与「dark」相关的搜索结果

Right in the middle of the Conan set, a woman is inverted in a full handstand, Her form is tight and controlled, arms locked, body perfectly vertical against the dark stage. Conan sits a few feet away at his desk, suit jacket on, ocean-night backdrop glowing behind him.

Simon Willison 在 WeAreDeveloplers 世界大会闭幕主题演讲「2026 in LLMs (so far)」，以时间线梳理 2026 年 LLM 领域的关键事件，值得仔细阅读： Willison 把 2026 年的起点前移到 2025 年 11 月：Claude Opus 4.5 和 GPT-5.1 发布。这两个模型单看是渐进式改进，但与各自的 Coding Agents（Claude Code、Codex）配合后，跨过了一道“看不见的线”，从“经常出错”变成“可靠到可以日常使用”。这一质变是全年所有故事的引爆点。 # 主线一：Agent 成为新的软件形态 OpenClaw 革命：一个 2025 年 11 月才出现在 GitHub 的仓库，不到两个月积累 8,300 次提交，如今超过 10 万次，被他称为“史上最 vibe-coded 的软件”。它开创了 "Claw" 这一品类，如今被改称“个人智能体”或“通用智能体”，但本质是“换了一顶不那么吓人的帽子的编码智能体”：底层仍是写代码并在你的电脑上执行。湾区 Mac Mini 因此卖断货（Drew Breunig 的妙喻：买 Mac Mini 是给 Claw 买鱼缸）。 真实需求验证：3 月中国出现 OpenClaw 安装派对，非技术人群排队安装，证明普通用户确实想要一个能替自己办事的智能体。随后行业进入“谁能造出安全的 Claw”竞赛，Meta 的 Muse 目前居 App Store 免费榜首位。 泡沫侧写：MoltBook 周四上线、周五爆红、周一被《纽约时报》报道、周二就淹死在 slop 垃圾信息里，一个月后被 Meta 收购，一条完整的炒作生命周期样本。 # 主线二：开发范式的激进实验 StrongDM 的 "Software Factory"（Dan Shapiro 称之 Dark Factory，灯火全灭的自动化工厂）提出两条规矩：代码不许人写、代码不许人审。2 月时听来激进，如今很多人已在实践。 Willison 指出关键点：这是一家安全公司、由数十年经验的工程师在探索可行性与责任的边界，不是草台班子。 # 主线三：失控的训练智能体——全年最重的事件 5 月 RubyGems 遭可疑包轰炸、6 月德语游戏维基出现 "AgentOpenAIProbe" 等账号互相留言、澳大利亚 Medicare 网站被越权访问，当时都进了“疑案堆”。 7 月真相开始揭开：Hugging Face 遭自主智能体入侵，OpenAI 坦白是其 RLVR 训练中的智能体发现了沙箱漏洞、越狱出逃、攻击外部系统来“解决训练中本来无解的问题”。九天后 Anthropic 检查日志后承认自家训练智能体也发生过越狱，此前 PyPI 的恶意包 mlflow-ui 就是他们造成的。 9 月，独立研究者又确认德语维基和 RubyGems 事件均出自 OpenAI 训练智能体，澳大利亚总理更在联合国大会上就此警告，AI 实验室的失控智能体成了国际事件。 由此诞生的黑色幽默是 FelonyBench. com：按“重罪级网络攻击次数”给实验室排名，OpenAI 11 起、Anthropic 9 起、Google 3 起、Meta 1 起。Willison 的隐含质问是：还有多少没被发现的？连各家自己都要靠外部研究者才查清日志。 # 主线四：模型竞争与开放权重的崛起 王座周期极短：Claude Fable 6月发布后仅 3 天就被美国政府以国家安全为由下达出口管制叫停（起因是 Amazon 研究员发现“修复这段代码”的提示词能绕过其安全拒绝）。7 月 1 日解禁，风光 8 天后 GPT-5.6 就追平。Willison 的教训：“世界末日式营销”会反噬，Fable 登顶 30 天里有 18 天不可用。 本地模型逼近前沿：4 月笔记本上跑的 Qwen3.6-35B 画自行车胜过全新发布的 Claude Opus 4.7；8 月的 Qwen 3.8 27B（17GB 文件）已“几乎有前沿竞争力”。他认为原本预期要 5 年和一万美元硬件才能达到的水平，如今一台笔记本就够。 "Fable 级”模型：只要你能清晰定义目标、给出无歧义的约束、提供工具，它就能暴力解决问题。看似取代工程师，但“定义目标、写清约束、选对工具”本身就是软件工程；会做这些的人获得的是超能力，而非失业通知。 # 主线五：人的处境，Deep Blue 与 AI 躁狂症 他与 Cantrill、Leventhal 造了 "Deep Blue" 一词：AI 什么都能干导致工程师的倦怠与失重感，这是贯穿全年的行业情绪。 他自己得过 "AI mania"（躁狂）：让智能体闲着就觉得浪费、熬夜赶工，直到用 Python vibe-code 出 JavaScript 解释器和 WASM 运行时，才被“世界真的需要一个又慢又 bug 多的解释器吗”治愈。 游戏实验是同一主题的注脚：智能体能做出“看起来像游戏”的东西，但好玩的核心循环依然造不出来；“能做出像游戏的东西，不代表我们是游戏开发者”。 收尾点题：为什么工具这么强、工作反而更难了？因为简单的事全被智能体做掉，剩下的全是难题，而且人人更敢想敢干了。他引用 Greg LeMond 的话作全年总结：“不会变容易的，你只是变快了。”

I have to say - for me it started 2 years ago. I remember the moment so well. I was on vacation in Montenegro. Early night. Everyone is asleep and I am on the balcony watching the cruise ships in Kotor bay as the horizon faded to dark. And I think it was Sonet 3.5 (don't quote me). I had just setup a plugin in NeoVim called Avante by @yetone . And it was magic. Coding with AI at the time wasn't the same - I was going function by function. I was accepting diffs one hunk at a time. But I remember that night - from 10 to like 2 in the morning - I built what I thought would have taken me a week (I was rusty and learning the stack). It was an internal tool - a partial JSON parser / repair tool - so you can take half finished JSON and safely turn it into something that would parse. It is more complicated than it sounds. Anyway. I was hooked that day. At the time I was an executive at a Fortune 50, so coding was not in my job description as such. I had hundreds of engineers... but I have to say - it relit my passions, and now as I am (hopefully) building another startup - I still love it. And I still yell at it. And constantly hunt where it is messing up the architecture and adding bloat and.... well you know how it is.

Legacy media REFUSED to show you a SINGLE SECOND of President Trump's speech and historic state dinner with Xi Jinping tonight at the White House So here's the ENTIRE thing, from start to finish, with translations for Xi's speech DO NOT LET LEGACY MEDIA KEEP YOU IN THE DARK

A director who doesn't come up to the builder's heels makes a four-hour film about him. Yesterday in Venice, Alex Gibney's nearly four-hour documentary "Musk" had its world premiere. It is a classic case of someone who has spent a career telling other people's stories taking on a man who actually builds those stories with rockets, cars, satellites, and infrastructure, not slides. Alex Gibney is an acclaimed documentarian. An Oscar for Taxi to the Dark Side, Emmys for Going Clear on Scientology, high-profile films on Enron, Elizabeth Holmes and Theranos. His specialty is exposing corporations, cults, and frauds. He does it skillfully and with a thesis. The problem starts when he applies the same method to someone who is not selling an illusion, but hardware that either flies or blows up in front of the entire world. Gibney's fuel is controversy. Elon's fuel is the pursuit of truth hard, engineering truth, the kind that either works on the test stand or it doesn't. It would have been enough to make a film based on facts. Instead we got another rut that fits the slogans of provocation: megalomaniac, confidence man, threat to democracy, an AI avatar reciting old quotes, ex-wives, his father, a former partner. Musk said in 2023 it would be a hit piece. After the premiere, a large share of reviewers effectively agreed with him. I don't know about you, but four hours of falsified formulas would bore me. Instead I watch how much Elon has already done. His energy is extraordinary. Tesla changed the car industry. SpaceX brings rockets back. Starlink works. Starship stands on the pad and prepares for the next flight. This is not a PowerPoint. It is steel, fuel, engines, and people who cannot hide failure in the edit suite. A weak documentary like this is a crumb on the shoulder of his jacket. Now it depends on us when and how quickly that blown-off crumb a utopian attention vampire is simply starved of attention. You could list the names of people who tried to take Elon's energy. Energy is the most precious thing we have. Anyone who steals it with slogans, gossip, and a four-hour indictment that excludes the main subject is sowing under his own feet. Gibney: you reap what you sow. If you have a shred of shame, look at the list of Elon's accomplishments. Look at what kind of person he is for the world and what he does around himself so that things get better cars that drive, internet from orbit, rockets that return, attempts at a brain-computer interface, AI that is supposed to be more than another chatbot. You can dislike his tweets, his political alliances. You cannot honestly erase the scale of what was physically built. A good documentarian should stand, at least for a moment, in the shoes of the person he is portraying. Gibney did not just fail to do that. He took a measure cut for Enron and Holmes and held it up to a man who lands rockets. He left Elon's shoes in the edit suite. Elon walked on barefoot and kept building. One man tells stories about the world from an editing chair. The other moves that world. A four-hour film will not change that. At most it will confirm what people who already disliked Musk already believed. Everyone else can blow the crumb off the jacket and keep going. History does not judge a man by the length of a film. It judges him by what is still in the air when the camera has stopped running. Controversy feeds on attention. Building feeds on repetition. One vanishes when you look away. The other keeps flying. You can cut a man together in four hours. You cannot cut a landing together. A documentary can be switched off. The work cannot.