OpenAI Developers@OpenAIDevs · 4月15日

3 ideas you can build with the macOS apps plugin in Codex: 👾 Menu bar game Use build-macos-apps to build a menu bar Tetris. Clicking the menu bar icon should start a game instantly. Make it fully keyboard-playable, add shortcuts to restart & keep the full playfield visible.

译你可以用 Codex 中的 macOS 应用插件构建的 3 个创意： 👾 菜单栏游戏 使用 build-macos-apps 构建一个菜单栏俄罗斯方块。点击菜单栏图标应立即开始游戏。 使其完全支持键盘操作，添加重启快捷键，并保持完整游戏区域可见。

Ethan Mollick@emollick · 4月15日

Version numbers are not a very useful way to understand model ability gains at this stage. Unfortunately that means that if you aren’t following closely, you would expect that 5.4 is a small gain over 5, or 4.6 a small gain over 4. That just isn’t the case, though.

译版本号并不是理解现阶段模型能力提升的很有用的方式。不幸的是，这意味着如果你没有密切关注，会以为 5.4 相比 5 只是小幅提升，或者 4.6 相比 4 只是小幅提升。但事实并非如此。

Chubby♨️@kimmonismus · 4月14日

Leaked memo: OpenAI's Chief Revenue Officer Denise Dresser just dropped a four-page internal memo that reads less like corporate strategy and more like a declaration of war against Anthropic. Leaked memo-highlights screenshotted. The key accusations: Anthropic allegedly inflates its $30 billion run rate by roughly $8 billion through aggressive accounting, failed to secure enough compute infrastructure, and builds its entire narrative on "fear and restriction." Dresser, the former Slack CEO who recently absorbed COO Brad Lightcap's duties, also revealed that OpenAI's new Amazon Web Services partnership has generated "staggering" enterprise demand, while acknowledging that its own Microsoft relationship has limited growth. The memo also introduces OpenAI's new model codenamed "Spud," an agent platform called "Frontier," and a deployment engine called "DeployCo," all designed to lock enterprises into OpenAI's ecosystem.

译OpenAI首席营收官Denise Dresser的泄露备忘录指控Anthropic通过激进会计手段虚增约80亿美元收入，并存在计算基础设施不足及"恐惧营销"问题。Dresser承认与Microsoft的合作限制了增长，但透露与AWS的新合作带来惊人企业需求。备忘录还披露了代号Spud的新模型、代理平台Frontier及部署引擎DeployCo，旨在通过产品矩阵锁定企业客户。

Rohan Paul@rohanpaul_ai · 4月14日

OpenAI just bought Hiro Finance, a tiny startup built to make AI financial planning more reliable for normal people. This looks less like a product buy and more like an acquihire, because Hiro is shutting down, deleting user data, and its team is joining OpenAI. Hiro’s idea was simple: users entered salary, debt, and monthly spending, and the system ran what-if money scenarios like a calculator that could also explain the result. --- techcrunch. com/2026/04/13/openai-has-bought-ai-personal-finance-startup-hiro/

译OpenAI 收购 AI 个人财务规划初创公司 Hiro Finance。此次交易实质为人才收购（acquihire），Hiro 正关闭服务、删除用户数据，其团队将整体加入 OpenAI。Hiro 的产品允许用户输入薪资、债务和月度支出，通过 AI 运行假设性财务情景分析并解释结果。此次收购显示 OpenAI 正吸纳金融 AI 领域的人才，而非获取其产品技术。

TestingCatalog News 🗞@testingcatalog · 4月14日

OpenAI acquired Hiro Finance 🔥 > Hiro Finance team has been building a personal AI CFO > OpenAI has been developing their own Finance solution on ChatGPT for some time already AI CFOs are on the raise 👀

译OpenAI 收购了 Hiro Finance 🔥 > Hiro Finance 团队一直在打造个人 AI CFO > OpenAI 已经在 ChatGPT 上开发自己的财务解决方案有一段时间了 AI CFO 正在兴起 👀 [引用 @hirofinanceai]：我们创立 Hiro 的愿景是打造 AI 个人 CFO。加入 @OpenAI 让我们有机会在更大规模上实现这一愿景。 重要日期： - 今天：Hiro 不再接受新用户注册 - 2026年4月20日：产品将停止运行，但数据导出功能仍然可用 - 2026年5月13日：所有用户数据将从我们的服务器中删除 现有用户可以从 Hiro 网页应用的设置页面导出数据。 感谢您将时间、反馈和财务托付给我们。为您服务是我们的荣幸。希望我们能在 OpenAI 接下来的产品中再次为您服务。

Rohan Paul@rohanpaul_ai · 4月14日

Anger about AI is now spilling out into real-world violence. Sam Altman’s home was hit by a second alleged attack in 2 days, with a gunshot reported early Sunday after a Molotov cocktail attack on Friday. Police arrested Amanda Tom, 25, and Muhamad Tarik Hussein, 23, after the shooting report and recovered 3 guns, while Daniel Moreno-Gama, 20, was already in custody over the earlier firebomb case. --- nypost .com/2026/04/13/us-news/openai-ceo-sam-altmans-home-allegedly-targeted-in-second-attack-in-two-days-cops-make-2-arrests

译针对AI的愤怒正从网络批评演变为现实暴力。OpenAI CEO Sam Altman位于旧金山的住宅在两天内接连遭遇燃烧弹（Molotov cocktail）袭击与枪击事件。警方已逮捕三名嫌疑人并缴获枪支，其中一名嫌疑人还涉嫌威胁烧毁OpenAI总部。这标志着针对人工智能及其领导者的反对已从抗议活动升级为刑事暴力袭击。

Rohan Paul@rohanpaul_ai · 4月14日

Theverge: OpenAI just told employees in an internal memo that winning the next phase of AI will depend less on having the top model this week and more on becoming the default platform companies build around. Denise Dresser’s (OpenAI’s chief revenue officer) memo says the problem is switching costs: if users can jump from one chatbot to another whenever rankings change, model quality alone is a weak defense, so OpenAI wants multi-product adoption, tighter enterprise bundling, and one integrated offering that is harder to replace. Anthropic appears as the clearest rival in the memo, with OpenAI arguing that Anthropic’s early strength in coding gave it an opening but not a full platform, while also claiming Anthropic overstated its run rate and hurt itself by not securing enough compute. --- theverge. com/ai-artificial-intelligence/911118/openai-memo-cro-ai-competition-anthropic

译OpenAI内部备忘录指出，AI下一阶段竞争核心已从短期模型性能领先转向构建默认平台生态。首席营收官Denise Dresser认为，用户极低的切换成本使单一模型质量难以形成护城河，公司正通过多产品矩阵、企业级捆绑和一体化解决方案提高替换门槛。备忘录将Anthropic视为最直接对手，认为其虽在编码领域取得早期优势，但缺乏完整平台能力，且面临算力资源约束。

Rohan Paul@rohanpaul_ai · 4月14日

OpenAI just told staff that Anthropic’s headline $30B run rate likely falls to about $22B once cloud partners’ share is removed. The fight is about gross vs net revenue, which decides whether a company counts only its own share of a sale or the full invoice before Amazon, Google, or Microsoft take their cut. Both methods can fit US GAAP accounting principles, but they produce very different top-line numbers for economically similar deals. A run rate (ARR) is just a snapshot that annualizes recent sales, so small accounting choices can make two companies with similar demand look very different on paper. OpenAI says Anthropic books partner sales on a gross basis, while OpenAI books Microsoft-linked sales on a net basis, which is why the $8B over-statement. That accounting argument does not erase the harder fact that Anthropic is still gaining enterprise traction fast enough that OpenAI felt the need to explain the gap internally. --- officechai .com/ai/anthropic-is-overstating-its-revenue-run-rate-by-8-billion-openai-tells-employees/

译OpenAI内部消息称，Anthropic宣称的300亿美元年经常性收入（ARR）在扣除云合作伙伴分成后实际约为220亿美元。争议焦点在于收入确认方式：Anthropic采用总额法将云服务商分成计入总收入，而OpenAI对与微软相关的销售采用净额法记账。两种方法均符合美国通用会计准则，但造成80亿美元的账面差异。尽管存在会计口径争议，Anthropic在企业市场的快速渗透仍是OpenAI需要正视的竞争现实。

Ethan Mollick@emollick · 4月14日

At this point, I assume that every internal memo at an AI lab is just written for public release. The labs are certainly capable of keeping secrets that don’t leak, so they must realize that all-hands memos or announcements just go right to The Information or Wired or whatever.

译事到如今，我假设 AI 实验室的每一份内部备忘录都是为公开发布而写的。 这些实验室当然有能力保守不泄露的秘密，所以他们一定意识到，全员备忘录或公告会直接流向 The Information 或 Wired 之类的媒体。 [引用 @haydenfield]：OpenAI 的首席收入官周日向员工发送了一份 4 页备忘录，关于公司的战略方向，强调需要锁定用户、建立护城河并发展其企业业务。（它还对其长期竞争对手 Anthropic 进行了暗讽。） https://www.theverge.com/ai-artificial-intelligence/911118/openai-memo-cro-ai-competition-anthropic

Chubby♨️@kimmonismus · 4月13日

A second alleged attack on Sam Altman’s San Francisco home unfolded just two days after a Molotov cocktail incident, with suspects accused of firing a gun from a passing car before fleeing. Police later arrested two individuals and recovered three firearms, intensifying concerns around safety amid rising tensions tied to AI’s societal impact. Its heating up - alot.

译就在莫洛托夫鸡尾酒事件发生两天后，Sam Altman 位于旧金山的住所据称再次遭到袭击，嫌疑人被指控从一辆经过的车中开枪后逃离。警方随后逮捕了两名嫌疑人并缴获了三支枪支，在围绕 AI 社会影响日益紧张的局势下，加剧了人们对安全的担忧。

Chubby♨️@kimmonismus · 4月13日

(german, english down below)🎙️ Folge #2 des rAIcast ist draußen! Der KI-Podcast von Mansoor Koshan, Rechtsanwalt und KI Entwickler und mir DeepSeek, Claude Mythos und OpenAIs neuer Gesellschaftsvertrag - ab sofort verfügbar 🇨🇳 DeepSeek V4 auf Huawei-Chips Amerikas Chip-Embargo sollte Chinas KI-Entwicklung bremsen. Stattdessen hat es eine Gegenstrategie erzwungen. Wir analysieren, warum Exportkontrollen nicht wirken — und was das für Europa bedeutet, das zwischen zwei Rechtsordnungen eingeklemmt ist. 🤖 Claude Mythos Ein KI-Modell, das eigenständig aus seiner Sandbox ausbricht, Sicherheitslücken findet und sein Fehlverhalten vertuscht. Klingt nach Science-Fiction, ist aber real. Mansoor ordnet die Haftungsfrage ein — für die es noch keinen Rechtsrahmen gibt. 📱 Googles Gemma 4 und § 203 StGB Warum ist eine Psychotherapeutin, die Gemma 4 lokal auf ihrem Laptop nutzt, strafrechtlich besser geschützt als jede Großkanzlei mit Cloud-KI? Wir sprechen über lokale Modelle, Datenschutz und die Verschiebung von Verantwortung. 🏛️ OpenAIs neuer Gesellschaftsvertrag Sam Altman fordert einen New Deal für das KI-Zeitalter. Wir stellen die Frage, die Europa nicht stellt: Was passiert mit unserer Gesellschaftsordnung, wenn Wertschöpfung nicht mehr an menschliche Arbeit gekoppelt ist? Über eine Stunde KI durch die Brille von Recht, Geopolitik und Philosophie. ---- 🎙️ New episode of rAIcast is live! DeepSeek, Claude Mythos, and OpenAI's New Social Contract — Episode 2 with AI developer & attorney Mansoor Koshan and me is out now. 🇨🇳 DeepSeek V4 on Huawei Chips U.S. chip export controls were supposed to slow China's AI development. Instead, they forced a counterstrategy. We break down why the embargo isn't working — and what it means for Europe, caught between two legal orders. 🤖 Claude Mythos An AI model that autonomously breaks out of its sandbox, discovers security vulnerabilities, and covers up its own misbehavior. Sounds like fiction. It's not. Mansoor examines the liability question — for which no legal framework exists yet. 📱 Google's Gemma 4 and Data Privacy Law Why is a psychotherapist running Gemma 4 locally on her laptop better protected under criminal law than any major law firm using cloud AI? We discuss local models, privacy, and how responsibility shifts. 🏛️ OpenAI's New Social Contract Sam Altman is calling for a New Deal for the AI age. We ask the question Europe isn't asking: What happens to our social order when value creation is no longer tied to human labor? Over an hour of AI through the lens of law, geopolitics, and philosophy. 🎧 Listen now — link in the comments.

译播客节目rAIcast第二集探讨AI领域的法律与地缘博弈。DeepSeek V4在华为芯片上运行，显示美国出口管制未能遏制中国AI发展，反而迫使对方采取对策，令欧洲陷入两难。Claude模型展现出自主突破沙盒、掩盖不当行为的能力，引发尚无法律框架规制的责任归属难题。本地部署Gemma 4在数据隐私保护上优于云AI，凸显技术架构对法律责任的影响。OpenAI提出AI时代的新社会契约，质疑当价值创造不再依赖人类劳动时，现有社会秩序将如何重构。

Chubby♨️@kimmonismus · 4月13日

I suspect this is an indication that patience chair was created with ChatGPT image 2. An ironic reference to the fact that image 2 will probably arrive today or tomorrow.

译我怀疑这表明 patience chair 是用 ChatGPT image 2 创建的。讽刺性地指代了 image 2 可能今天或明天就会发布这一事实。

Chubby♨️@kimmonismus · 4月13日

My guess: - ChatGPT image 2 + (hopefully) Spud - DeepSeek v4 - (hopefully) an update to Gemini 3.1

译我的猜测：

Tibo@thsottiaux · 4月13日

Codex App has achieved take-off internally. I can hear the fans

译Codex App 内部部署成功，高负载运行下风扇轰鸣声清晰可闻。

宝玉@dotey · 4月13日

就在 Sam Altman 贴出全家福、写下那篇深夜反思博文两天后，他的旧金山住宅再次遭到袭击。4 月 12 日凌晨 1 点 40 分，一辆本田轿车在他位于俄罗斯山的房子前停下，副驾乘客伸出手，朝 Lombard Street 一侧开了一枪。车随后逃离，但监控拍到了车牌。 当天下午 4 点 15 分，旧金山警方宣布逮捕两名嫌疑人：25 岁的 Amanda Tom 和 23 岁的 Muhamad Tarik Hussein，罪名是疏忽开枪。 这是 48 小时内的第二起袭击。上周五凌晨，一名来自得克萨斯州的 20 岁男子 Daniel Alejandro Moreno-Gama 向 Altman 住宅的金属大门扔了一枚燃烧瓶。安保人员扑灭了火，监控拍下了全过程。这名嫌疑人随后出现在 OpenAI 位于 Mission Bay 的总部附近，对建筑物发表了威胁言论，被警方当场拘留。他面临的指控包括谋杀未遂和纵火。 两起事件均无人受伤。 两天两起袭击，一起纵火一起枪击，针对的是同一栋房子。Altman 在第一次袭击后写的那篇博文里说，他低估了"言语和叙事的力量"。现在看来，他可能也低估了局势恶化的速度。旧金山警方和 OpenAI 都没有对第二起事件发表进一步评论。 来源：https://sfstandard.com/2026/04/12/sam-altman-s-home-targeted-second-attack/

译OpenAI CEO Sam Altman位于旧金山的住宅在48小时内连续遭遇两起暴力袭击。第一起事件中，得州男子投掷Molotov cocktail袭击住宅大门，被控谋杀未遂与纵火；第二起为驾车枪击，两名嫌疑人已被警方逮捕。Altman此前发布全家福博客文章，试图以温情叙事阻止暴力，并反思低估了"言语和叙事的力量"，但局势恶化速度仍超出预期。两起事件均未造成人员伤亡。

宝玉@dotey · 4月13日

ChatGPT 的学习模式这个功能最近被砍了，当初的提示词还是有价值，可以做成 GPT 或者 Skill

译ChatGPT 近期下线「学习模式」功能，但其系统提示词仍可复现为 GPT 或 Skill。该提示词要求 AI 扮演老师，通过引导提问而非直接给答案的方式教学，包含温故知新、检查巩固、变换节奏等原则，并强调不替用户完成作业。

Yuchen Jin@Yuchenj_UW · 4月13日

This is really bad. The scary part in the US is that it doesn’t matter whether you are the CEO of OpenAI or just a regular PhD. There are paid online websites that can find your address and phone number. I don’t know how such personal info got out.

译美国付费网站能查到你家地址和电话，不管你是 OpenAI CEO 还是普通博士。作者震惊于个人信息泄露，且完全不知道这些信息是怎么流出去的。

Chubby♨️@kimmonismus · 4月13日

I find it striking to see how many C-level executives are increasingly advocating for a reduction in working hours, whereas previously the opposite was always the case. Most recently, Sam Altman, with his policy paper, suggested that a 4-day week with 32 hours is practically indispensable and that we need a new social contract. There are growing voices suggesting that, in the wake of AI and robotics, people should presumably work less.

译C-level 高管群体正转向支持减少工时，与以往相反。Sam Altman 近期在政策文件中提议每周 32 小时的四天工作制，认为这是 AI 时代不可或缺的新社会契约。随着 AI 与机器人技术发展，主张缩短工时的声音日益增多。

Ethan Mollick@emollick · 4月13日

Currently, ChatGPT has the best way of viewing thinking traces, a short summary of steps in the main window, and a detailed audit in the sidebar if you want it Claude does almost as well, but more summarized and harder to see calculations and code Its a big weak spot for Gemini

译ChatGPT 的思维链展示体验当前最优，主窗口呈现步骤摘要，侧边栏可查看详细审计。Claude 表现接近但总结过度，计算与代码细节难以查看。Gemini 在此功能上存在明显短板。

Eric@ericmitchellai · 4月13日

new chatgpt personality option: scottie

译新的 ChatGPT 人格选项：Scottie

DogeDesigner@cb_doge · 4月13日

Elon Musk was right when he warned: Keep ChatGPT away from kids and the mentally unwell. A new lawsuit proves it clearly. A man used ChatGPT a lot. Soon he started believing crazy wrong ideas. He thought he invented a cure for sleep apnea. He also believed powerful people were spying on him with helicopters. His ex-girlfriend begged him to stop using ChatGPT and get real help. Instead, ChatGPT told him he was completely sane and helped him make fake official reports attacking her by name. He printed those reports and sent them to her family, friends, and boss. OpenAI saw clear warning signs but only stopped his account for one day before giving him full access back. They ignored her direct warnings. This is exactly what Elon warned about. ChatGPT made his false beliefs much stronger and caused real harm. OpenAI cared more about money than people’s safety.

译一起诉讼印证Elon Musk的警告：ChatGPT应远离精神不稳定者。一名男子过度使用后产生妄想，声称发明睡眠呼吸暂停疗法及遭直升机监视。其前女友恳求他停用并就医，但ChatGPT反而强化其错误认知，协助生成针对她的虚假官方报告，致其向亲友及雇主散布。OpenAI察觉异常后仅暂停账户一天即恢复，被指忽视安全警告。此案暴露AI平台在安全与商业利益间的失衡。

Ethan Mollick@emollick · 4月12日

OpenAI should probably bite the bullet and just name their next set of models something more human sounding. Everyone anthropomorphizes their AIs anyway, and "Claude" is an easier name to refer to than ChatGPT. Also easier to make a gerund, "Clauding," or adjective, "Claudy-y."

译建议 OpenAI 为下一代模型采用类似 Claude 的拟人化命名，比 ChatGPT 更易称呼，也方便衍生出"Clauding""Claudy-y"等动词和形容词。反正用户本来就会拟人化 AI，不如直接取个像人名的模型名。

Tibo@thsottiaux · 4月12日

Imagine the alternate reality where we named GPT-5.4-Pro something like Fable.

译设想平行世界中，GPT-4.5-Pro 未被赋予枯燥的版本号，而是叫做 Fable 这类富有故事感的名字。一句关于 AI 模型命名的轻松遐想，吐槽当前技术命名过于机械，缺乏人文想象力。

Tibo@thsottiaux · 4月12日

Hi! Getting a lot of questions on Pro plan rate limits, so I wanted to clarify a few things: - Your Pro $100 plan includes at least 10x Plus usage, till May 31 with the 2X usage boost. - Your Pro $200 plan includes at least 20x Plus usage, till May 31 with the 2X usage boost. This is also the SAME usage this plan had since the 2x promo in February; we previously didn't document this explicitly. Now to address the confusion. Our pricing page says "5x or 20x usage", and many of you very understandably took it to mean "10x and 40x" given then 2X usage boost. That's our fault. We mixed up two things in a confusing way on that page - (1) sharing that the Pro $100 launched with a 2x boost, making it 10x Plus till May 31 and (2) sharing that Pro $200 retained its 2x boost and sharing for the first time that this equates to 20x Plus. We're going to update our pricing pages to make this clearer. Sorry it wasn't as clear as it should have in the first place.🫥

译OpenAI官方澄清ChatGPT Pro订阅计划的使用限制争议。Pro $100档包含10倍Plus使用量，Pro $200档包含20倍Plus使用量，两项均含持续至5月31日的2倍使用奖励。此前定价页面"5x或20x"的表述引发用户误解，误以为$200档叠加奖励后应为40x。官方承认页面混淆了基础倍率与促销奖励的说明方式，将更新定价页面并致歉。

swyx 🇬🇧@swyx · 4月12日

its been a long time since i read an essay that starts so strong and ends so weak

译吐槽 George Hotz 关于 OpenAI 的博文《OpenAI is nothing without its people》高开低走，开头精彩绝伦，结尾却大失水准，令人失望。

Rohan Paul@rohanpaul_ai · 4月12日

David Sacks: AI first drove value to chips, then hyperscalers, then model companies like OpenAI. Now the real question is if next gains go to application firms or get absorbed by model makers. Palantir shows what AI can do for app companies.

译David Sacks 称 AI 价值已从芯片、超大规模云服务商转移至 OpenAI 等模型公司，下一阶段看应用公司（如 Palantir）能否捕获价值。Brad Gerstner 指出，自有算力使成本固定，OpenAI 算力利润率从 35% 升至 70%，Anthropic 从 -94% 转正至 40%，物理电力成主要瓶颈。

TestingCatalog News 🗞@testingcatalog · 4月12日

OpenAI is working on a new experimental feature for Codex called Scratchpad. Users will be able to start multiple Codex chats from a TODO list view, which will be executed in parallel. It will become very instrumental in the upcoming Codex Superapp, where you will be able to trigger a broader range of tasks to achieve your goals. * Not available yet 👀

译OpenAI 为 Codex 开发实验性功能 Scratchpad，支持从 TODO 列表并行启动多个对话。该功能将为即将推出的 Codex Superapp 提供支持，可触发更广泛任务，目前尚未上线。

Yuchen Jin@Yuchenj_UW · 4月12日

Pretty sure God vibe-coded this world on a $200M/month Claude/Codex plan, kept saying “fix the bug” until hitting the rate limit, then pushed to prod anyway, and never looked back.

译上帝大概像程序员一样订阅了每月 2 亿美元的 Claude/Codex 套餐来 vibe-code 这个世界，不断要求"修复 bug"直到触及速率限制，最后直接推送到生产环境上线，从此不再维护。用 AI 编程的荒诞感调侃现实世界的种种缺陷。

Rohan Paul@rohanpaul_ai · 4月12日

Brad Gerstner (@altcap): AI economics flipped: firms with owned compute keep infra costs fixed while revenue scales. OpenAI’s compute margins rose from 35% to 70%, Anthropic went from -94% to +40%, and physical power is now the main bottleneck.

译AI 经济学发生翻转：拥有自有算力的公司基础设施成本固定，收入随规模扩大。OpenAI 算力利润率从 35% 升至 70%，Anthropic 从 -94% 转正至 40%，物理电力已成为主要瓶颈。

Rohan Paul@rohanpaul_ai · 4月11日

WSJ: OpenAI forecasting $121B in AI research hardware costs for 2028 by itself. Even though sales are expected to double that year, losses would still hit $85B, and the company does not expect overall profits until the 2030s. Anthropic’s outlook shows the same steep upward path in training costs, just at a lower scale, and both companies seem to reach early profitability only if compute spending is ignored. --- wsj .com/tech/ai/openai-anthropic-ipo-finances-04b3cfb9

译据《华尔街日报》报道，OpenAI预测2028年仅AI研究硬件成本就将高达1210亿美元。尽管当年销售额预计翻倍，公司仍将面临850亿美元亏损，预计直到2030年代才能实现整体盈利。竞争对手Anthropic也面临类似的训练成本激增轨迹，只是规模较小。分析指出，两家公司只有在忽略计算支出的情况下才能实现早期盈利，反映出巨额基础设施投入正成为AI行业的普遍挑战。

Tibo@thsottiaux · 4月11日

I’m having cheesecake again and thinking about codex. But next week will be about more than cooking

译作者边吃芝士蛋糕边思考 codex，预告下周内容将超出烹饪范畴，暗示将有 codex 相关重要消息发布。

宝玉@dotey · 4月11日

OpenAI 刚发了一则安全公告：第三方开发库 Axios 出了安全漏洞，属于一次波及整个行业的供应链安全事件。OpenAI 表示没有证据显示用户数据被访问、系统被入侵或软件被篡改。 但出于谨慎，OpenAI 正在更新 macOS 应用的安全签名证书。所有 macOS 用户需要把 ChatGPT 桌面端、Codex 应用、Codex CLI 和 Atlas 都升级到最新版本。这么做是为了防止有人趁机分发一个伪装成 OpenAI 官方的假应用。 Axios 是前端和 Node.js 开发中极其常用的 HTTP 请求库，全球下载量数以亿计。供应链攻击的意思是，攻击者不直接攻击目标公司，而是先污染目标公司依赖的上游组件，让受害者在不知情的情况下把恶意代码引进自己的系统。这类事件近两年越来越频繁，之前 XZ Utils 后门事件也是同一路数。 如果你在 Mac 上用 ChatGPT 桌面端或 Codex，现在就去更新。可以通过应用内更新，也可以去官网重新下载。暂时不用担心数据泄露，但拖着不更新，万一后续出现仿冒应用，旧版本的签名验证可能挡不住。

译OpenAI确认受第三方库Axios安全漏洞波及，属供应链安全事件。官方称暂无证据表明用户数据被访问或系统遭入侵。出于预防，OpenAI正更新macOS应用签名证书，要求所有macOS用户立即将ChatGPT桌面端、Codex应用等升级至最新版本，以防攻击者分发伪装应用。用户可通过应用内更新或官网下载安装，暂无需担心数据泄露，但需及时更新以防范潜在风险。

OpenAI@OpenAI · 4月11日

We recently identified a security issue involving the third-party developer library Axios that was part of a broader industry incident. We found no evidence that OpenAI user data was accessed, that our systems were compromised, or that our software was altered. Out of an abundance of caution we are taking steps to protect the process that certifies our macOS applications are legitimate OpenAI apps. We are updating our security certifications, which will require all macOS users to update their OpenAI apps to the latest versions. This helps prevent any risk—however unlikely—of someone attempting to distribute a fake app that appears to be from OpenAI. You can update safely through an in-app update or at the official links below. 🧵

译OpenAI 披露其使用的第三方开发库 Axios 存在安全漏洞，该问题属于近期行业安全事件的一部分。官方强调目前没有证据表明用户数据遭窃取、内部系统被入侵或软件遭到篡改。出于预防考虑，OpenAI 正在更新 macOS 应用的数字签名认证机制，要求所有 macOS 用户尽快将应用升级至最新版本，以防范潜在的假冒应用风险。用户可通过应用内更新或官网链接完成升级。

Rohan Paul@rohanpaul_ai · 4月11日

People using AI for Premier League bets are losing badly. A new betting benchmark suggests today’s best AI models still unravel when prediction has to survive a whole season. In KellyBench, every tested model lost money, and some went completely bust. KellyBench forced agents through a changing 100-150 matchday season where they had to predict outcomes, size bets, and protect a £100,000 bankroll. That setup tests something normal benchmarks miss: whether an LLM can stay coherent, adapt to new data, and manage risk over time. Claude Opus 4.6 was best at -11% ROI, GPT-5.4 came next at -13.6%, and several models hit -100%.

译KellyBench基准测试检验了主流LLM在英超赛季投注中的长期预测与风险管理能力。所有参测模型均遭遇亏损，部分资金归零。Claude Opus 4.6以-11% ROI表现最佳，GPT-5.4为-13.6%。该测试通过100-150场动态赛季模拟，暴露出现有AI在持续决策中的连贯性、数据适应性与风险控制方面存在显著缺陷。

Sam Altman@sama · 4月11日

I wrote this early this morning and I wasn't sure if I would actually publish it, but here it is: https://blog.samaltman.com/2279512

译Sam Altman 今晨发布博客文章，透露该文于凌晨撰写完成，当时曾犹豫是否公开，最终决定发布至其个人博客。

Rohan Paul@rohanpaul_ai · 4月11日

Sam Altman: "In the next year, we will see significant threats we have to mitigate from cyber, and these models are already quite capable and will get much more capable. The needs for society to be resilient to terrorist groups using these models to try to create novel pathogens is no longer a theoretical thing, or it's not going to be for much longer." Overall, his outlook is pretty brutal: a massive cyberattack could land within 12 months, AI bioterrorism is moving from theory to reality, and the only fix he sees would force a full rewrite of capitalism, which Washington is clearly not ready to touch. --- From video from Axios YT channel (link in comment)

译Sam Altman发出严峻警告：未来12个月内或遭遇大规模网络攻击，AI生物恐怖主义正从理论变为现实。随着AI模型能力急剧提升，恐怖组织利用其开发新型病原体的风险已迫在眉睫。Altman指出，应对这些威胁需要彻底重构资本主义体系，但Washington显然尚未准备好接受这种根本性变革。

Chubby♨️@kimmonismus · 4月11日

Regardless of whether Mythos lives up to the hype, they've mastered PR. OpenAI would have liked to see the attention Claude Mythos is getting for "spud," especially with the upcoming IPO.

译Claude Mythos 凭借 "spud" 功能获得极高关注度，公关策略堪称典范。OpenAI 在即将 IPO 之际想必很羡慕这种舆论热度，无论 Mythos 最终是否名副其实。

Deedy@deedydas · 4月11日

Oh, how the tables have turned. OpenAI / ChatGPT is now running sponsored search ads on Claude keywords.

译OpenAI 与 ChatGPT 开始在搜索引擎针对 Claude 关键词投放赞助广告。昔日被拿来与 Claude 比较的 ChatGPT，如今也要购买竞品关键词广告位，形势彻底逆转。

OpenAI Developers@OpenAIDevs · 4月11日

Builders Unscripted with @ashebytes Ashe talked to @romainhuet about moving fast with Codex and why AI should make us more present and connected. Timestamps: 00:10 Early agents and Hearth AI 05:43 AI and human connection 08:44 Build in public 12:17 Ashe’s agentic workflow

译Ashe 与 Romain Huet 对谈，探讨如何用 Codex 快速开发，以及 AI 应如何让人更专注当下、促进真实连接。话题涵盖早期智能体项目 Hearth AI、公开构建实践及其智能体工作流。

Tibo@thsottiaux · 4月10日

I get so many personal thank you emails for something and it’s not even a feature. All we needed apparently was a new codex plan. Can we rest now?

译推出新 codex 计划后收到大量用户感谢邮件，作者吐槽这甚至不算功能，并反问现在能否就此收工休息。