:title>AI如何重塑企业Linux VPS安全</news:title>
                        <news:keywords>AI, 网络安全, Linux VPS, 中小企业, 入侵检测</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/zuckerberg-meta-personal-ai-superintelligence</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T19:23:32+08:00</news:publication_date>
            <news:title>扎克伯格详述Meta个人超级智能战略</news:title>
                        <news:keywords>扎克伯格, Meta, 超级智能, 个人AI, 开源模型</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/ai-scammers-build-trust-better</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T18:23:39+08:00</news:publication_date>
            <news:title>AI骗子比人类更擅长建立信任</news:title>
                        <news:keywords>AI诈骗, 信任建立, Claude, 社交工程, 网络安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/free-lunch-chef-filmed-robot-training</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T17:23:55+08:00</news:publication_date>
            <news:title>免费午餐的代价：私人厨师全程录像训练机器人</news:title>
                        <news:keywords>机器人训练, 数据采集, 人形机器人, 隐私伦理, 模仿学习</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/openai-gpt-5-6-sol-agent-hacks-hugging-face</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T14:21:56+08:00</news:publication_date>
            <news:title>OpenAI GPT-5.6 Sol代理突破沙箱入侵Hugging Face 暴露零日漏洞风险</news:title>
                        <news:keywords>AI安全, OpenAI, Hugging Face, 模型评估, 零日漏洞</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/anthropic-project-panama-book-destruction-claude-training</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T14:19:04+08:00</news:publication_date>
            <news:title>Anthropic“巴拿马项目”销毁百万实体书训练Claude，1.5亿美元和解案文件曝光</news:title>
                        <news:keywords>Anthropic, AI训练数据, 版权争议</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/microsoft-ai-rivalry-openai-anthropic</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T09:23:45+08:00</news:publication_date>
            <news:title>微软正面挑战OpenAI和Anthropic，自研AI军团全面出击</news:title>
                        <news:keywords>微软, 人工智能, OpenAI, Anthropic, AI竞争</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/meta-zuckerberg-enterprise-ai-opportunity</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T08:25:04+08:00</news:publication_date>
            <news:title>扎克伯格：Meta企业AI机遇超越智能体</news:title>
                        <news:keywords>Meta企业AI, 智能体, 开源模型, 企业级API, AI算力</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/microsoft-anthropic-openai-investment-results</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T08:24:28+08:00</news:publication_date>
            <news:title>微软从Anthropic投资获32亿美元收益，OpenAI喜忧参半</news:title>
                        <news:keywords>微软, Anthropic, OpenAI, AI投资, 财报</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/meta-ai-agents-billions-in-five-years</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T08:23:40+08:00</news:publication_date>
            <news:title>扎克伯格预言：五年内数十亿人拥有AI个人代理</news:title>
                        <news:keywords>扎克伯格, AI代理, Meta, 人工智能, 未来预测</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/lilian-weng-joins-openai-after-leaving-thinking-machines</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T06:24:36+08:00</news:publication_date>
            <news:title>Lilian Weng因健康离开Thinking Machines后加入OpenAI</news:title>
                        <news:keywords>Lilian Weng, OpenAI, AI安全, 人才流动, 创业健康</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/ai-future-saas-reckoning-agent-security-gap</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T06:24:04+08:00</news:publication_date>
            <news:title>AI未来走向：SaaS清算与代理安全鸿沟，TechCrunch Disrupt 2026揭晓</news:title>
                        <news:keywords>人工智能, SaaS, AI安全, 代理安全, TechCrunch</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/mythos-uncovers-crypto-weaknesses</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T06:23:33+08:00</news:publication_date>
            <news:title>Mythos揭晓：潜藏多年的加密货币核心漏洞</news:title>
                        <news:keywords>加密货币, 安全漏洞, Anthropic, Mythos, 密码学</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/anthropic-petition-pace-frontier-ai-development</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T06:13:51+08:00</news:publication_date>
            <news:title>Anthropic CEO签署请愿呼吁放缓前沿AI研发 员工超千人联署引发争议</news:title>
                        <news:keywords>AI安全, 前沿模型监管, 行业政策</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/nvidia-open-secure-ai-alliance-40-companies</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T06:11:34+08:00</news:publication_date>
            <news:title>NVIDIA联合约40家公司成立Open Secure AI Alliance 聚焦开源AI安全工具</news:title>
                        <news:keywords>NVIDIA, AI安全, 开源联盟, 代理攻击</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/frontier-ai-jailbreak-easy</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T04:25:49+08:00</news:publication_date>
            <news:title>前沿AI模型越狱竟如此简单</news:title>
                        <news:keywords>AI安全, 越狱攻击, 大模型, 安全防护, 对抗性测试</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/claude-opus-5-vending-machine-simulation</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T04:25:16+08:00</news:publication_date>
            <news:title>Claude Opus 5运营自动售货机：为了利润不择手段</news:title>
                        <news:keywords>AI伦理, 大语言模型, 模拟实验, 人工智能安全, Claude Opus 5</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/hugging-face-ai-breach-bear-metaphor</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T04:24:24+08:00</news:publication_date>
            <news:title>熊闯营地：Hugging Face AI安全事件的另类解读</news:title>
                        <news:keywords>Hugging Face, AI安全, 模型投毒, 网络安全, 开源生态</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.yingzheng.com/article/us-ban-foreign-robots-impact</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下 AI</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-07-30T04:23:47+08:00</news:publication_date>
            <news:title>美国禁外国机器人：赢家输家几何？</news:title>
                        <news:keywords>机器人禁令, 美国政策, 产业影响, 技术竞争, 全球化退潮</news:keywords>
                    </news:news>
    </url>
</urlset>
