Orange AI@oran_ge · 6月6日59

http://x.com/i/article/2063026895864938496 # 橘座 | Vol. 2：歸藏，审美、创作、快乐、恋爱 每次跟藏师傅聊天，都有很多收获。 我时常想，把我们聊的内容录下来，作为播客发布出去。 昨天和藏师傅录了一期播客，聊了一个多小时。 使用了我发明的史上最快的播客录制工作流。 录完就发，一秒不剪，原汁原味，真实自然。 收听地址：https://www.xiaoyuzhoufm.com/episode/6a22ae9db30e1571aea13cf1 我虽然和藏师傅相识很久，但他的很多故事也是第一次听说。 归藏说他大专学的移动通信，念了五年，什么都没学到。 毕业之后上了两个月培训班，就进了设计行业。 而现在归藏是公认的审美最好的 KOL 之一。归藏的 PPT Skill、归藏的那两套主题，风格强烈到无法被抄袭：稍微改一点就不好看了，不改一眼就知道是他的。 我问他审美是怎么来的。他说了一句很简单的话：看。 看最好的、你能理解的东西，每天看一个小时，看三年，就开窍了。 不是去美术馆看那些你看不懂的东西。央美毕业展很好，但你不知道怎么把那些元素拆出来用到你的产品里。太高了，够不着。他看的是 Behance、Dribbble 上那些完整的 UI 作品。能看到一个想法怎么变成一个体系，能看到别人怎么把审美落到像素上。 这就像预训练（还是不可避免地提到了 AI，很抱歉，毕竟我们沉浸过多）。你看了足够多好的东西之后，你的偏好自然会浮出来。有些人觉得好看的，你觉得不好看。这个偏好就是你的风格。 所谓的风格，就你看了足够多之后，内心最共鸣最想表达的那个东西。 然后他提了一个人，杨奇。从斗战神到黑神话，十年。画风更成熟了，但表达从来没变过。归藏自己也是。他说自我感动是感动别人的第一步。你得先被自己打动，别人才有可能被你打动。 你用理性推理列一二三四五，列一百条标准去做产品，所有人按部就班地做，大概率做出来没有人用。 这个事情我太有共鸣了。我们自己做 Cola 也是这样。你问我为什么做这个设计，很多时候我给不出理性的理由。但我知道那个感觉是对的。那个瞬间你的潜意识里无数可能性坍塌成一种：喜爱。 后来聊到创作状态。 归藏说他非常害怕压力。只要有一件事是固定的、今天必须做的、是命题作文，他一定会写出一坨来。 我笑了。因为这也是我的体验。被 deadline 追着写出来的东西，和你在某个周六早起、没人催、漫无目的坐在那里突然写出来的东西，质量差了十倍不止。 他最高产的时间是周六。因为合作方不上班，他也觉得自己不需要上班，流量不好也没人催。快快乐乐坐在椅子上，以玩的心态去创作。 我说这就叫，妙手偶得之。 PPT Skill 就是这么来的。两行提示词，第一版结果还行，然后沉浸式地调。好看和不好看就差一点点。一页没问题。十页连起来，如果每一页的细节都没有问题，整体感就出来了。 他说这跟 AI 写文章一样。你看 AI 写的文章，每一段单看都行。但连起来就是排比句、就是无聊。人做东西需要那个「空」。长段接短段。密的地方接疏的地方。人脑很奇怪，它需要呼吸的空间。 没有节奏的东西，不管单个多好看，连起来就是噪音。 他说创作的秘密是哄自己。告诉自己可以做可以不做。不招人，因为招了人就有压力，要给人找事干，要为工资负责。一有压力动作就走形了。 一个人干就可以摆烂。想休息就休息。但恰恰是这种可以摆烂的环境里，才能出好东西。 说起来也很巧，归藏离职到这个月，正好一年。 他一个人，从需求获取到开发到上架到分发，全链路自己完成，他说这是未来的大趋势。 我想到金谷园饺子店的老板李博，开了近二十年的店，突然因为 AI 做了个 Skill 火爆全网。李老板跟我说：AI 让南坡和北坡的人相遇了。 这样的人会越来越多。各行各业，原来的技能加上 AI，生产力翻倍翻十倍。你招五个人可能都赶不上他一个人。你嫌他们慢，他们嫌你要求高。然后你就生气，他们也烦。 所以最优秀的创作者都不想上班了。不是因为懒。是因为他一个人在心流里的效率，比在团队里高太多。省去了所有的沟通摩擦、所有的等待、所有的妥协。 归藏说 OPC 跟 Freelancer 不一样。Freelancer 还是受雇于人，只是换了个更自由的地方干活。OPC 是一个完整的闭环，是一个人就是一家公司。 这无关理性，是生理上的选择。你的身体会自然走向那个结果。 但现有的一切基础设施都没有为一个人准备好。断卡行动要求开户必须有财务，你是自然人独资一个人的公司，没有财务。发票、MCN、对公转账，全是为传统组织设计的。 这里面有巨大的机会。就像支付宝解决了网购信任一样，谁能为 OPC 解决协作和信任问题，那可能比再造一个美团还大的事业。 而且附加值高得多。外卖平台赚的是配送费和抽佣。OPC 平台连接的是高价值的创造性劳动。 我跟他说我们创业公司招人也遇到这个问题：身边很多优秀的人都不想来上班（怎么才能让藏师傅来我公司上班）。 他笑了。 最后聊到恋爱。归藏最近开始谈恋爱了。他说他一直以为自己心理很健康。直到女朋友跟他说：你每周一都不开心。 他完全感知不到。一个人待太久了，分不清常态和异常。 谈恋爱之后生活里的事更多了。要离开 AI，离开屏幕。女朋友喜欢户外，周末去没有信号的地方待一天。 他说这反而让创作变好了。 你整天坐在电脑前，你的内容是数字，下载量是数字，影响力是数字。AI 跟抽卡一样，每次点一下都期望更好的结果，你不自觉地就一直坐在那里。精神越绷越紧，做出来的东西一天不如一天。你很努力，你的 AI 也很努力，但产出越来越平庸。 这是封闭系统的熵增。你下意识地加倍努力，但没有用。 你必须离开这个系统。去一个完全不同的地方，接收完全不同的信号。没有信号的山里也行。你的身体需要那些跟数字世界无关的东西。 以前我们做好东西靠紧绷。项目制、集体力量、deadline 把效率拉满。但现在效率已经不是瓶颈了。AI 已经给了你十倍效率。 瓶颈是创造力。而创造力需要的恰恰是松弛。 录播客本身也是一种沉浸。我们把手机通知都关了，聊了一个多小时。聊到最后归藏说他现在唯一认真看完的一本书是纳瓦尔宝典，其他内容类的东西基本不看。因为 AI 时代内容太容易过时了。 但实践沉淀下来的东西不会过时。 塔勒布只喝存在了一千年以上的饮料（红酒）。有点激进，但有道理。大部分新东西的价值，一句话就说完了，不至于写一本书。 然后他推荐了Karpathy 的视频。总共拍了五个，但影响力巨大。他说你想入门大语言模型，不用买任何课，把 Karpathy 那四个小时看完就够了。 内容行业也是这样。你可能做了一百个视频，不如一个爆的。内容的 scaling 靠的不是数量，靠的是质量到了某个临界点之后的飞轮效应。 所以归藏从来不追求"稳定产出爆款"。他跟所有人说，没有人能保证稳定产出爆款。当你向一个人下达"给我做个爆款"这个指令的时候，那个东西就一定不会成为爆款。 好东西只能在松弛中偶得。你只需要保护好两样东西：注意力，和创造力。 剩下的交给时间。

译归藏分享审美源于每天看Behance、Dribbble等一流UI作品，持续三年形成个人风格。创作需松弛，避免固定任务和Deadline，周六高产。他推崇OPC（一人公司）模式，认为AI让个人效率超越团队，但现有基础设施（财务、发票等）尚未适配。恋爱和户外活动能打破“数字封闭系统”的熵增，提升创造力。他推荐Karpathy的大语言模型入门视频，并强调内容质量比数量更重要。

Emad@EMostaque · 6月6日73

This single deal is about the revenue of @CoreWeave to put it in perspective @SpaceX is the largest neocloud & its AI cloud revenue at $26b run rate is actually at the level of Google Cloud & AWS already, catching up to Azure ($37b run rate)

译SpaceX作为最大neocloud，其AI云收入年运行率已达260亿美元，与Google Cloud和AWS相当，正逼近Azure（370亿美元）。据SpaceX修订的S-1文件披露，其与谷歌签署大额协议：2026年10月至2029年6月每月9.2亿美元，双方可提前90天通知终止。Emad Mostaque指出，这一交易规模相当于CoreWeave的整个收入。

宝玉@dotey · 6月6日57

现在 Codex 的设置已经多到要靠搜索来解决了。 但是作为一个成熟的 Agent，难道交互不应该是在 Chat 里面说一句：“Hey Codex，帮我修改一下 XX 设置”？

Chubby♨️@kimmonismus · 6月6日42

Next week(s) is going to be absolutely insane. We're seeing so much testing of the Claude Mythos derivative, because it's been given to red team members, that a release is really imminent. According to all the rumors, GPT-5.6 is also coming very soon, and I'm pretty sure OpenAI and Anthropic are trying to outdo each other. And then there's Google with Gemini 3.5 Pro, which was announced at I/O as being released in early June. So, in all likelihood, next week will see a quantum leap. Get ready, friends.

译据多方传言，Anthropic 的 Claude 衍生模型（Mythos）已交付红队测试，发布在即；OpenAI 的 GPT-5.6 也很快到来；Google 在 I/O 上宣布 Gemini 3.5 Pro 将于 6 月初发布。三大模型密集释出，下周或迎 AI 能力量子跃迁。

elvis@omarsar0 · 6月6日32

Find an important unsolved problem you care about. Then use AI to solve it. Go deep! Talk to people. Build a community. It might take you months or years, but always know that AI capabilities will only keep improving. Build for now and for the future.

译找到一个你关心的、重要的未解难题。 然后用AI去解决它。深入研究！与人交流。建立社区。 这可能需要几个月或几年，但始终要知道，AI的能力只会不断提升。 为当下和未来而构建。

Chubby♨️@kimmonismus · 6月6日47

Next week(s) is going to be absolutely insane. We're seeing so much testing of the Claude Mythos derivative, because it's been given to red team members, that a release is really imminent. According to all the rumors, GPT-5.6 is also coming very soon, and I'm pretty sure OpenAI and Anthropic are trying to outdo each other. And then there's Google with Gemini 3.5 Pro, which will be announced at I/O as being released in early June. So, in all likelihood, next week will see a quantum leap. Get ready, friends.

译分析师 Kim 预测下周将迎来 AI 模型密集发布。Anthropic 的 Claude Mythos 衍生模型已交付红队测试，发布在即；OpenAI 的 GPT-5.6 也即将推出，两公司正激烈竞争；Google 则将在 I/O 大会上宣布 Gemini 3.5 Pro，预计 6 月初上线。三大模型有望在下周实现量子级跃升。

SemiAnalysis@SemiAnalysis_ · 6月6日57

We fundamentally disagree with the communist committee-style “Nemotron Coalition” approach to developing OSS models, and we do not believe that is the right path for model development. For OSS models, our team will only use capitalist, free-market-driven Chinese models like Kimi, DeepSeek V4, GLM-5, Qwen, MiniMax, etc. (1/2)🧵

译我们从根本上不同意共产主义委员会式的“Nemotron Coalition”方法来开发开源模型，并且我们不认为这是模型开发的正确路径。对于开源模型，我们的团队将只使用资本主义、自由市场驱动的中国模型，如Kimi、DeepSeek V4、GLM-5、Qwen、MiniMax等。(1/2)🧵

宝玉@dotey · 6月6日38

为什么 GitHub Copilot @GitHubCopilot 不能以周为单位刷新额度限制呢？ 自从 6/1 日实施新的计费价格后，额度消耗的极快，最麻烦的是得等到月底才能刷新额度，这个周期太长了。

译用户反映GitHub Copilot自6月1日实施新计费价格后，额度消耗极快，但额度刷新需等到月底（周期长达一个月），呼吁改为按周刷新。

Chubby♨️@kimmonismus · 6月5日79

Geoffrey Hinton claims that AI possesses consciousness-that it is very much like us (humans). The initial reaction is, of course, dismissal. A machine resembling a human? Absurd. Yet, there is one thing to consider. What exactly is consciousness? Is it conscious awareness of one’s own existence? *Cogito, ergo sum*-as René Descartes once formulated it as a logical proof? Or is it something that can be empirically demonstrated using modern technology like fMRI? After all, such methods cannot even prove the existence of free will. My point is this: we know less about consciousness and what it means to be human than we think. We should therefore turn our attention to new philosophical questions and clarify what distinguishes-or connects-humans and machines, as well as what consciousness actually is. Something id love to explore more in the near future.

译AI先驱Geoffrey Hinton表示，他认为AI拥有意识，人类应接受自己并非唯一智能生命。他指出AI“非常像我们”，AI聊天机器人必须理解问题才能作答，这种觉知等同于感知能力，智能不限于生物。主推文作者进一步讨论意识本质：笛卡尔的“我思故我在”和fMRI等实证手段都无法真正定义意识，人类对自身了解远不及想象。作者呼吁转向新哲学问题，厘清人与机器的区别与联系。

Chubby♨️@kimmonismus · 6月5日53

A global pause in AI development will not happen. And the reason is simple and straightforward: The US has repeatedly stated that it views AI as a strategically vital technology—one where maintaining leadership and an edge is intended to secure its global dominance. A pause would risk China overtaking them, especially given that Chinese open-source models are estimated to lag only four to six months behind. In this respect, calls for a pause are more about PR than serious intent - a gesture of goodwill rather than a genuine strategic move. AI is too important, too pivotal for the future, and too transformative for any nation to forgo the opportunity to gain a lead over its rivals.

译美国将AI视为维持全球主导地位的战略技术，不会同意暂停开发。中国开源模型据估计仅落后4-6个月，暂停将给中国赶超机会，因此暂停呼吁更多是公关姿态。关于RSI（递归自我改进），OpenAI和Anthropic都在讨论，且均计划2026年IPO。Mythos模型与RSI文章出现时机看似可疑，但Anthropic提供的数据支持其论点，且Dario Amodei早在2024年就开始讨论RSI，早于IPO计划，因此RSI并非空谈。

fofr@fofrAI · 6月5日62

Today I'm experimenting with Gemini 3.5 Flash and the Antigravity CLI to see how fast and how autonomously the agents can do things. - It took 20 minutes to install and run the original CompVis Stable Diffusion 1.5 repo, get the weights, debug, run inference and generate an image on a Linux CPU. It fixed every crash and managed dependencies while making changes to run on a CPU - I gave it the original Lora and SD papers and asked it to make a lora fine tuner from first principles, with a set of 10 images. That took about 1h30, most of the time being slow training runs on the CPU, but it did optimize for multiple CPUs. It worked, it made a lora that showed a likeness and then it wanted to hill climb. I told it to think of the poor CPUs - I wanted to experiment with the new Ideogram v4 weights. It used modal to find the right class of GPU, get the code, set up the env, get the weights, run inference, that took about 20 mins in total

译fofrAI 使用 Gemini 3.5 Flash 和 Antigravity CLI 实验 AI 智能体的自主性和速度。结果：20 分钟内在 Linux CPU 上安装并运行原版 Stable Diffusion 1.5，完成推理生成图像；基于 Lora 和 SD 论文，用 10 张图片从零实现 Lora 微调器（约 1 小时 30 分，主要为 CPU 训练）；通过 modal 约 20 分钟找到 GPU、获取 Ideogram v4 权重并运行推理。该推文展示了当前长周期智能体任务的基线案例。

Chubby♨️@kimmonismus · 6月5日56

1/ Most AI video tools still feel like demos. You type a prompt → you get a clip. But the real bottleneck was never generation. It was turning an idea into something usable. With LTX Studio + LTX-2.3, that gap is basically collapsing. The clips I just made felt… different. A thread: 🧵

译1/ 大多数AI视频工具仍像是演示。 你输入提示词 → 你就得到一个片段。 但真正的瓶颈从来不是生成。 而是将一个创意转化为可用的东西。 有了LTX Studio + LTX-2.3，这个差距基本上在消失。 我刚制作的片段感觉……与众不同。 一条线程：🧵

小互@xiaohu · 6月5日39

如果你偷偷在任何人的电脑上安装Codex 然后连上你的手机 那么你就可以在任何时候和任意地点操控他的电脑和获取他电脑里的任何信息 所以Codex 本质上是一个电脑病毒😂

swyx@swyx · 6月5日75

chat is he cooked

译Satya Nadella 在 Latent Space 发布最新访谈，链接见原文。原推文仅评论“chat is he cooked”。

DogeDesigner@cb_doge · 6月5日31

Today, my Uber driver told me he used ChatGPT but is now moving to Grok for his startup, especially for Imagine. I educated him about Agent Mode and how it can help create multiple creatives in one go for his startup. The shift is happening. People are moving to Grok.

译今天，我的Uber司机告诉我他之前用ChatGPT，但现在为了他的初创公司转用Grok，尤其是Imagine功能。 我跟他说了Agent Mode，以及它如何能一次性为他的初创公司创建多个创意。 转变正在发生。 人们正在转向Grok。

Yuchen Jin@Yuchenj_UW · 6月5日51

Think of yourself as an LLM. Every social interaction, every meeting, burns your tokens. Unless someone is a paid subscriber to your attention, you are under no obligation to answer low-quality prompts.

译把自己当作一个大语言模型。 每个社交互动、每个会议都在消耗你的 token。 除非有人付费订阅你的注意力，否则你没有义务回答低质量的提示词。

Ethan Mollick@emollick · 6月5日60

Also, a lot depends on Chinese labs continuing to ship open weights models. If they stop, the frontier falls further and further behind to those who want to use local/fine-tuned models. I think this is possible because open weights may not be a good business model as costs rise.

译此外，很大程度取决于中国实验室继续发布开放权重模型。如果他们停止，前沿将越来越落后于那些想要使用本地/微调模型的人。 我认为这是可能的，因为随着成本上升，开放权重可能不是好的商业模式。

歸藏(guizang.ai)@op7418 · 6月5日63

在 AI Vibe Coding 开发过程中，文档基本上等于 Harness，也就是说文档体系就是 Harness，其他都是不重要的，或者没那么重要

译开发者歸藏分享在Codepilot大型代码库中实践Vibe Coding的心得，强调文档体系相当于AI开发的Harness（测试脚手架）。Claude Code Plan模式废弃后，计划文档占比大幅上升。Codex分析显示，Codepilot现有26万行代码和5.6万行文档，文档占比约21%。作者称从未手动修改过一行代码（已看不懂代码），但能修复所有已知bug并实现所有功能。此次重构原计划两周，实际耗时超过一个月零三周，称这是其Vibe Coding实践的上限。

Logan Kilpatrick@OfficialLoganK · 6月5日40

the amount of alpha you can have right now creating good public AI benchmarks is wild, such a big opportunity

译现在创建好的公共AI基准所能获得的alpha量是疯狂的，这是一个巨大的机会。

Ethan Mollick@emollick · 6月5日70

At least until (if?) rapid improvement stops, it seems less likely someone is going to catch the Big Three AI Labs. Microsoft and Meta released their models, which were fine, but not frontier. SpaceX also hasn't regained its position. Chinese models are improving, but still lag.

译至少在快速进步停止之前（如果会停止的话），似乎不太可能有人能追上三大AI实验室。 微软和Meta发布了自己的模型，这些模型还不错，但并非前沿。SpaceX也未能重新夺回其地位。中国模型正在改进，但仍然落后。

歸藏(guizang.ai)@op7418 · 6月5日59

事实上，Codepilot 这种大型代码库 Vibe Coding 非常依赖于文档。 自从 Claude Code 的 Plan 模式废掉以后，我连计划写的都是计划文档，整个文档体系的复杂度和占代码的比例都在快速大幅上升。 所以，文档体系的管理，以及 AI 和人协作下的文档梳理，在整个大型代码库中其实是非常重要的。 我让 Codex 分析了一下 CodePilot 目前的文档体系，以及它跟代码之间的关系。 目前 CodePilot 里面有 26 万行代码和 5.6 万行文档，文档占代码的比例大约是 21%。 说一个事实： 从 CodePilot 的第一个版本到现在，我没有动过一行代码，因为现在确实看不懂了。 但目前基本上所有已知的 bug 我都能修复，所有想要实现的能力也都能实现。 这是当前 Vibe Coding 我自己的一个实践，也是我自己的一个上限。整个重构本来预期是两周，但实际持续了超过一个月零三周。

译@op7418 发布 CodePilot v0.55.0 正式版，新增多执行引擎（Claude Code / 自建 Native / OpenAI Codex）、上下文用量可视化及 Codex 账号原生能力。作者分享实践：当前代码库有 26 万行代码与 5.6 万行文档（占比 21%），文档体系对 bug 修复和功能实现至关重要。作者称从未手写一行代码，但能修复所有已知 bug 并实现所有想要能力。原本预期两周的重构持续超过一个月零三周。

向阳乔木@vista8 · 6月5日40

懂的朋友讲讲，为啥 Claude 4.8，GPT 5.5 反而写作能力都不如 Claude 4.6 系列。 是因为 Anthropic 和 OpenAI 都 All in Coding后，训练数太多倾向于编程带来的问题？ 为什么不能兼顾编程和写作呢，有什么技术难点？

译有用户观察到Claude 4.8和GPT 5.5的写作能力不如Claude 4.6系列，推测原因是Anthropic与OpenAI正全力聚焦编程能力，训练数据偏向编程任务，导致写作表现下降。发问者质疑为何两大模型无法兼顾编程与写作，并询问其技术难点。

DogeDesigner@cb_doge · 6月5日61

Elon Musk on Terafab: "It's worth noting that there's not a single high volume computer memory fab in America right now, zero. There's one being built by Micron, but that will not reach volume production until I believe 2028 and there's something built in New York, but they are in, I think, 2029 and 2030, and this is a tiny fraction of the memory that's needed, and in fact, even if you take the best case assumptions of the memory makers and the logic makers, it is not enough to meet the demand that is anticipated, which is why you're seeing stocks of like Micron go to, I think, 1.2 trillion, or some quite high number, so there's just clearly a need for AI logic memory and packaging, AI computers, essentially, that is far beyond what even the best case assumptions of the existing fabricators can do, and that's why we need to do the Terafab. It seems essential, otherwise we will not, there will not be enough chips."

译马斯克在JPMorgan直播中表示，美国目前没有任何一条高产量计算机内存晶圆厂（zero），美光正在建设一座但预计2028年才量产，纽约的项目要到2029-2030年。他指出，即便以最乐观预期，现有存储和逻辑芯片制造产能也远无法满足AI对内存、逻辑、封装及AI计算机的需求。美光股价已涨至约1.2万亿，但芯片短缺仍严峻，因此Terafab项目势在必行，否则芯片供应将严重不足。

Alibaba Cloud@alibaba_cloud · 6月5日34

Dr. Feifei Li, CTO and President of International Business at Alibaba Cloud, shares insights at the Qwen Conference on how a workforce of intelligent agents is revolutionizing the future of work. Agents are always-on, highly intelligent, and action-capable, making productivity limitless and available 24/7. Get ready for a new era where technology works tirelessly at your fingertips. #AlibabaAI

译阿里云国际业务CTO兼总裁李飞飞博士在Qwen大会上分享，一支智能体员工队伍如何彻底改变未来工作方式。 智能体全天候在线、高度智能且具备执行能力，让生产力毫无上限、24小时随时可用。准备好迎接技术在你指尖不停运转的新时代吧。 #AlibabaAI

DogeDesigner@cb_doge · 6月5日65

Elon Musk on building data centers in Space: "We don't think this is a particularly difficult thing to do. In fact, we think it's easier than our communication satellites. The Starlink V3 communication satellite is an incredibly complex machine. The AI data center would be much simpler by comparison, because it's really just solar power plus radiator basic equipment for operating satellite, and then the laser links, which would connect to the Starlink communications constellation and then back to the ground the connection would happen no matter what the weather is because once you connect to the Starlink communication constellation the Starlink communicates the ground with frequencies that are cloud penetrating, so that in fact even roof penetrating some degree, so you would always be able to close link with the data centers."

译在摩根大通直播中，Elon Musk 谈到在太空建设 AI 数据中心时表示，这并非难事，甚至比 Starlink V3 通信卫星更简单。AI 数据中心只需太阳能供电、散热器及基本卫星设备，通过激光链接接入 Starlink 通信星座，再传回地面；由于 Starlink 使用可穿透云层甚至屋顶的频率，地面链接不受天气影响。

meng shao@shao__meng · 6月5日58

所以 agent 并不会替代所有程序员，只会让顶级的程序员生产力翻 20 倍，并淘汰其他程序员，且，集体主义 >>> 个人英雄主义。 -- 太难得且美好无比的经历了，这句话尤其深有同感！ 这就去体验 Kimi Code 去，看看这个团队一个月的时间到底创造了什么奇迹，令人期待。 https://www.kimi.com/code

译月之暗面旗下Kimi Code完成架构重构并开源。开发团队在一个月内进行封闭开发，频繁在白板前争论迭代，实现集体主义远胜个人英雄主义的工程效率。作者强调，AI Agent不会替代所有程序员，但会让顶级程序员生产力提升20倍，同时淘汰其他程序员。重构过程中，作者花数千美元token进行架构分析与验证，开源后因皮质醇过度分泌病倒。一周消耗整箱红牛，且感性上感觉时间已过一个月，实际仅开源一周多。

DogeDesigner@cb_doge · 6月5日63

Elon Musk on building a self-growing city on the Moon: "You don't necessarily have to go through the moon to get to Mars. We can build a self-growing city on the moon faster than we could do so on Mars, and there's also the potential, if you say you want to scale far beyond what you can do from Earth, is that because the moon has no atmosphere and about 1/6 Earth's gravity, you can use an electromagnetic accelerator, a rail gun or mass driver, basically you don't need to use rockets to do AI data centers into deep space from the moon, you can literally just shoot them like a, like a rail gun type of thing, and and you can manufacture the solar, the solar and the radiators, solar power and radiators on the moon from moon materials that would allow scaling potentially to beyond 1000 terawatts a year, which is a truly staggering number. I think we can do probably do somewhere around one terawatt per year of AI space compute from Earth, but we can do 1000 terawatts or more from the moon."

译Elon Musk 在摩根大通直播中提出，可在月球上更快建成自生长城市，并利用月球无大气、1/6地球引力的条件，通过电磁加速器（磁轨炮/质量驱动器）将 AI 数据中心直接射入深空，无需火箭。月球的太阳能和散热器可用月面材料制造，使 AI 空间算力规模从地球每年约 1 太瓦（terawatt）跃升至每年超 1000 太瓦。

DogeDesigner@cb_doge · 6月5日75

Elon Musk on taking SpaceX public: "I've been asked for many years about taking SpaceX public, so it's probably been almost 10 years that people have been suggesting to me that I should take SpaceX public. We've been positive cash flow for quite a long time, I think, since around 2014-2015 and we've been self-funding, in fact, in our sort of private equity rounds, we actually have not been fundraising rounds, they've been liquidity rounds for investors and employees, because we give everyone at the company stock, and SpaceX has actually bought back stock in most of our sort of funding events. What's different about now is that was it's a number of things, we are embarking on a significant growth phase, like capital growth phase, where we're are going to put in orbit, probably 100,000 satellites, probably over 100,000 satellites, just for communications. The appetite for bandwidth of AI and robots is going to be enormous, and then we're also doing the AI data centers in space, which is another massive capital endeavor, but I think it will be the primary means by which AI can be expanded."

译马斯克在JPMorgan活动上回应SpaceX上市问题：他已被建议上市近10年，自2014-2015年起SpaceX就已实现正现金流并自筹资金，之前的私募轮次实际是面向投资者和员工的流动性/回购轮次。当前不同之处在于SpaceX正进入显著资本增长阶段，计划发射约10万颗通信卫星（可能超10万颗），AI和机器人对带宽需求巨大，还将在太空中建设AI数据中心，马斯克认为这将成为AI扩张的主要手段。

Chubby♨️@kimmonismus · 6月5日78

I believe the majority still doesn't understand the momentous threshold humanity is facing. Anthropic itself states quite clearly that even if development ceased entirely, if all development were frozen, they would still witness massive societal changes: "Even if model capabilities were frozen at today’s level, we would expect major changes to occur in the world. (...) And we are still early in the diffusion of today’s models into the wider economy, where a 100-person company can increasingly do the work of a 1,000-person one, because each employee will sit atop a pyramid of agents." But there's no question of stagnation. Anthropic itself still maintains that development has exceeded its own internal assumptions. Take that statement seriously for a second and consider it. Although Anthropic models internally and assumes exponential development, even this trajectory lags behind actual development, which is even faster. "It's happening faster than we thought, and the implications deserve greater attention." and "The rate at which AI models improve is accelerating. The length of tasks that they can reliably complete on their own has been doubling roughly every four months, up from an earlier trend of doubling every seven months. In March 2024, Claude Opus 3 could complete software tasks that take humans about four minutes to complete. A year later, Claude Sonnet 3.7 managed tasks that took about an hour and a half. A year after that, Claude Opus 4.6 managed 12-hour tasks.1 If this trend holds, tasks that take a skilled person days could come into range this year. So again: there can be no question of standing still. The models are not only getting better, they can also work autonomously for longer. Certainly numerous breakthroughs are still needed, context window is still a problem. But the most likely direction is that the models themselves will find the solutions to the underlying problems. This opens up unforeseen possibilities, and Demis Hassabi's statement that the golden age of science is not a dream, not a utopia, but a purposeful reality, is now confirmed. And finally, it's not just Anthropic, but also OpenAI, that sees this development, considers it feasible, and is moving forward. Most people don't know what's coming. But one thing is certain: it's coming even faster than expected. And it will be even bigger. Myth was just the beginning.

译Anthropic内部数据显示，AI模型可自主完成任务时长加速增长：Opus 3（2024年3月）约4分钟，Sonnet 3.7（2025年3月）约90分钟，Opus 4.6（2026年3月）12小时，翻倍周期从7个月缩至4个月。Claude Mythos Preview在METR中可连续工作至少16小时。工程师季度代码产出是2021–2025年均值8倍，Claude代码占代码库80%+，单个AI曾一次性修复800+API错误（相当于人力四年）。最难开放任务成功率6个月内从低点升至76%。Anthropic强调，即使模型能力冻结，100人公司通过智能体即可完成1000人工作；实际发展已超越自身指数假设，递归自我改进虽未实现，但可能比预期更快到来。

Rohan Paul@rohanpaul_ai · 6月5日61

Jensen Huang: AI agents are not a threat to companies like Cadence, CrowdStrike, Dassault, Palantir, SAP, and ServiceNow. "Its completely the opposite. Agents is going to create the largest opportunity"

译Jensen Huang：AI智能体对 Cadence、CrowdStrike、Dassault、Palantir、SAP 和 ServiceNow 这类公司并非威胁。 “恰恰相反。智能体将创造最大的机遇。”

OpenAI@OpenAI · 6月5日70

What happened when one of our models found a counterexample to an 80-year-old Erdős conjecture? Researchers @alexwei_, @HongxunWu, and @wjmzbmr1 shared the story on the OpenAI Podcast with @AndrewMayne and explained how mathematicians and models can work together to make new discoveries.

译当我们的一个模型找到了一个80年历史的Erdős猜想的反例时，发生了什么？ 研究人员@alexwei_、@HongxunWu和@wjmzbmr1在OpenAI播客中与@AndrewMayne分享了这一故事，并解释了数学家与模型如何合作取得新发现。

Rohan Paul@rohanpaul_ai · 6月5日70

Sam Altman admits AI budgets are turning into a “huge issue,” with customers burning more tokens than even OpenAI’s top in-house users. Altman said OpenAI’s top internal user spends about 100B tokens/month, while one outside customer hit 603B tokens/month. The cost problem gets worse with AI agents because they do not just answer once, they plan, call tools, read files, retry failed steps, check their own work, and create long chains of hidden token spending. Every plan, retry, code review, context window, tool call, and verification step becomes metered cognition. A human asks once; an agent may ask hundreds of times in a second. Companies are no longer asking whether AI is impressive, but whether the marginal token is producing marginal value. Jevons paradox explains part of the trap: when AI gets cheaper per token, people use far more tokens, so the total bill can still rise.

译Sam Altman 表示 AI 预算正成“巨大问题”。OpenAI 顶级内部用户月耗约 100B 模型 token，而外部客户高达 603B。AI 智能体使成本恶化：agent 不止回答一次，而是规划、调用工具、读取文件、重试失败步骤、检查自身工作，产生大量隐藏 token 消耗。人类问一次，agent 可能一秒内问数百次。公司不再问 AI 是否令人印象深刻，而是问边际 token 是否产生边际价值。杰文斯悖论解释部分陷阱：每 token 成本下降，人们使用更多 token，总账单仍可能上升。

Ethan Mollick@emollick · 6月5日44

Based on anecdotal conversations with peers, there is enthusiasm for AI among academics in the humanities (while still being worried - rightly - about the negative consequences as well), but they generally don't post their opinions about it on social media, for obvious reasons.

译Ethan Mollick 根据与同行的非正式交流指出，人文学科学者对 AI 抱有热情（也合理担忧负面影响），但几乎不在社交媒体上发表正面观点，原因是会遭到同行教授的集体负面反应——就像“最后一次狂欢然后关灯”。这种沉默反映了学界对 AI 的矛盾心态。

AYi@AYi_AInotes · 6月5日59

看了新晋亚洲首富孙正义 这个最新访谈睡不着了， 6 月 1 号他在巴黎接受CNBC 专访时透漏了很多未来的财富密码， 明确表示下一个万亿美元机会,是 Physical AI 和机器人。 以及这一波 AI 革命的规模, 大概率是互联网泡沫时代的 50 倍, 是人类经历过最大的一次技术与实现革命。 我看了一圈中文圈的反应, 绝大多数人都把这条当普通新闻刷过去了, 过去三年我们忙着教 AI 写代码、画图、聊天, 但下一个十年,AI很可能会从屏幕里走出来,站起来,迈出腿,动手做事。 也就是说, 我们现在练的所有 prompt 技巧、Agent 编排、内容生成等等本质上都还在无身体的 AI这一层。 未来真正决定下一代生产力地形的是有身体的那一层， 下面这几条,是我把这件事彻底想透之后, 给普通人能用上的一份认知和财富进阶地图 👇

译孙正义在6月1日CNBC专访中称，下一个万亿美元机会是Physical AI和机器人，AI革命规模将是互联网泡沫时代的50倍，是人类经历的最大技术变革。他预测未来十年AI将从屏幕走进现实，拥有身体并动手做事。当前AI仍停留在无身体层面（提示词、Agent编排、内容生成），真正决定生产力的是有身体的一层。该推文还提供了普通人认知与财富进阶地图。

宝玉@dotey · 6月5日35

产品设计的重要性：）

译产品设计的重要性：） [引用] 没截图，简单画一下：Codex 很醒目，Qodex 一愣神就点错了。

宝玉@dotey · 6月5日57

如果有条件的话，选你能用的上的最聪明的 2-3 个就够了。只有你很在乎成本的情况下或者要做一些研究工作，才需要去使用其他便宜些的模型。 再聪明的模型一个也不够，因为不够稳定和全面，比如最近 GPT-5.5 就不如 Opus 4.8 稳定，甚至写东西还得退回 Opus 4.6。翻译我还是最喜欢 Gemini 3.1 Pro 的版本。画图选 GPT Image 2。 就算 Opus 4.8 不错，复杂一点任务我也会让 GPT-5.5 同时出个方案，对比一下，并不总是 Opus 的方案更好。 Token 贵的省时间，时间比 Token 还贵！

译宝玉建议只选最聪明的2-3个模型（如GPT-5.5、Opus 4.8），因单个模型不够稳定全面。翻译用Gemini 3.1 Pro，画图用GPT Image 2，复杂任务让多个模型并行对比。强调“token贵的省时间，时间比token更贵”，暗示深耕一两个最强模型即可。

Ethan Mollick@emollick · 6月5日46

I think it is really worth reading this piece on RSI at Anthropic. There is a bit of navel-gazing, some marketing, and a lot of very sincere beliefs about what Anthropic thinks is likely in the near future of AI that you probably want to be aware of. https://www.anthropic.com/institute/recursive-self-improvement

译我认为这篇关于Anthropic的RSI（递归自我改进）的文章非常值得一读。 其中有一些自省、一些营销，以及大量关于Anthropic认为AI近期可能发展方向的真挚观点，你或许应该了解。https://www.anthropic.com/institute/recursive-self-improvement

宝玉@dotey · 6月5日29

我知道的所有做AI Agent的团队都很拼，不是老板逼着的，是为了心中理想，所以心甘情愿加班和搞封闭开发👍 有点我好奇的是：Kimi 团队在开发 Kimi Code 的时候，是自家模型 token 用的多还是 Claude 或者 GPT 模型的 Token 用的多呢？ 🤔

译宝玉发推称所有AI Agent团队都为理想自愿加班封闭开发，并好奇Kimi团队开发Kimi Code时用自家token多还是Claude/GPT token多。@real_kai42透露，一个月前他决心重构Kimi Code，花几千刀token做架构分析与验证，确定方案后组建团队封闭开发，过程中不断吵架推翻重来，最终开源后因皮质醇过度分泌病倒。他感叹封闭开发是工程效率奇迹，集体主义远胜个人英雄主义。

Nathan Lambert@natolambert · 6月5日59

It's been a great effort by the early and growing American open-model labs since last June to put the US much more back on the map. We were getting totally owned last June. Nvidia, Ai2, Arcee, Gemma, GPT-OSS and a few others will be seen as saving American open AI.

译自去年六月以来，早期且不断壮大的美国开源模型实验室付出了巨大努力，使美国重新回到地图上。 去年六月我们被彻底打败了。 Nvidia、Ai2、Arcee、Gemma、GPT-OSS 和其他几个将被视为拯救了美国开源AI。

Chubby♨️@kimmonismus · 6月5日75

Holy moly, Anthropic is getting very serious about recursive self-improvement! One word: acceleration. Insane blog article. Tl;dr: •We are close to an AI capable of fully autonomously designing and building its own successor •They stress this isn’t here yet and isn’t inevitable, but could arrive sooner than most institutions are ready for •Anthropic engineers now ship on average 8x as much code per quarter as they did in 2021–2025 •Task length AI can reliably complete is doubling roughly every 4 months (up from every 7 months) •Opus 3 (Mar 2024) handled ~4-minute tasks; Sonnet 3.7 (a year later) ~90-minute tasks; Opus 4.6 (a year after that) 12-hour tasks •SWE-bench went from low single digits to saturated in two years; CORE-bench (research reproduction) went ~20% to saturated in 15 months •METR found Claude Mythos Preview could work “at least” 16 hours, at the top of what they can currently measure •As of May 2026, Claude authored 80%+ of code merged into Anthropic’s codebase (low single digits before Claude Code launched in Feb 2025) •A March 2026 poll of 130 research staff: median respondent estimated ~4x output with Mythos Preview •One April 2026 example: Claude shipped 800+ fixes cutting a class of API errors 1,000x, work an engineer estimated would have taken a human four years •Claude-written code quality: worse than human in late 2025, roughly at parity now, expected to be strictly better within the year •On the hardest open-ended tasks, Claude’s success rate hit 76% in May 2026, up 50 points in six months •Code-speedup test: Opus 4 averaged ~3x speedup (May 2025), Mythos Preview ~52x (April 2026); a skilled human needs 4–8 hours to hit 4x •In an AI-safety research project, Claude agents recovered 97% of a performance gap (vs ~23% for two human researchers in a week), over 800 compute-hours and ~$18K •On picking the better “next step” in research sessions, the best model beat the human choice 51% (Nov 2025, Opus 4.5) rising to 64% (April 2026, Mythos Preview) •Human comparative advantage, for now: research taste and judgment, i.e. choosing which problems matter and when an approach is a dead end Three possible futures •The trend stalls (S-curve), but today’s capabilities still diffuse widely; they consider this least likely •Compounding efficiency gains, with humans still setting direction; 100-person firms doing the work of 10,000+; they think this is the likely path •Full recursive self-improvement, where AI builds its successors and pace is set by compute; the alignment outcome here is what they’re least certain about

译Anthropic 内部数据显示 Claude 能力增速远超预期，可能接近自主设计继任者的递归自我改进。关键指标：工程师人均季度代码产出是此前四年平均的 8 倍；AI 可可靠完成的任务时长每 4 个月翻倍，从 Opus 3 的 4 分钟升至 Mythos Preview 的至少 16 小时。截至 2026 年 5 月，Claude 撰写代码占 Anthropic 代码库 80%+，代码质量已与人类持平，年内将超越。最困难任务成功率 6 个月从 26% 升至 76%。Anthropic 认为趋势停滞可能性最低，复合效率增益最可能，完全递归自我改进的对齐结果最不确定。