🐙

🚨OpenAI CFO Says Company Not Ready For IPO 2026, Altman Stopped Inviting Her To Financial Meetings with Investors >be sam altman >commit to $600 billion in cloud spending over 5 years >will burn $200B+ before generating cash revenue >but revenue growth already slowing >cfo tells colleagues she has concerns >says company won't be ready for 2026 ipo >your move? >stop inviting her to financial meetings >change her reporting structure so she no longer reports to you >problem solved CFO Sarah Friar has been telling colleagues OAI won't be ready for 2026 IPO as revenue growth will not support the spending. CEO Sam Altman wants to IPO before Anthropic does. So, he excluded her from compute spending meetings with top investors. Investors said her absence was "noticeable and awkward." OpenAI tried to own data centers via Stargate $500B joint venture with Oracle and SoftBank. Lenders said no, wouldn't back "a company with an unproven business model burning billions annually." So now they are forced to rent instead. Remember the $122B they just raised? They actually only got $25B; the rest is conditioned to "if you IPO" Dario Amodei: "I kind of get the impression that some of the other companies have not written down the spreadsheet, that they don't really understand the risk they're taking." OpenAI might not actually make it to IPO...

🚨 重磅！Google Gemma-4-31B 模型被彻底破解！ Google 最新 Gemma-4-31B 基础模型出现越狱版本，已完全移除安全限制。 模型名称：Gemma-4-31B-JANG_4M-CRACK HarmBench 得分：93.7%（149/159） 采用 18GB 混合精度 MLX 量化（加载约 22.7GB），支持 Apple Silicon，并原生支持视觉多模态。 目前已在 Hugging Face 开放下载： huggingface.co/dealignai/Gemm… 本地无审查强力模型又多一个选择了。 这破解速度真快，可以为所欲为了😘！

太卷了，谷歌凌晨直接王炸。 他们发布了最新开源模型 Gemma 4，并强调其可在个人设备本地运行。 与依赖云端调用不同，Gemma 4 支持在用户本地硬件上运行，同时具备函数调用的能力，意味着模型不仅能对话，还可以直接调用工具、浏览网页、执行代码甚至接入 API，具备 AI 代理的能力。 这一变化的核心在于，推理成本被大幅压缩，甚至接近于零。用户无需再为每次调用付费，只需承担本地硬件成本。 市场观点认为，这将冲击以调用计费为主的 AI 商业模式，尤其是 OpenAI、Anthropic 等依赖 API 收费的厂商。 目前相关中文文档页面，也同步发布了。

🚨BREAKING: The CEO of the most valuable private company in history is being sued by his own sister for s*xual abuse. A judge just allowed the case to move forward. Annie Altman filed an amended lawsuit on April 1, 2026 in St. Louis federal court accusing her brother Sam Altman of sexually abusing and raping her between 1997 and 2006. She says the abuse started when she was 3 years old and he was 12. A judge dismissed the original claims because the statute of limitations expired in 2008. But he allowed her to refile under Missouri's Childhood Sexual Abuse law. She did. The case is now active. Sam Altman denied everything. Called it extortion. Filed a defamation countersuit against his own sister based on social media posts she made between 2021 and 2024. His family says she has mental health challenges. Annie posted videos saying she was "touched by older siblings" and that "an almost tech billionaire" molested her. The judge said those statements make it reasonable to infer she meant Sam Altman. This is the same man whose own board of directors fired him in 2023. Former board member Helen Toner said two senior executives came forward with screenshots and documentation. They used the words "psychological abuse." Said he was lying and manipulating people. Said they had no belief he could change. The board fired him secretly because they knew he'd try to undermine them. He came back five days later. Employees were told either Sam comes back or the company dies. This wasn't his first time. The management team at his first startup Loopt went to the board twice asking to fire him for "deceptive and chaotic behavior." He was reportedly pushed out of Y Combinator too. Every safety leader who clashed with him left. Musk. Sutskever. Amodei. When Jan Leike resigned he said "safety culture has taken a backseat to shiny products." OpenAI was forcing departing employees to sign agreements saying if they ever criticize the company they lose all their equity. Millions of dollars. When it leaked Altman said he didn't know. The CEO didn't know what was in his own exit contracts. He asked Scarlett Johansson to voice ChatGPT. She said no. Twice. They made a voice that sounded just like her anyway. Altman tweeted "her" on launch day. She had to hire lawyers to get them to stop. Forbes says he's worth $3.3 billion. The Musk trial starts April 27. His sister's case is now active. His own board fired him for lying. His own executives called it psychological abuse. And this is the guy the world is trusting with AI.

Vikings in the Mist! 30-second extended video from Grok Imagine Update your Grok app ! It just gets better and better with every update.

Switching to Gemini from other AI apps just got easier. Starting to roll out today on desktop, you can now bring your preferences and chat history into Gemini, so you can pick up right where you left off in just a few clicks. 🧵

@soraofficialapp RIP SORA!!! x.com/evanluthra/sta…

哈哈哈，这不靠谱的公司干什么都不行，都撤了吧，openAI谁沾边儿谁倒霉，替代可以试试seedance

要推销不动了，看看还有没有冤大头上赶着陪葬🤭

@OpenAI OpenAI monitoring everyone’s submissions from Codex logs

On our way to OpenAI!

youtu.be/pFqmqrtOB5w?si… #QuitGPT

关于scam altman 统治世界的计划: ​第一步： 用 OpenAI 研发的 AI 砸掉普通人的饭碗。 ​第二步： 告诉失业的你，别担心，我会发 UBI（基本收入）救济。 ​第三步： 但领救济的前提，是把你的虹膜交出来，换取我的数字 ID。 #OpenAI #SamAltman #QuitGPT

OpenAI 现在的迭代速度比不上 Anthropic，它们已进化成两种不同的生物。 过去7天，Anthropic发布了什么： Dispatch、Channels、Voice mode、/loop、1M上下文GA、MCP elicitation、移动端持久化Cowork、Excel/PowerPoint跨应用上下文、inline charts、64k默认输出token。 每一个功能都是工程师账号发推、GitHub直接release。 同期OpenAI发布了什么： GPT-5.4 mini和nano。重新设计了model picker。砍掉了"Nerdy"人格预设。宣布了三起收购。 这是AI行业最被低估的差距。 Claude Code负责人Boris Cherny每天发10-30个PR，自去年11月起没写过一行手敲代码。团队每天60-100次内部发布。Cowork用Claude Code自己10天搭完。工具在构建下一版工具。 Anthropic的PM不写PRD。原型直接替代文档。工程师有权直接发布和宣布。 OpenAI有个职位叫"CEO of Applications"。这个title存在的原因，是工程师没有权限在没有高管审批链的情况下发布。 一个循环在复利。另一个在开会。🫠

Encyclopedia Britannica and Merriam Webster have just sued Sam Altmans OpenAI. These legendary publishers accuse OpenAI of stealing nearly 100000 copyrighted articles and dictionary entries to train ChatGPT. The AI now copies their content freely while crushing their website traffic and revenue that built centuries of real knowledge. This is not innovation or progress. This is blatant industrial scale theft from the guardians of human knowledge. Sam Altman preaches ethics and safety while building his empire on plagiarism and lies. The lawsuits are only getting started.

@OpenAINewsroom @astral_sh 🌚

Btw, the proceeds of any legal victory in the OpenAI case will be donated to charity. I will in no way enrich myself.

55000人被裁，留下的人却快疯了？哈佛揭穿AI时代的“残酷真相”裁员5万，涨薪56%！

今天科技圈最热门的话题之一： Yann LeCun的AMI Labs融资$10.3亿，欧洲史上最大种子轮，专门挑战GPT/Claude/Gemini的自回归架构 图灵奖得主LeCun离开Meta，用4个月创立了AMI Labs。 他要做的事只有一件：证明整个LLM路线走错了。 LeCun的核心判断：GPT、Claude、Gemini这类模型只是在预测下一个token，本质上是统计相关性，无法真正理解世界。 AMI的架构叫JEPA——不预测文字，而是建立物理世界的内部模型，能理解3D现实、连续信号和因果关系。 同期，李飞飞的World Labs也刚融了$10亿——三周内两位顶级AI科学家合计筹了$20亿，全部押注LLM是死路。 投资方包括NVIDIA、贝索斯、丰田、三星、淡马锡。 他们在等的不是产品，是一个可能颠覆整个行业的架构。