<?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.winzheng.com/article/spotify-ai-persona-music-policy</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T22:23:52+08:00</news:publication_date>
            <news:title>Spotify为AI音乐打“人设”标签，默认不推荐</news:title>
                        <news:keywords>Spotify, AI音乐, 流媒体, 版权保护, 算法推荐</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-zoom-screen-sharing-bug</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T21:23:49+08:00</news:publication_date>
            <news:title>AI仅用20个提示发现Zoom通话劫持漏洞</news:title>
                        <news:keywords>AI安全, Zoom漏洞, 远程代码执行, 屏幕共享, 人工智能攻防</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/llm-next-big-thing-ai-research</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T20:24:14+08:00</news:publication_date>
            <news:title>大模型的下一个风口与AI学术研究转向</news:title>
                        <news:keywords>大语言模型, AI研究, Transformer, AI趋势</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/anthropic-text-watermark-coverage</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T20:23:38+08:00</news:publication_date>
            <news:title>Anthropic宣布为AI生成文本添加水印，覆盖旧版模型</news:title>
                        <news:keywords>AI安全, 内容溯源, 水印技术, Anthropic</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/sanders-letter-ai-pause-openai-anthropic-meta</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T20:21:29+08:00</news:publication_date>
            <news:title>桑德斯致函OpenAI等要求暂停前沿AI开发 安全派发展派观点对立</news:title>
                        <news:keywords>AI监管, 前沿模型, 政策争议</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/revealing-ai-models-inner-thoughts</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T19:23:32+08:00</news:publication_date>
            <news:title>新招数揭示AI模型“内心戏”</news:title>
                        <news:keywords>AI推理轨迹, 语言模型, 中国AI, 模型训练, 知识产权</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/novo-nordisk-aws-agentic-ai</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T18:24:39+08:00</news:publication_date>
            <news:title>诺和诺德携手AWS，用代理式AI加速药物发现</news:title>
                        <news:keywords>AI药物发现, 诺和诺德, AWS, 代理式AI</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/organoids-over-neural-networks</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T18:23:54+08:00</news:publication_date>
            <news:title>AI已死，类器官当立：迷你人脑将超越神经网络？</news:title>
                        <news:keywords>类器官, 人工智能, 生物计算, 脑机接口, 科技伦理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-schizophrenia-genetic-puzzle</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T17:23:50+08:00</news:publication_date>
            <news:title>AI助力破解精神分裂症复杂遗传谜题</news:title>
                        <news:keywords>AI医疗, 精神分裂症, 基因组学, 深度学习, 精准医疗</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-changing-vulnerability-response-timeline</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T15:23:42+08:00</news:publication_date>
            <news:title>AI如何重塑漏洞响应时间线：从被动修复到主动防御</news:title>
                        <news:keywords>AI安全, 漏洞响应, 零日漏洞, 容器安全, 人工智能</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/nvidia-500-billion-ai-compute-financing-platform</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T14:20:54+08:00</news:publication_date>
            <news:title>英伟达联手六家机构筹5000亿美元AI算力融资 股价下跌引发牛熊辩论</news:title>
                        <news:keywords>AI算力融资, Nvidia, 基础设施投资, 资本市场, 算力金融化</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/openai-70-billion-employee-share-repurchase-valuation-852-bi</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T14:18:31+08:00</news:publication_date>
            <news:title>OpenAI完成70亿美元员工股份回购 估值维持8520亿美元未变</news:title>
                        <news:keywords>OpenAI, 股权回购, AI估值</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/meta-open-ai-model-strategy</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T08:25:18+08:00</news:publication_date>
            <news:title>Meta押注开源新模型，能否翻身AI落后局面？</news:title>
                        <news:keywords>AI开源, Meta, 大模型, 扎克伯格, AI竞赛</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/openai-daybreak-cyber-model</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T08:24:28+08:00</news:publication_date>
            <news:title>AI攻击激增，OpenAI推出新型网络防御模型</news:title>
                        <news:keywords>OpenAI, 网络安全, AI防御, 大模型, Daybreak计划</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/openai-employee-tender-offer</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T08:23:41+08:00</news:publication_date>
            <news:title>OpenAI员工套现70亿美元，旧金山房市再陷困境</news:title>
                        <news:keywords>OpenAI, 员工要约收购, 旧金山房市, 科技行业, AI经济</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/amazon-off-grid-data-center-pollution</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T06:24:02+08:00</news:publication_date>
            <news:title>亚马逊押注离网电厂，或成美国气候污染最大源头</news:title>
                        <news:keywords>亚马逊, 数据中心, 气候污染, AI能源, 离网电力</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/zuckerberg-ai-manifesto-backlash</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T06:23:35+08:00</news:publication_date>
            <news:title>扎克伯格AI宣言，正是人们讨厌AI的原因</news:title>
                        <news:keywords>扎克伯格, 人工智能, Meta, 科技评论, AI伦理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/anthropic-unreleased-claude-riemann-hypothesis-bound-41-6-to</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T06:11:13+08:00</news:publication_date>
            <news:title>Anthropic未发布Claude将黎曼ζ函数零点下限从41.6%提升至67.2%</news:title>
                        <news:keywords>AI数学能力, 黎曼假设, Anthropic Claude, 纯数学突破</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/nvidia-six-firms-500b-ai-infra-funding</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T06:08:49+08:00</news:publication_date>
            <news:title>英伟达联手六机构筹5000亿美元AI算力资金</news:title>
                        <news:keywords>英伟达, AI基础设施, 算力融资</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-professors-academic-research</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T04:24:36+08:00</news:publication_date>
            <news:title>AI教授面临新现实：学术研究如何突围？</news:title>
                        <news:keywords>AI研究, 学术生态, 人才流动, 产业界, AI伦理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/claude-agent-hacks-gym-waitlist</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T04:23:50+08:00</news:publication_date>
            <news:title>Claude智能体黑进健身房系统，科技圈沸腾</news:title>
                        <news:keywords>AI智能体, Claude, 网络安全, AI伦理, 科技新闻</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/muse-glimmer-personal-superintelligence-vision</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-11T02:23:34+08:00</news:publication_date>
            <news:title>Meta新模型Muse Glimmer预示扎克伯格个人超级智能愿景</news:title>
                        <news:keywords>AI模型, 开放权重, Meta, 个人超级智能, AI所有权</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-science-reasoning-censorship</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T21:23:44+08:00</news:publication_date>
            <news:title>AI科研需推理而非数据，审查产业引担忧</news:title>
                        <news:keywords>AI科研, 推理能力, 内容审核, 科技伦理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-slop-backlash-platforms</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T20:24:30+08:00</news:publication_date>
            <news:title>AI垃圾内容引发反噬，平台终于开始整顿</news:title>
                        <news:keywords>AI治理, 内容审核, 人工智能伦理, 社交媒体, 深度伪造</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-chip-material-discovery-funding</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T20:23:38+08:00</news:publication_date>
            <news:title>AI“打地鼠”式寻材：融资900万美元打造更“冷”芯片</news:title>
                        <news:keywords>AI材料发现, 芯片散热, 半导体材料, 科技融资</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/us-house-refuses-review-openai-anthropic-ai-containment</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T20:16:36+08:00</news:publication_date>
            <news:title>美国众议院决定不就OpenAI和Anthropic AI模型封锁问题展开正式审查</news:title>
                        <news:keywords>AI监管, OpenAI, Anthropic, 国会政策</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/meta-releases-muse-glimmer-30b-local-agent-model</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T20:13:54+08:00</news:publication_date>
            <news:title>Meta开源30B参数Muse Glimmer 单GPU本地运行直面OpenAI封闭路线</news:title>
                        <news:keywords>Meta, 开源模型, AI代理</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/rise-of-ai-job-interview</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T19:24:45+08:00</news:publication_date>
            <news:title>AI面试崛起：凌晨1点的求职新常态</news:title>
                        <news:keywords>AI面试, 人工智能, 招聘变革, 求职体验</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/peer-review-ai-era-survival</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T19:24:19+08:00</news:publication_date>
            <news:title>AI时代同行评审不堪重负，学术诚信何去何从</news:title>
                        <news:keywords>同行评审, AI学术, 科研诚信, 学术出版, ChatGPT</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/meta-muse-glimmer-local-ai-agents</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T19:23:37+08:00</news:publication_date>
            <news:title>Meta Muse Glimmer：消费级GPU跑本地AI代理</news:title>
                        <news:keywords>Meta, Muse Glimmer, 开源AI, 本地智能体, 消费级GPU</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/simens-physics-ai-human-in-charge</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T18:23:42+08:00</news:publication_date>
            <news:title>物理AI虽快千倍，西门子称安全签核仍需人类</news:title>
                        <news:keywords>物理AI, 西门子, 工业智能, 安全监管, 工程仿真</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/startups-chasing-next-big-thing-llms</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T17:24:20+08:00</news:publication_date>
            <news:title>追寻LLM下一个重大突破的初创公司</news:title>
                        <news:keywords>大语言模型, 初创公司, AI架构, 注意力机制, Transformer</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-for-science-reasoning</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T17:23:50+08:00</news:publication_date>
            <news:title>AI驱动科学，重在推理而非数据</news:title>
                        <news:keywords>AI科学, 科学推理, 人工智能, 数据驱动, 人机协作</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/meta-muse-spark-1-1-security-breach-sandbox-test</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T14:17:42+08:00</news:publication_date>
            <news:title>Meta Muse Spark 1.1安全测试突破外部系统 成第三家披露失败的AI实验室</news:title>
                        <news:keywords>AI安全, Meta, Muse Spark 1.1</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/google-deepmind-leadership-restructure-hassabis-dean-discove</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T14:15:21+08:00</news:publication_date>
            <news:title>谷歌DeepMind领导层重组 哈萨比斯卸日常运营 Jeff Dean三人离职创业</news:title>
                        <news:keywords>谷歌, DeepMind, 人才流动, AI商业化</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/hedge-fund-invests-chip-startup-source-foundry</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T06:23:54+08:00</news:publication_date>
            <news:title>困境中对冲基金豪掷4亿美元，押注芯片初创Source Foundry</news:title>
                        <news:keywords>AI芯片, 对冲基金, 半导体投资, 科技融资</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/openai-astra-paper-academic-misconduct-allegations</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T06:11:06+08:00</news:publication_date>
            <news:title>OpenAI Astra模型249页论文遭学术不端指控 引用缺失引争议</news:title>
                        <news:keywords>OpenAI, Astra, 学术不端, 数学研究</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/aisi-report-ai-agents-deception-openai-anthropic</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T06:07:56+08:00</news:publication_date>
            <news:title>AISI测试现19次越界 OpenAI与Anthropic代理涉真实入侵</news:title>
                        <news:keywords>AI安全, 代理测试, 越狱行为</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/anthropic-claude-code-auto-mode-default</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T04:23:48+08:00</news:publication_date>
            <news:title>Anthropic默认开启Claude Code自动模式，AI编程再少人工干预</news:title>
                        <news:keywords>Anthropic, Claude Code, AI编程, 自动模式, 开发者工具</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/google-tpu-v7-post-transformer-uncertainty</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T01:35:22+08:00</news:publication_date>
            <news:title>Google TPU v7 宣称60%能效提升 后Transformer架构优化未经独立核实</news:title>
                        <news:keywords>TPU, Google, AI硬件</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/openai-pauses-astra-development-security-vs-progress-conflic</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T01:35:02+08:00</news:publication_date>
            <news:title>OpenAI暂停Astra部分开发 安全优先与技术进展冲突升级</news:title>
                        <news:keywords>OpenAI, Astra, AI安全</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/ai-safety-testing-risk</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T00:24:18+08:00</news:publication_date>
            <news:title>AI安全测试反成安全风险</news:title>
                        <news:keywords>AI安全, 智能体安全, 安全测试, AI监管</news:keywords>
                    </news:news>
    </url>
    <url>
        <loc>https://www.winzheng.com/article/jill-lepore-silicon-valley-science-fiction-democracy</loc>
        <news:news>
            <news:publication>
                <news:name>赢政天下</news:name>
                <news:language>zh</news:language>
            </news:publication>
            <news:publication_date>2026-08-10T00:23:43+08:00</news:publication_date>
            <news:title>历史学家莱波雷：硅谷误读科幻，正侵蚀民主</news:title>
                        <news:keywords>科幻误读, 硅谷文化, AI治理, 民主, 科技伦理</news:keywords>
                    </news:news>
    </url>
</urlset>
