VOL.2026.05.06·81 篇·每日精选
AI 日报
AI DAILY二〇二六年五月六日 星期三每日自动更新
01AI 动态今日动态79 篇
马斯克曾希望 OpenAI 走商业化道路，Greg Brockman 作证
Google AI NewsElon Musk Wanted OpenAI to Go Commercial, Greg Brockman Testifies
ChatGPT免费模型升级：幻觉减半/记忆增强/回答更精简
量子位刚刚，ChatGPT免费模型升级了：幻觉砍半/记忆更强/回答更简洁
首次使用裸盖菇素后的人脑变化
Hacker NewsHuman brain changes after first psilocybin use
Xbox CEO 结束 Copilot AI 开发并改组领导层
Hacker NewsXbox CEO ends Copilot AI development and overhauls leadership
美国大使 Burch 表示美国‘欢迎’教廷对 AI 监管的意见
Google AI NewsUS ‘Welcomes’ Vatican Input on AI Regulation, Says Ambassador Burch
青少年 AI 安全研究
Google AI NewsResearch into youth Ai safety
为什么只有富家子弟能在音乐界取得成功
Hacker NewsWhy Only Rich Kids Make It in Music Today
出版商称扎克伯格亲自授权 Meta 的版权侵权行为
Hacker NewsZuckerberg 'personally authorized' Meta's copyright infringement, publishers say
电信工人呼吁对人工智能使用进行限制
Google AI NewsTelecom workers call for restrictions over use of artificial intelligence
Common Sense Media 创建独立监督机构测试儿童 AI 安全
Google AI NewsCommon Sense Media creates independent watchdog to test AI safety for kids
OpenAI 推出 ChatGPT 5.5 Instant 作为新默认模型
Google AI NewsOpenAI rolls out ChatGPT 5.5 Instant as the new default model for everyone
X用户诱骗Grok向其发送20万美元
Hacker NewsX user tricks Grok into sending them $200k
广泛使用的Daemon Tools磁盘应用在长达一个月的供应链攻击中被植入后门
Hacker NewsWidely used Daemon Tools disk app backdoored in monthlong supply-chain attack
美国将对谷歌、微软和 xAI 的新 AI 模型进行安全测试
Google AI NewsUS to safety test new AI models from Google, Microsoft, xAI
IBM CEO Arvind Krishna 警告华盛顿必须在 AI 监管上找到‘金发姑娘’般的中间立场
Google AI NewsIBM CEO Arvind Krishna warns Washington must find 'Goldilocks' middle ground on AI regulation
AI 政策应关注贡献而非仅关注保护
Google AI NewsMake AI policy about contribution, not just protection
宾夕法尼亚州众议院委员会通过 AI 医疗监管法案
Google AI NewsAI health care regulation bill advances out of Pennsylvania House committee
埃塞俄比亚人工智能研究院在美国 SelectUSA 投资峰会上展示 AI 能力
Google AI NewsEthiopian Artificial Intelligence Institute showcases AI capabilities at SelectUSA Investment Summit in the United States
马斯克诉 OpenAI 案在 AI 控制权辩论中产生的争议多于启示
Google AI NewsThe Elon Musk-OpenAI trial is producing more heat than light in the debate over who should control AI
草中之工具：培养下一代工匠
Hacker NewsTools in the Grass: Raising the next generation of crafts person
AI 监管：全球规则制定竞赛中美国、欧盟和中国领先
Google AI NewsAI regulation: Global race to shape rules as US, EU, China take lead
OpenAI 推出 ChatGPT 5.5 Instant 作为新默认模型
Google AI NewsAnthropic CEO Dario Amodei warns some software companies will 'completely go bust'
加州果农在 Del Monte 破产后将销毁 42 万棵桃树
Hacker NewsCalifornia farmers to destroy 420k peach trees following Del Monte bankruptcy
Show HN：探索灵感源自 3000 件大师绘画作品的配色方案
Hacker NewsShow HN: Explore color palettes inspired by 3000 master painter artworks
SubQ: 为12M-token上下文构建的次二次方LLM
Hacker NewsSubQ: Sub-quadratic LLM built for 12M-token context
扎克伯格‘亲自授权并鼓励’Meta 侵犯版权
Hacker NewsZuckerberg 'Personally Authorized and Encouraged' Meta's Copyright Infringement
OpenAI 的新测量工具和广告管理器能否证明其作为广告渠道的价值？
Google AI NewsWill OpenAI’s New Measurement Tools And Ads Manager Prove Its Worth As An Ad Channel?
GLM-5V-Turbo：迈向多模态智能体的原生基础模型
Hacker NewsGLM-5V-Turbo: Toward a Native Foundation Model for Multimodal Agents
Anthropic CEO 警告网络'危险时刻'，称 AI 暴露了数千个安全漏洞
Google AI NewsAnthropic CEO warns of cyber ‘moment of danger’ as AI exposes thousands of vulnerabilities
RadixArk 融资一亿美元以缓解计算瓶颈
Google AI NewsRadixArk raises $100M to ease computing bottlenecks
Anthropic 加速金融领域布局，CEO Amodei 警告软件颠覆风险
Google AI NewsAnthropic deepens finance push as CEO Amodei warns of software disruption
Anthropic 推出面向金融服务公司的新型 AI 智能体
Google AI NewsAnthropic Releases New AI Agents for Financial Services Firms
Amazon CEO Andy Jassy 的一则更新，每位人工智能投资者都应关注
Google AI News1 Update Amazon CEO Andy Jassy Just Said That Every Artificial Intelligence (AI) Investor Should Hear
为什么航空公司总是濒临破产
Hacker NewsWhy airlines are always going bankrupt
GPT-5.5 Instant
Hacker NewsGPT‑5.5 Instant
Tessera Labs完成A轮融资
Google AI NewsTessera Labs Raises Series A
Tessera Labs完成A轮融资 StartupHub.ai
像对待开发人员一样对待你的编程代理
Hacker NewsTreat your coding agents like developers
消息称 OpenAI、Anthropic 正洽谈收购 AI 服务公司
Google AI NewsOpenAI, Anthropic ventures in talks to buy AI services firms, sources say
加速 Gemma 4：使用多令牌预测草稿器实现更快的推理
Hacker NewsAccelerating Gemma 4: faster inference with multi-token prediction drafters
Mark 和 Mary Stevens 捐赠 2 亿美元助力 USC 全方位推进 AI 研究
Google AI NewsMark and Mary Stevens give $200 million to power AI research across USC
Coinbase今日裁员：加密巨头裁减数百个工作岗位，CEO 称 AI 正在'改变我们的工作方式'
Google AI NewsCoinbase layoffs today: Crypto giant cuts hundreds of jobs as CEO says AI is 'changing how we work'
南非的 AI 政策遭到 AI 本身的破坏
Google AI NewsSouth Africa’s AI Policy Got Sabotaged by AI Itself
AI 的三大逆定律
Hacker NewsThree Inverse Laws of AI
科技：商务部扩大 AI 安全测试
Google AI NewsTech: Commerce expands AI safety testing
Show HN: 一个用于自动剥离K8s日志中PII的变异Webhook
Hacker NewsShow HN: A Mutating Webhook to automatically strip PII from K8s logs
金融服务业和保险业的智能体
Hacker NewsAgents for financial services and insurance
AI 在律师-客户关系和法律实践中的演进
Google AI NewsThe evolution of AI in lawyer-client relations and legal practice
Show HN：Airbyte Agents——为跨多个数据源的智能体提供上下文
Hacker NewsShow HN: Airbyte Agents – context for agents across multiple data sources
Anthropic 随着与 OpenAI 竞争升级推出新型 Claude 金融智能体
Google AI NewsAnthropic Rolls Out New Claude Financial Agents as OpenAI Rivalry Escalates
Anthropic 通过新型 AI 智能体、完整 Microsoft 365 集成和与 Moody's 的数据合作深化对华尔街的布局
Google AI NewsAnthropic deepens push into Wall Street with new AI agents, full Microsoft 365 integration, Moody's data partnership
大气微塑料可能正在导致全球变暖
Hacker NewsAirborne Microplastics May Be Warming the Planet
联邦安全网络尚未为人工智能做好准备
Google AI NewsThe Federal Safety Net Isn’t Ready for Artificial Intelligence
SubQ: 具备12M-token上下文的次二次方LLM
Hacker NewsSubQ: a sub-quadratic LLM with 12M-token context
新融资可能使Anthropic成为全球最具价值的AI初创公司
Google AI NewsFresh Funding Could Make Anthropic The World’s Most Valuable AI Startup
南加州大学启动变革性 AI 倡议，获大学历史上最大额捐赠之一
Google AI NewsUSC launches transformational AI initiative with one of the largest gifts in university history
SAP斥资10亿欧元收购AI创企Prior Labs，解锁企业结构化数据
Google AI NewsSAP Acquires AI Startup Prior Labs With €1B Investment to Unlock Enterprise Structured Data
不是 AI 删除了你的数据库，是你自己删的
Hacker NewsAI didn't delete your database, you did
AI 产品墓地
Hacker NewsAI Product Graveyard
Cyborg 与 Austin Artificial Intelligence 合作，在生产中提供端到端安全 AI
Google AI NewsCyborg Partners with Austin Artificial Intelligence to Deliver End-to-End Secure AI in Production
估值14亿美元的AI初创公司Blitzy挑战Claude Code和Codex
Google AI News$1.4 Billion AI Startup Blitzy Is Taking On Claude Code and Codex
Joe Lonsdale 谈 AI 监管：国家审查应尽可能有限且有针对性
Google AI NewsJoe Lonsdale on AI regulation: National review should be as limited and targeted as possible
CAISI 与 Google DeepMind、Microsoft 和 xAI 签署关于前沿 AI 国家安全测试的协议 - 美国国家标准与技术研究院
Google DeepMindCAISI Signs Agreements Regarding Frontier AI National Security Testing With Google DeepMind, Microsoft and xAI - National Institute of Standards and Technology (.gov)
CAISI 与 Google DeepMind、Microsoft 和 xAI 签署前沿 AI 国家安全测试协议
Google AI NewsCAISI Signs Agreements Regarding Frontier AI National Security Testing With Google DeepMind, Microsoft and xAI
SAP 向德国 AI 初创公司 Prior Labs 投资十亿欧元
Google AI NewsSAP bets €1B on German AI startup Prior Labs, which raised only €9M pre-seed
ZyG在推出仅一年后以5亿美元估值完成6000万美元A轮融资
Google AI NewsZyG raises $60 million Series A at $500 million valuation just a year after launch
数十亿美元的 AI 融资轮推动四月成为一年中初创公司融资第三高的月份
Google AI NewsBillion-Dollar AI Rounds Push April To Third-Highest Startup Funding Month In A Year
调查：联邦资金、可负担性和AI是初创生态系统面临的主要挑战
Google AI NewsSurvey: Federal funding, affordability and AI are startup ecosystems’ main challenges
前ironSource高管为AI创企Zyg获得5亿美元估值融资，在展示公开表现前下注，这完全取决于谁在构建它
Google AI NewsFormer ironSource Executives Raised at a $500 Million Valuation for AI Startup Zyg Before Showing Public Traction and That Bet Is Entirely About Who Is Building It
Microsoft、Google 推动 AI 智能体治理进入企业 IT 主流
Google AI NewsMicrosoft, Google push AI agent governance into enterprise IT mainstream
当每个人都有 AI，而公司仍然学不到任何东西时
Hacker NewsWhen everyone has AI and the company still learns nothing
律师透露如何在4天内为其AI原生律师事务所筹集900万美元
Google AI NewsA lawyer reveals how he snagged $9 million in 4 days for his AI-native law firm
Google Chrome 在未经同意的情况下静默在您的设备上安装了一个 4GB 的 AI 模型
Hacker NewsGoogle Chrome silently installs a 4 GB AI model on your device without consent
智能体编程的经验教训：当代码变得廉价时我们应该做什么
Hacker NewsLessons for Agentic Coding: What should we do when code is cheap?
一个完整的 Llama2 推理引擎，可放入 1356 字节的 x86 汇编代码中
Hacker NewsA complete Llama2 inference engine that fits in 1356 bytes of x86 assembly
因Meta收购紧张局势，中国阻止美国对AI创企投资
Google AI NewsChina blocks US funding for AI startups amid Meta acquisition tensions
从零开始训练你自己的 LLM
Hacker NewsTrain Your Own LLM from Scratch
Think 2026：随着 AI 差距扩大，IBM 提供 AI 运营模型蓝图
Google AI NewsThink 2026: IBM Delivers the Blueprint for the AI Operating Model as the AI Divide Widens
CVE-2026-31431：Copy Fail 对比无根容器
Hacker NewsCVE-2026-31431: Copy Fail vs. rootless containers
Monarrch 在纳什维尔活动上推出新型 AI 监管技术
Google AI NewsMonarrch unveils new AI regulation technology at Nashville event
03人力资源动态职场与人才2 篇