Greg Brockman@gdb · 5月22日34

Daybreak for helping cybersecurity at major banks:

译OpenAI的Daybreak项目为大型银行的网络安全防御带来曙光： [引用 @TheRealAdamG]：https://www.afr.com/companies/financial-services/major-banks-use-openai-s-daybreak-for-cybersecurity-defence-20260519-p5zyn9

Rohan Paul@rohanpaul_ai · 5月22日54

Bloomberg: The Pentagon is testing rival AI models after moving to replace Anthropic’s Claude inside military workflows. 25 power users are comparing models that could handle the department’s most demanding AI tasks. --- bloomberg. com/news/articles/2026-05-21/pentagon-tests-rival-ai-models-in-race-to-replace-anthropic

译彭博社：五角大楼在军事工作流程中替换Anthropic的Claude后，正在测试竞品AI模型。 25名核心用户正在比较可能处理该部门最苛刻AI任务的模型。

Rohan Paul@rohanpaul_ai · 5月22日82

California’s Governor Newsom signs first-of-its-kind executive order to prepare workers and businesses for potential AI disruption. The order treats AI job loss as a policy problem, not just a company problem, and asks agencies to study severance, subsidized jobs, employment insurance, stock compensation, worker ownership, and faster warning signs before layoffs spread. The technical issue is that AI can replace parts of a job before replacing the whole job, so payroll, hiring, and task data may show damage earlier than official unemployment numbers. California wants a new AI labor dashboard, updated WARN Act rules, and business feedback in monthly jobs reports so the state can see which sectors are being hollowed out first. The deeper shift is about who gets the productivity gain when software does more work: companies can keep it as margin, or policy can push some gains toward workers through equity, ownership, training, or wage support. The hard part is that training alone may not solve displacement if AI removes demand for the roles people are being trained to enter. --- gov. ca .gov/2026/05/21/governor-newsom-signs-first-of-its-kind-executive-order-to-prepare-workers-and-businesses-for-potential-ai-disruption/

译加州州长纽森签署行政令，首次将AI引发的失业问题正式列为公共政策议题，要求各部门研究遣散费、就业保险及员工持股等保障措施。该命令认识到AI可能渐进式替代岗位任务，而非一次性取代整个职位，因此计划建立新的劳动力数据看板，以更早捕捉行业受到的冲击。政策核心在于探讨如何将AI带来的生产力红利，通过股权、薪酬支持等方式在企业与员工间进行更公平的分配。同时指出，单纯的职业培训可能无法解决被AI彻底取代的岗位所带来的根本性就业错配问题。

Epoch AI@EpochAIResearch · 5月22日63

For frontier AI chips, memory is the largest and fastest-growing component cost. High-bandwidth memory (HBM) has grown from 52% to 63% of total AI chip component spending between Q1 2024 and Q4 2025.

译对于前沿AI芯片而言，内存是最大且增长最快的组件成本。 高带宽内存（HBM）在AI芯片组件总支出中的占比，已从2024年第一季度的52%增长至2025年第四季度的63%。

AI Notkilleveryoneism Memes ⏸️@AISafetyMemes · 5月22日73

Not only are AIs suddenly solving the hardest math problems in history, they're winning fiction writing awards "AIs can't write" is obvious cope, sorry. The biggest tell of AI writing (besides a few overused ticks) is that it's TOO well-written, too articulate

译AI不仅能解决史上最难数学题，还在小说写作比赛中获奖 “AI不会写作”显然是自欺欺人，抱歉。 AI写作最大的破绽（除了几个过度使用的习惯用语）是它写得**太好**，表达**太清晰**。

Rohan Paul@rohanpaul_ai · 5月22日65

The Information: Anthropic is currently in early-stage talks to lease and deploy Microsoft's custom AI chips for inference workloads. Microsoft is pitching Maia 200 as a cheaper way to run some AI inference, and claims maia 200 is more cost-effective than nvidia chips for certain inference jobs. Maia 200 is Microsoft’s second-generation AI accelerator, built on TSMC 3nm, with FP8/FP4 math, 216GB HBM3e, 7TB/s bandwidth, and 272MB SRAM, which makes it aimed at feeding large models fast rather than teaching them from scratch. Anthropic already committed $30B to Azure, Microsoft may invest up to $5B in Anthropic, and Claude is already tied into Microsoft’s Copilot stack, so the chip talks are also a customer-supplier feedback loop. IMO, Maia does not need to beat Nvidia everywhere to matter, because a cheaper chip for narrow, high-volume inference jobs can still shift billions of tokens away from GPUs. --- theinformation .com/articles/anthropic-talks-use-microsofts-ai-chips

译据The Information报道，微软正向AI公司Anthropic推销其第二代AI芯片Maia 200，强调该芯片在特定推理任务中比NVIDIA芯片更具成本效益。Maia 200专注于高速推理而非训练，双方已有深度合作基础：Anthropic已在Azure承诺300亿美元支出，且Claude已整合进微软Copilot。此次芯片合作旨在深化协同。分析认为，Maia 200无需全面超越NVIDIA，只要能在高量推理中提供更低成本选项，便可能将部分计算需求从GPU转移。

Rohan Paul@rohanpaul_ai · 5月22日68

WSJ: Anthropic is now projected to hit its first operating profit in Q2, with revenue jumping 130% to $10.9B after expecting profitability only around 2028. The shock is not only the profit, but the speed at which enterprise AI spending is converting into real revenue. Anthropic made $4.8B in Q1, then told investors it expects $559M in operating profit in the June quarter. The driver is agentic coding, where Claude does longer software tasks instead of only answering one prompt at a time. That changes the product from a chatbot into rented digital labor, so customers can justify much bigger bills. The cost story also changed, because Anthropic spent 71 cents of every revenue dollar on compute in Q1 but expects 56 cents in Q2. Compute is still the main tax on AI, but lower compute per dollar means scale is starting to work in Anthropic’s favor. Anthropic also benefits from using more Google and Amazon chips, a smaller free consumer base, and fewer giant data-center promises than OpenAI. --- wsj. com/tech/ai/mind-blowing-growth-is-about-to-propel-anthropic-into-its-first-profitable-quarter-7edbf2f4

译据《华尔街日报》报道，Anthropic预计将大幅提前盈利，计划于2026年第二季度实现首次运营利润，而非此前预计的2028年。该公司收入预计同比增长130%，达到109亿美元，二季度运营利润可达5.59亿美元。核心驱动力是“代理式编程”，使Claude能处理更长的软件任务，提供可租赁的数字劳动力，从而显著提升客户价值。同时，成本结构持续改善，计算成本占收入比预计从一季度的71%降至56%，规模效应开始显现。此外，更广泛地使用谷歌和亚马逊芯片、较小的免费用户群体以及更少的超大型数据中心投资承诺，也为其带来了竞争优势。

Rohan Paul@rohanpaul_ai · 5月22日59

President Trump postponed the planned AI executive order because he feared parts of it could slow US AI companies while China is racing to catch up. "We're leading everybody, and I don't want to do anything that's going to get in ​the way of that lead"

译特朗普总统推迟了计划中的AI行政令，因为他担心其中部分内容可能会拖慢美国AI公司的步伐，而中国正在奋力追赶。 “我们领先于所有人，我不想做任何会阻碍这种领先的事情。”

Krea@krea_ai · 5月22日26

this happening in 1h! join us to chat about LoRAs and new updates coming to Krea 2.

译1小时后开始！ 加入我们，一起聊聊LoRAs和Krea 2即将到来的新更新。 [引用 @krea_ai]：我们明天将举办首次社区空间活动！ 通过X Spaces加入我们，讨论Krea 2的更新、分享反馈或展示你的作品。 链接在下方 👇

宝玉@dotey · 5月22日20

之前 ClawHub 名称被占用的问题，官方已经解决了，非常感谢 @chenjingtalk @pat_erichsen @steipete 的支持🙏

Rohan Paul@rohanpaul_ai · 5月21日72

Standard Chartered just made AI job replacement an official banking strategy. Will replace "lower-value human capital" with AI, with 7,500+ jobs at risk as it cuts more than 15% of back-office roles by 2030. The target is not the bank teller or the client banker first, but the large internal machine that checks documents, routes approvals, handles compliance files, reconciles records, and keeps banking systems moving. Banks are especially exposed because much of their back office is structured language work, where LLMs can read forms, compare policies, flag missing data, summarize cases, and pass clean tasks to humans. --- news. sky. com/story/standard-chartered-to-replace-lower-value-human-capital-with-ai-13545809?dcmp=snt-sf-twitter

译渣打银行正式将AI工作替代列为官方战略，计划到2030年通过AI自动化超过15%的后台支持岗位，影响逾7500个职位。该银行的目标并非直接替代柜员或客户经理，而是首先针对内部处理文档、审批流程、合规档案及系统运营的庞大支持机器。银行业因其后台工作多为结构化语言任务而尤为脆弱，大语言模型能够高效处理表单阅读、政策比对、数据校验及案例总结等工作，从而接管大量此类规则化流程，实现运营自动化。

Rohan Paul@rohanpaul_ai · 5月21日80

WSJ: OpenAI is preparing a confidential IPO filing in the coming weeks. Goldman Sachs and Morgan Stanley are helping draft the prospectus. OpenAI was recently valued at $ 852B. The Musk legal win removed one IPO roadblock, but public investors will still test OpenAI’s structure, Microsoft relationship, spending commitments. Investors continue to scrutinize OpenAI’s financial position and that of its partners, while Altman keeps announcing its web of tech partnerships. --- wsj .com/tech/ai/openai-ipo-filing-date-0ec95af5

译OpenAI正准备在数周内秘密提交IPO申请，由高盛和摩根士丹利协助起草招股书，此前估值高达852亿美元。尽管马斯克的法律诉讼胜利消除了一重障碍，但公众投资者仍将严格审视OpenAI的复杂公司治理结构、与微软的深度关系以及其高昂的支出承诺。市场持续关注OpenAI及其合作伙伴的财务状况。

Alibaba Cloud@alibaba_cloud · 5月21日56

Alibaba Cloud has been named a Visionary in the 2026 Gartner® Magic Quadrant™ for CPaaS🚀 Our AI-native, API-first platform covers SMS, Voice, Chat App, Intelligent Contact Center & more — serving 200+ countries across 29 regions. We're not adding AI on top. We're rebuilding communications from the ground up. #CPaaS #Gartner #AlibabaCloud #AI #CloudCommunications

译阿里云在2026年Gartner® CPaaS魔力象限™中被评为愿景者🚀 我们的AI原生、API优先平台覆盖短信、语音、聊天应用、智能联络中心等——服务于29个区域的200多个国家。 我们不是在现有基础上添加AI，而是从底层重构通信。 #CPaaS #Gartner #阿里云 #AI #云通信

Rohan Paul@rohanpaul_ai · 5月21日60

Every bank you talk to, AI is right at the top of their agenda. Blackstone President and COO Jon Gray made it very clear. Any rule-based businesses, like accounting, legal, finance, will be completely disrupted by AI.

译你与每家银行交谈，AI都位居其议程首位。 黑石集团总裁兼首席运营官乔恩·格雷明确指出。 任何基于规则的业务，如会计、法律、金融，都将被AI彻底颠覆。

Alibaba Cloud@alibaba_cloud · 5月21日36

After the Monaco Grand Prix 2026, the action continues. 🏁 Monaco doesn't just celebrate speed — it pioneers what's next. This June, the city hosts its first-ever AI Film Festival, a landmark moment where storytelling and artificial intelligence converge to reimagine cinema. Alibaba Cloud is thrilled to be part of it. At Alibaba Cloud, we believe the AI era belongs to creators as much as engineers. Our cloud and AI infrastructure — from foundation model training to real-time rendering at scale — is built to give filmmakers, artists, and visionaries the tools to experiment fearlessly and bring bold ideas to life, frame by frame. The future of film isn't just being imagined. It's being rendered. 📍 One Monte Carlo 🗓 June 9–10, 2026 🔗 https://int.alibabacloud.com/m/1000413073/ #AlibabaCloud #AIFilmFestival #GenerativeAI #FutureOfFilm #Monaco2026

译2026年6月，摩纳哥将举办首届AI电影节，标志着人工智能与电影叙事的深度融合。阿里云作为合作方，强调其云与AI基础设施（从基础模型训练到大规模实时渲染）致力于赋能创作者，为电影人、艺术家提供实验与实现大胆创意的工具，共同探索由AI技术驱动的电影制作新范式。

Rohan Paul@rohanpaul_ai · 5月21日48

StanChart is using AI to automate rule-based support work like compliance screening and customer data updates. plans to cut over 15% of support roles by 2030. CEO says this is about building a future-fit operation.

译渣打银行正在使用AI自动化合规筛查和客户数据更新等基于规则的支持工作。计划到2030年削减超过15%的支持岗位。 CEO表示，这是为了打造面向未来的运营模式。

Chubby♨️@kimmonismus · 5月21日78

Anthropic is paying SpaceX $1.25 billion per month for compute. Per month. That's $15 billion a year flowing to a company whose total annual revenue is $18 billion. One AI lab is about to account for the majority of SpaceX's commercial income. We only know this because SpaceX filed for an IPO today and had to disclose the terms. The deal was announced weeks ago with no financials attached. Source: Axios

译Anthropic每月向SpaceX支付12.5亿美元用于算力。每月。 这意味着每年有150亿美元流向一家年总收入为180亿美元的公司。一家AI实验室即将占据SpaceX商业收入的大部分。 我们之所以知道此事，是因为SpaceX今天提交了IPO申请，必须披露相关条款。该交易数周前已宣布，但未附带财务细节。来源：Axios

Greg Brockman@gdb · 5月21日19

ChatGPT team ships

译ChatGPT团队发布新功能

Elon Musk@elonmusk · 5月21日69

SpaceX is actively hiring world-class engineers/physicists for SpaceXAI, even if you have zero prior experience in AI. Smart humans figure it out fast. Please send an email with ~3 bullet points demonstrating evidence of exceptional ability to ai_eng@spacex.com.

译SpaceX正在为SpaceXAI积极招聘世界级工程师/物理学家，即使你此前没有任何AI经验。聪明的人能很快掌握。请发送邮件至ai_eng@spacex.com，附上约3个要点证明你的卓越能力。

Rohan Paul@rohanpaul_ai · 5月21日69

WOW, 🤯 A leaked audio from Meta’s April 30 all-hands. Meta is reportedly using its own engineers’ work traces to train coding AI while cutting thousands of jobs. Here Zuckerberg arguing that models learn better when they watch “really smart people” perform tasks, meaning Meta’s internal code, tool use, clicks, and problem-solving can become higher-grade training data than contractor-written examples. The idea is behavior cloning: instead of only feeding an AI finished code, Meta can feed it the step-by-step path a strong engineer takes, including edits, tests, mistakes, fixes, and tool choices. That can teach a model not just what correct code looks like, but how a skilled developer moves from a vague task to a working solution. Meta is reportedly cutting about 8,000 jobs, roughly 10% of its workforce, and additionaly moving about 7,000 employees toward AI-focused work, so the hard realy is that human expertise is being turned into training data before some of those humans leave. The story is not fully independently verified, but the shift is happening for sure: tech companies no longer see AI as a tool sitting beside workers, but as a system that can absorb worker patterns and then compress them into software.

译Meta正利用内部工程师的工作痕迹——如代码编写、工具使用和问题解决步骤——来训练其编程AI。CEO扎克伯格认为，让AI观察“聪明人”执行任务（行为克隆），比使用外部承包商代码样本更有效。同时，Meta正裁员约8000人，并计划让约7000名员工转向AI相关岗位。此举反映科技行业新趋势：公司正将人类专业知识直接转化为训练数据，AI不再只是工具，而是能吸收并压缩员工工作模式的系统。

Alibaba Cloud@alibaba_cloud · 5月21日31

Join us on May 26, 2026, at the Alibaba Cloud Claw Talks offline roadshow during Qwen Conference 2026, and discover the next wave of AI agent applications with us: https://click.qwencloud.com/m/20000000190/ Meet advanced AI Agents — QoderWork, JVS Claw, MuleRun, Quick BI, QwenPaw, and Accio Work — built to work, think, and scale with you. Explore how they power automation, analytics, and collaboration to drive the next wave of productivity innovation.

译诚邀您于2026年5月26日参加2026年Qwen大会期间的阿里云Claw Talks线下巡展，与我们一同探索下一波AI智能体应用浪潮：https://click.qwencloud.com/m/20000000190/ 认识先进的AI智能体——QoderWork、JVS Claw、MuleRun、Quick BI、QwenPaw和Accio Work——它们专为与您协同工作、思考和扩展而构建。探索它们如何赋能自动化、分析和协作，以推动下一波生产力创新。

Kling AI@Kling_ai · 5月21日67

Kling AI Debuts at Cannes — RAPHAEL RAPHAEL, a 100% AI-generated feature film project, is a large-scale production being developed in collaboration between Mateo AI Studio and MBC C&I’s AI Content Lab, a leading force in AI video production in Korea. Currently in production with the goal of a theatrical release in 2026, this project is leveraging Kling AI’s powerful video model throughout the production process to maximize distinctive visual effects and deliver a differentiated cinematic experience. It proves the industrial viability of pure AI filmmaking and marks a new trend for AI-native theatrical cinema.

译AI电影项目RAPHAEL在戛纳亮相。该片由Mateo AI Studio与韩国MBC C&I的AI内容实验室联合开发，全程使用Kling AI视频模型进行制作，旨在实现独特的视觉效果与差异化的观影体验。项目计划于2026年登陆院线，其大规模制作旨在证明纯AI电影制作的工业可行性，标志着AI原生院线电影新趋势的开端。

SenseTime@SenseTime_AI · 5月21日64

🏆 @SenseTime_AI 𝗵𝗮𝘀 𝗯𝗲𝗲𝗻 𝗻𝗮𝗺𝗲𝗱 𝗖𝗵𝗶𝗻𝗮’𝘀 𝗡𝗼. 𝟭 #VisionAI 𝗹𝗲𝗮𝗱𝗲𝗿 𝗯𝘆 @IDC 𝗳𝗼𝗿 𝘁𝗵𝗲 𝟭𝟬𝘁𝗵 𝗰𝗼𝗻𝘀𝗲𝗰𝘂𝘁𝗶𝘃𝗲 𝘆𝗲𝗮𝗿. Behind the milestone: 🚀 𝗖𝗩 𝟮.𝟬 𝗘𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻: Spearheaded the transition to next-generation Vision AI architectures, natively embedding GenAI capabilities. 🌏 𝗢𝘃𝗲𝗿𝘀𝗲𝗮𝘀 𝗚𝗿𝗼𝘄𝘁𝗵: Expanded to 12 international markets, serving 500+ enterprise clients with exceptional retention. 💹 𝗣𝗿𝗼𝗳𝗶𝘁𝗮𝗯𝗹𝗲 𝗚𝗿𝗼𝘄𝘁𝗵: Anchored by accelerating annual recurring revenue (ARR). 🔄 𝗔𝗜 𝗮𝘁 𝗦𝗰𝗮𝗹𝗲: Transitioning from project-based deployments to a highly scalable AI Platform. A decade of market leadership cements our foundation to lead the next wave of technology advancement, including the much-anticipated convergence of Vision AI and Genrative AI.

译商汤科技被IDC连续十年评为中国计算机视觉市场第一。其关键驱动因素包括：主导推进CV 2.0架构演进，原生嵌入生成式AI能力；海外市场扩展至12个国际区域，服务超500家企业客户并保持高留存率；以年度经常性收入增长为锚点实现盈利性增长；并从项目制交付转型为高度可扩展的AI平台。这一十年市场领导地位为其引领计算机视觉与生成式AI的融合浪潮奠定了基础。

小互@xiaohu · 5月21日78

FSD来了… 官宣进入大陆…

译FSD来了… 官宣进入大陆…

Krea@krea_ai · 5月21日27

we're hosting our first Community Spaces tomorrow! join us through X Spaces to chat about updates coming to Krea 2, share feedback, or showcase your work. link below 👇

译我们将于明天举办首次社区空间活动！ 欢迎通过 X Spaces 加入我们，一起讨论 Krea 2 即将推出的更新、分享反馈或展示你的作品。 链接如下 👇

OpenAI Developers@OpenAIDevs · 5月21日58

Show us what you’re building with realtime voice. Join the OpenAI team in SF on May 27 for a demo showcase using the latest voice models. We’re looking for prototypes and products that are interesting, useful, creative, and technically ambitious. Top projects will present onstage, win prizes, and be featured by @OpenAIDevs & @cerebral_valley for a community vote.

译展示你正在构建的实时语音应用。 加入OpenAI团队，于5月27日在旧金山参加使用最新语音模型的演示展示。 我们正在寻找有趣、实用、富有创意且技术上具有雄心的原型和产品。 优秀项目将上台展示、赢得奖品，并由@OpenAIDevs和@cerebral_valley进行社区投票。

🚨 AI News | TestingCatalog@testingcatalog · 5月21日69

Anthropic 🤝 SpaceX Anthropic is getting up to GB200 of capacity in Colossus 2 in June as a part of the expanded agreement with SpaceX. Partnerships are huge unlocks 👀

译Anthropic 🤝 SpaceX 作为与SpaceX扩大协议的一部分，Anthropic将于6月在Colossus 2中获得高达GB200的算力容量。 合作伙伴关系是巨大的解锁因素 👀

Elon Musk@elonmusk · 5月21日80

As the recently expanded partnership with @AnthropicAI demonstrates, @SpaceX is offering AI compute as a service at significant scale. We are in discussions with other companies to do the same. Over time, especially with orbital data centers, we expect to serve AI at extremely high scale.

译正如最近与@AnthropicAI扩大合作所展示的，@SpaceX正在大规模提供AI算力服务。 我们正在与其他公司进行类似合作的讨论。 随着时间的推移，特别是通过轨道数据中心，我们预计将以极高的规模提供AI服务。

NotebookLM@NotebookLM · 5月21日37

We're back in action! Back to your regularly programmed studying, learning, and research. (Thank you SO much for your patience and understanding 🙏)

译我们已恢复正常运行！请继续您的日常学习、研究和探索。 （非常感谢您的耐心与理解 🙏）

Chubby♨️@kimmonismus · 5月21日65

Holy: After Anthropic secured compute capacity from Colossus 1, it is now also getting access to compute from Colossus 2. But to be honest: I somehow expected to his. Grok certainly doesn’t need all that compute.

译Holy：继从Colossus 1获得算力后，Anthropic现在也将获得Colossus 2的算力支持。 但说实话：我多少预料到了这点。Grok显然不需要那么多算力。

Chubby♨️@kimmonismus · 5月21日64

Even bigger win for Anthropic: Anthropic’s revenue is set to more than double to $10.9 billion in the second quarter, an explosive rate of growth that will help it turn an operating profit for the first time. About $500m in profit. Nuts.

译Anthropic迎来更大胜利： Anthropic第二季度营收预计将翻倍至109亿美元，这一爆发式增长将助其首次实现营业利润。 利润约5亿美元。 难以置信。

Sam Altman@sama · 5月21日84

a general-purpose model solved a major open problem in mathematics. we'll be saying this a lot over the coming years, but this is a kinda big milestone. i'm very excited for AI to greatly extend our understanding of the world, but still, i have complicated feelings today.

译一个通用模型解决了数学领域的一个重大开放问题。 未来几年我们会经常说这句话，但这确实是一个相当重要的里程碑。 我非常期待AI能极大地拓展我们对世界的理解，但今天，我的心情依然很复杂。

Replit ⠕@Replit · 5月21日35

LA-born new media artist whose installations, public sculptures, and live AV works blend AI, memory, and urban landscape, @NateMohlerArt is the founder of Projekt Blank. See his workshop and installation on day one of Vibecon. NYC, June 17–18. Get your tickets at http://vibecon.ai

译出生于洛杉矶的新媒体艺术家，其装置、公共雕塑和现场视听作品融合了AI、记忆与城市景观，@NateMohlerArt 是 Projekt Blank 的创始人。 他将在 Vibecon 第一天展示其工作坊与装置作品。 纽约，6月17-18日。 购票请访问 http://vibecon.ai

SemiAnalysis@SemiAnalysis_ · 5月21日14

SemiAnalysis: relentlessly releasing god models

译SemiAnalysis：持续发布神级模型

Jeff Dean@JeffDean · 5月21日18

Four questions in under 2 minutes. Thanks to Google's social team for suggesting this little interlude at #googleio yesterday!

译两分钟内回答四个问题。感谢Google社交团队在昨天#GoogleIO上建议的这个小环节！

OpenAI Developers@OpenAIDevs · 5月21日51

Come join us at a Codex meetup 🌏 https://developers.openai.com/community/meetups

译来参加我们的Codex线下聚会吧 🌏 https://developers.openai.com/community/meetups

swyx🛬 SFO@swyx · 5月21日63

btw we did a bake off of Exa vs competitors and it took all of 1.5 hrs for the team to unanimously converge on exa lol. so proud to see my former landlords crush it - time travel back to last year and listen to a pre pmf @WilliamBryk to understand how to spot companies on a generational tear

译Exa完成2.5亿美元融资，估值达22亿美元，由a16z领投。该公司致力于为AI智能体优化搜索，服务超5000家公司和50万开发者。其在代码、公司等多个垂直领域搜索效果顶尖，能减少90%返回文本且几乎不影响RAG质量，并构建性价比最优的网页智能体。创始人强调搜索创新对对抗虚假信息至关重要，此轮融资将用于改善社会信息环境。推文作者团队在快速对比测试后，一致认可Exa的领先优势。

Artificial Analysis@ArtificialAnlys · 5月21日37

Artificial Analysis benchmarks were featured in yesterday’s Gemini 3.5 Flash launch Yesterday @GoogleDeepMind released Gemini 3.5 Flash at Google I/O '26 and our benchmarks were used by @sundarpichai to highlight the model’s leading position on the Intelligence vs. Speed Pareto frontier. We have full benchmarks across intelligence, speed, token usage and cost on the Artificial Analysis website

译Artificial Analysis基准测试被纳入昨日Gemini 3.5 Flash的发布 昨日@GoogleDeepMind在Google I/O '26上发布了Gemini 3.5 Flash，我们的基准测试被@sundarpichai用于展示该模型在智能与速度帕累托前沿的领先地位。我们在Artificial Analysis网站上提供了涵盖智能、速度、token使用量和成本的完整基准测试数据。

Rohan Paul@rohanpaul_ai · 5月21日75

CNBC: Meta starts major cuts with 8,000 layoffs as AI shakes the tech giant. Along with the layoffs, around 7,000 employees will be shifted into new AI-focused roles. Meta is not only trimming costs, it is changing its internal shape around AI infrastructure, foundation models, and AI monetization, which means the company wants more people building the systems that train models, the models themselves, and the products that turn those models into revenue. --- cnbc .com/2026/05/20/meta-layoffs-zuckerberg-says-success-isnt-a-given-in-memo.html

译Meta启动大规模裁员，计划削减约8000个岗位。与此同时，约7000名员工将被调配至新的AI相关职位。此次调整并非单纯的削减成本，更是公司围绕AI进行内部结构重塑的关键举措，重点将资源集中于AI基础设施、基础模型的构建以及AI技术的商业化，旨在打造从模型训练、产品开发到实现盈利的完整链条。

Rohan Paul@rohanpaul_ai · 5月21日74

Mark Zuckerberg told employees in a Wednesday memo that laying off 8,000 workers was necessary because “success isn’t a given.” The full memo, as published by businessinsider. "Hey everyone, I want to express my gratitude to everyone leaving today for all of the hard work you've put into serving our community. It's always sad to say goodbye to people who have contributed to our mission and to building this company. I feel the weight of that, and I'm spending a lot of time making sure we manage this as well as possible. This is the most dynamic I have seen our industry. I'm optimistic about everything we're building to give billions of people the power to express themselves and connect with the people they care about. I'm also optimistic about delivering personal superintelligence to everyone. We've always focused on putting power in people's hands. This is how we believe progress is made in the world. These values are what makes us different, and they are why Meta has been successful. But success isn't a given. AI is the most consequential technology of our lifetimes. The companies that lead the way will define the next generation. We're transforming our company to make sure it will always be the best place for talented people to have the greatest impact. People tell us that they appreciate the ability to take greater ownership and execute their vision with less bureaucracy and management to navigate. At the same time, we also want to provide everyone with as much stability as possible. We won't always get this balance right, but I care deeply about this so we'll keep adjusting and work hard to do right by people along the way. To that end, I want to be clear that we do not expect other company-wide layoffs this year. I also want to acknowledge that we haven't been as clear as we aspire to be in our communication, and that's one area I want to make sure we improve. I'm confident in what we're all building together. We are one of the few companies positioned to help define the future. Meta has the talent, the infrastructure, the apps and distribution, and the business model. We have a lot of work ahead, but what's on the other side is going to be extraordinary. Once again, I'm grateful to those leaving today. And I'm grateful to everyone around the company for all of the historic work we will continue doing together. Mark" --- businessinsider .com/heres-what-mark-zuckerberg-said-about-future-layoffs-at-meta-2026-5

译扎克伯格在给员工的备忘录中表示，裁员8000人是因为“成功并非必然”，AI是决定性技术，领先企业将定义下一代。此次裁员与公司重组相关：Meta在裁减岗位的同时，将约7000名员工转向AI相关的新岗位，重塑公司结构以聚焦AI基础设施、基础模型开发及AI变现。扎克伯格对离开的员工表示感谢，并重申Meta拥有定义未来的实力。他承诺今年不会再有全公司范围的裁员，同时承认公司沟通方式有待改进。