共收录 22 篇深度内容
📌 今日概览
- AI 产品转向可执行任务：ChatGPT 上线财务功能，Gemini 推出批量决策 API 与自托管沙箱接入。
- 科研与机器人模型同步推进：AlphaGenome 2 开放 DNA 变异预测 API，GR00T 2 展示长程灵巧操作。
- AI 智能体成为基础设施议题：编排工具、B2A 流量与企业自托管环境受到集中讨论。
- 芯片行业围绕黄仁勋去留争论，游戏界继续就 AI 参与制作的标准存在分歧。
🔥 必读 Top3
- Jim Fan 作者展示 GR00T 2 驱动真实灵巧手完成化学实验并自行清理的全程视频，称其为自主、未剪辑、1 倍速。引用内容称该模型为面向机器人的开放 System One 模型，完全在机器人在载算力运行，支持 20 步以上、5 分钟以上的长程灵巧操作与全身控制。
入选理由：提供自主、未剪辑、1 倍速的真实灵巧手实验视频，是当日机器人模型进展中最具验证价值的条目。 - Demis Hassabis 哈萨比斯宣布 AlphaGenome 2 DNA 序列模型，称其可预测 DNA 变异在特定细胞类型中的影响，助力理解复杂疾病，学术研究可通过 API 使用。
入选理由：头部实验室发布科研专用 DNA 序列模型并开放学术 API，直接面向复杂疾病机理研究。 - ChatGPT ChatGPT 宣布桌面应用上线 Finances 功能，可查看收支与投资、寻找省钱方式并自建财务工具，美国 Free、Go、Plus、Pro 用户可用。
入选理由：消费级 AI 产品新增财务场景，覆盖美国多个订阅档位，影响面广且可实际使用。
🧭 主题分组
- AI 产品（4 条）—— 大厂把 AI 从问答推向财务、批量决策与可自托管的任务执行。
- 科研机器人（3 条）—— DNA 变异预测与机器人长程操作两项进展均强调可实际调用的能力。
- 模型评测（1 条）—— Gemini Decision 1 在决策基准领先，但成本高于同类，尚未达到前沿水平。
- 芯片行业（3 条）—— 围绕黄仁勋个人失误与去留的讨论，延伸到 AI 泡沫是否会被戳破。
- 游戏产业（2 条）—— AAA 游戏的 AI 辅助边界被重新讨论，独立工作室成为反 AI 风潮的主要承受方。
- AI 智能体（2 条）—— 智能体编排、B2A 流量与企业自托管接入共同显示 agent 正成为基础层。
- AI 应用（1 条）—— 艺术家用 Claude 编写代码生成视频装置，展示创作流程中的工具化使用。
- 黑客松（1 条）—— 收费筛选作弊者的设计误伤建设者，主办方改为免费并全额退款。
- 体育分析（1 条）—— 布雷迪开球前的手势规律经录像研究识别，并用于超级碗防守部署。
- 思维谜题（1 条）—— 狮子与毒牛排的答案取决于狮子数量的奇偶与完全理性的多层共识。
- 美国政治（1 条）—— 马库斯援引报道要求卢特尼克与特朗普辞职并应被起诉。
- 社会话题（1 条）—— 无家可归者威胁母子的帖子引发评论者责任归因争议。
- AI 观点（1 条）—— 以芒格旧语戏谑 AI 风险，属于低可信度的调侃。
📖 详细内容
AI 产品
大厂把 AI 从问答推向财务、批量决策与可自托管的任务执行。
ChatGPT
｜ @ChatGPT
｜ 影响力: 78.7万粉丝 · 互动分 3190.0
ChatGPT
｜ @ChatGPT
｜ 影响力: 78.7万粉丝 · 互动分 3190.0
💡 核心观点: ChatGPT 宣布桌面应用上线 Finances 功能，可查看收支与投资、寻找省钱方式并自建财务工具，美国 Free、Go、Plus、Pro 用户可用。
可信度: 70/100
事实核查:
- 原文支持: Finances is available to Free, Go, Plus and Pro users in the U.S.（官方账号公布的上线范围，属产品公告）
- 原文支持: You can also build custom financial tools.（为功能描述，原文未说明实现方式）
原文内容:
🤑🤑🤑 Cha-ching 🤑🤑🤑 Finances are now in the desktop app. Find ways to save, understand your spending, and make sense of your investments. You can also build custom financial tools. Update the desktop app to get started. Finances is available to Free, Go, Plus and Pro users in the U.S.
⏰ 19:46 | ❤️ 2302 点赞 | 📝 50词 | 入选: 高赞 2302 | 查看原文 →
Google AI Studio
｜ @GoogleAIStudio
｜ 影响力: 21.5万粉丝 · 互动分 3176.5
Google AI Studio
｜ @GoogleAIStudio
｜ 影响力: 21.5万粉丝 · 互动分 3176.5
💡 核心观点: Google AI Studio 发布 Gemini Decisions API，称基于多模态决策模型 Gemini Decision 1，可对最多 1000 个问题批量给出快速、校准后的决策。
可信度: 66/100
事实核查:
- 原文支持: powered by Gemini Decision 1, our new multimodal decision model（官方发布信息，原文未附基准或评测）
- 原文支持: get fast, calibrated decisions on up to 1000 questions at once（属官方性能宣称，无独立数据）
原文内容:
introducing the Gemini Decisions API powered by Gemini Decision 1, our new multimodal decision model get fast, calibrated decisions on up to 1000 questions at once across text, images, audio, video, and PDFs try it today in AI Studio and via the Gemini API https://t.co/h3Kmz2ypDi
⏰ 18:15 | ❤️ 2300 点赞 | 📝 45词 | 入选: 高赞 2300 | 查看原文 →
Google AI Studio
｜ @GoogleAIStudio
｜ 影响力: 21.5万粉丝 · 互动分 760.2
Google AI Studio
｜ @GoogleAIStudio
｜ 影响力: 21.5万粉丝 · 互动分 760.2
💡 核心观点: Google AI Studio 宣布：在 Gemini API 上使用 Managed Agents 构建应用的开发者，现在可以接入自托管环境，自带沙箱。
可信度: 75/100
事实核查:
- 原文支持: now you can bring your own sandbox with self-hosted environments（原文给出了该产品能力说明和上手链接，属于官方发布信息。）
原文内容:
building with Managed Agents on the Gemini API? now you can bring your own sandbox with self-hosted environments start building today: https://t.co/shgEhF5GRX https://t.co/xUBUB9qHWa
⏰ 20:10 | ❤️ 457 点赞 | 📝 28词 | 入选: 高赞 457 | 查看原文 →
Machina
｜ @EXM7777
｜ 影响力: 16.0万粉丝 · 互动分 2118.5
Machina
｜ @EXM7777
｜ 影响力: 16.0万粉丝 · 互动分 2118.5
💡 核心观点: 作者称 Hermes 桌面端改版已接近完成，并开放早期体验名额；引用推文中他表示要为 Hermes agent 打造最好的 UI/UX，并称会优于 Grok Bot、ChatGPT Dots 等产品。
可信度: 32/100
事实核查:
- 无法核实: almost done with the Hermes desktop redesign（仅有作者自述进度，未给出完成时间或可检查的版本。）
- 无法核实: it will be better than Grok Bot, ChatGPT Dots or anything else（这是对尚未发布产品的比较性断言，原文未提供依据。）
原文内容:
almost done with the Hermes desktop redesign... who wants an early access? https://t.co/lA3ZBG6PzG
引用 @EXM7777: i'm building the best UI/UX for Hermes agent
it will be better than Grok Bot, ChatGPT Dots or anything else
what's one thing you're missing in the current desktop app? https://t.co/HkTtUVOgZq
⏰ 14:00 | ❤️ 1300 点赞 | 📝 15词 | 入选: 高赞 1300 / 引用讨论 | 查看原文 →
科研机器人
DNA 变异预测与机器人长程操作两项进展均强调可实际调用的能力。
Demis Hassabis
｜ @demishassabis
｜ 影响力: 195.8万粉丝 · 互动分 4149.0
Demis Hassabis
｜ @demishassabis
｜ 影响力: 195.8万粉丝 · 互动分 4149.0
💡 核心观点: 哈萨比斯宣布 AlphaGenome 2 DNA 序列模型，称其可预测 DNA 变异在特定细胞类型中的影响，助力理解复杂疾病，学术研究可通过 API 使用。
可信度: 68/100
事实核查:
- 原文支持: It predicts the impact of DNA variants in different specific cell types（为官方发布方对产品能力的描述，原文未给出评测数据）
- 原文支持: Available for academic research via our API（属产品可用性说明，原文未提供访问细节）
原文内容:
Super proud to announce AlphaGenome 2 - our new DNA sequence model! 🧬 It predicts the impact of DNA variants in different specific cell types - unlocking crucial information for understanding complex diseases. Available for academic research via our API: https://t.co/dEyrvpTxIl
引用 @GoogleDeepMind: Your heart, brain, and immune system share the same DNA sequence. So how can a single-letter change affect disease in some cells, while sparing the rest?
AlphaGenome 2 is our new AI model that zooms in on specific cells to help scientists understand what drives disease. 🧵
⏰ 17:12 | ❤️ 2938 点赞 | 📝 40词 | 入选: 高赞 2938 / 引用讨论 | 查看原文 →
Jim Fan
｜ @DrJimFan
｜ 影响力: 60.5万粉丝 · 互动分 1015.26
Jim Fan
｜ @DrJimFan
｜ 影响力: 60.5万粉丝 · 互动分 1015.26
💡 核心观点: 作者展示 GR00T 2 驱动真实灵巧手完成化学实验并自行清理的全程视频，称其为自主、未剪辑、1 倍速。引用内容称该模型为面向机器人的开放 System One 模型，完全在机器人在载算力运行，支持 20 步以上、5 分钟以上的长程灵巧操作与全身控制。
可信度: 55/100
事实核查:
- 无法核实: Autonomous, uncut, 1x speed（视频为作者提供的展示素材，原文未给出第三方验证或原始录制信息）
- 原文支持: Long-horizon tasks: dexterous manipulation and whole-body control chained over 20+ steps and 5+ minutes（该表述见于作者引用的 GR00T 2 发布文案，原文确有此描述；具体性能无独立证据）
原文内容:
Autonomous, uncut, 1x speed. This is GR00T 2 controlling real dexterous hands to perform a chemistry lab experiment (and clean up after itself) end to end. Episode 1 of a new series! I’ll keep dropping real-time robot videos, with none of the “4x” or “8x” labels you usually see in the corner. No one can speed up reality, except Dr. Strange ;)
引用 @DrJimFan: Introducing GR00T 2: our most advanced, open System One model for robotics. It runs entirely on your robot’s onboard compute, fully untethered, and you own every weight of it. GR00T 2 features:
1. Long-horizon tasks: dexterous manipulation and whole-body control chained over 20+ steps and 5+ minut…
⏰ 15:32 | ❤️ 593 点赞 | 📝 57词 | 入选: 高赞 593 / 引用讨论 | 查看原文 →
Chubby♨️
｜ @kimmonismus
｜ 影响力: 15.9万粉丝 · 互动分 871.31
Chubby♨️
｜ @kimmonismus
｜ 影响力: 15.9万粉丝 · 互动分 871.31
💡 核心观点: Dario Amodei 称癌症、阿尔茨海默病、心脏病和精神疾病可能不到十年就成为历史。作者认为这发生在 Anthropic IPO 前，并且觉得这是真的，不只是 IPO 炒作。
可信度: 30/100
事实核查:
- 原文支持: Dario Amodei says cancer, Alzheimer's, heart disease and mental illness could be history in less than a decade（原文直接陈述了该预测，但未给出数据或来源支撑其可行性。）
- 原文支持: This comes right before Anthropics IPO（原文提到该预测出现在 Anthropic IPO 之前；引文中另有彭博社关于 IPO 超额认购及最高融资 904 亿美元、目标估值约 1.4 万亿美元的信息。）
- 无法核实: more than just IPO hype（这是作者对预测动机与可信度的个人判断，原文无相关证据。）
原文内容:
Dario Amodei says cancer, Alzheimer's, heart disease and mental illness could be history in less than a decade. This comes right before Anthropics IPO, but after Fable 5.2 release, I think its true and more than just IPO hype.
引用 @kimmonismus: Looking crazy good for Anthropic: Its IPO is already multiple times oversubscribed, according to Bloomberg, with investors seeking far more shares than are available.
The company and its co-founders aim to raise up to $90.4 billion through the IPO and a private placement, targeting a fully diluted…
⏰ 14:31 | ❤️ 629 点赞 | 📝 39词 | 入选: 高赞 629 / 引用讨论 | 查看原文 →
模型评测
Gemini Decision 1 在决策基准领先，但成本高于同类，尚未达到前沿水平。
Artificial Analysis
｜ @ArtificialAnlys
｜ 影响力: 17.7万粉丝 · 互动分 736.5
Artificial Analysis
｜ @ArtificialAnlys
｜ 影响力: 17.7万粉丝 · 互动分 736.5
💡 核心观点: 引用 Artificial Analysis 评测：谷歌首个决策模型 Gemini Decision 1 在决策指数得 66 分，知识与推理基准领先，但尚未达到前沿水平；每百万任务成本 345 美元，约为 Drex Max v1.1 的三倍。
可信度: 85/100
事实核查:
- 原文支持: scoring 66 on the Artificial Analysis Decision Index（原文明确给出该指数得分及版本 v2.1。）
- 原文支持: costs $345 per 1M Tasks, around three times higher than Drex Max v1.1 ($103)（原文同时列出两项价格，345 美元约为 103 美元的 3.35 倍，与「约三倍」表述一致。）
- 原文支持: leading in knowledge and reasoning benchmarks（原文称其领先知识与推理基准，但在 AA-Paperclip 等其他评测上落后于 Drex Max v1.1。）
原文内容:
Google has released its first decision model, Gemini Decision 1, scoring 66 on the Artificial Analysis Decision Index and leading in knowledge and reasoning benchmarks Gemini Decision 1 from @GoogleDeepMind scores 66 on the Artificial Analysis Decision Index v2.1, in line with Elephant 9B (@baselabs) and Jev 1.14 (@typesafeai), and 7 points behind leaders Drex Max v1.1 (@NaceAI) and GLiDE v1.2 (@fastinoAI), which both score 73. It is priced at $0.10 per 1M input tokens and costs $345 per 1M Tasks across the Index, around three times higher than Drex Max v1.1 ($103) due to its higher pricing Key results for Gemini Decision 1: ➤ Not yet frontier on Decision Index: Gemini Decision 1 scores 66 on the Artificial Analysis Decision Index. It leads knowledge and reasoning benchmarks, scoring 66 on AA-Omniscience-Decision and 49 on HLE-MultiChoice. It trails on other evaluations, including our new proprietary classification and extraction benchmark AA-Paperclip, where it answers all 15 questions correctly on 33% of documents vs. 63% for Drex Max v1.1 ➤ More than 3x the Cost per Task of leading models: Gemini Decision 1 costs $345 per 1M Tasks, compared with $103 for Drex Max v1.1, $156 for Jev 1.14. This difference is driven by its token price of $0.10 per 1M input tokens vs. $0.03 for Drex Max v1.1. Gemini Decision 1 has a similar Cost per 1M Tasks to GPT-6 Luna (Decisions, $339) ➤ Strong calibration for reasoning tasks: We measure calibration with the Brier score, the average squared gap between the probability a model assigns to its chosen answer and whether that answer was correct (lower is better). Gemini Decision 1 has a Brier score of 0.18 across the Index, between Drex Max v1.1 (0.16) and Jev 1.14 (0.19). Calibration varies by task. with its Brier score of 0.20 on HLE-MultiChoice among the lowest of all models in the Index, while its 0.14 on When2Call compares with 0.04 for Drex Max v1.1
⏰ 18:18 | ❤️ 500 点赞 | 📝 272词 | 入选: 高赞 500 / 长文 | 查看原文 →
芯片行业
围绕黄仁勋个人失误与去留的讨论，延伸到 AI 泡沫是否会被戳破。
Chubby♨️
｜ @kimmonismus
｜ 影响力: 15.9万粉丝 · 互动分 1806.0
Chubby♨️
｜ @kimmonismus
｜ 影响力: 15.9万粉丝 · 互动分 1806.0
💡 核心观点: 作者反对让 Jensen Huang 离开 NVIDIA。他承认黄仁勋似乎犯了错并承担责任，但认为黄从零打造 NVIDIA、本人就是公司的象征，其重要性超过这次失误，公司和行业发展不应因此受损。
可信度: 28/100
事实核查:
- 原文支持: Jensen Huang must stay.（这是作者的核心主张，属直接表态。）
- 原文支持: how he built NVIDIA from virtually nothing（原文以此作为黄的功绩依据，未给出具体历程或数据。）
- 无法核实: Yes, it appears he made a mistake.（作者仅称看来他犯了错，未说明错误内容，原文没有可核验细节。）
原文内容:
If I didn’t know any better, I’d call it clickbait. No, Jensen Huang must stay. Anyone familiar with his biography knows how he built NVIDIA from virtually nothing. He is more than "just" a CEO; he *is* NVIDIA, arguably the most important company in the world, providing the engine that powers the technological revolution. Yes, it appears he made a mistake. But he takes responsibility. And neither NVIDIA as a company nor the broader course of development should suffer because of it. Jensen is more significant than the mistake that occurred.
引用 @kuhhbin: Jensen Huang must go
We argue in @TheEconomist that his position is untenable
"Like many powerful leaders, he came to believe that he was too important for the normal rules to apply to him."
But he is not the only one. Howard Lutnick has failed to stem the flow of chips to China. More heads shou…
⏰ 21:28 | ❤️ 1357 点赞 | 📝 93词 | 入选: 高赞 1357 / 引用讨论 | 查看原文 →
@levelsio
｜ @levelsio
｜ 影响力: 99.1万粉丝 · 互动分 1316.0
@levelsio
｜ @levelsio
｜ 影响力: 99.1万粉丝 · 互动分 1316.0
💡 核心观点: 作者以反问评论纽约时报关于 Jensen Huang 第二次婚外情的报道，猜测这件事会不会成为戳破 AI 泡沫的导火索。
可信度: 16/100
事实核查:
- 无法核实: Nvidia CEO Jensen Huang had a second affair with VP of Health Care Kimberly Powell, per the New York Times.（引用文本称消息来自纽约时报，但未附链接或报道细节，输入内无法核实。）
- 无法核实: Wouldn't it be something if this was the thing that would pop the AI bubble（这是假设性猜测，原文没有因果依据。）
原文内容:
Wouldn't it be something if this was the thing that would pop the AI bubble
引用 @BullTheoryio: BREAKING: 🇺🇸 Nvidia CEO Jensen Huang had a second affair with VP of Health Care Kimberly Powell, per the New York Times. https://t.co/OyeMo2Xq1S
⏰ 22:56 | ❤️ 1029 点赞 | 📝 16词 | 入选: 高赞 1029 / 引用讨论 | 查看原文 →
Anthony Pompliano 🌪
｜ @APompliano
｜ 影响力: 244.0万粉丝 · 互动分 1210.0
Anthony Pompliano 🌪
｜ @APompliano
｜ 影响力: 244.0万粉丝 · 互动分 1210.0
💡 核心观点: 作者提出条件式判断：如果 NVIDIA 让 Jensen Huang 离开而牛市仍然延续，那么就没有什么能阻止 AI 浪潮继续让投资者获利。
可信度: 14/100
事实核查:
- 无法核实: If Nvidia ousts Jensen Huang and the bull market continues, there is literally nothing that will stop the AI tsunami from enriching investors.（这是带前提的预测，原文未提供数据或机制论证。）
原文内容:
If Nvidia ousts Jensen Huang and the bull market continues, there is literally nothing that will stop the AI tsunami from enriching investors.
⏰ 22:44 | ❤️ 894 点赞 | 📝 23词 | 入选: 高赞 894 | 查看原文 →
游戏产业
AAA 游戏的 AI 辅助边界被重新讨论，独立工作室成为反 AI 风潮的主要承受方。
Ethan Mollick
｜ @emollick
｜ 影响力: 41.8万粉丝 · 互动分 4715.5
Ethan Mollick
｜ @emollick
｜ 影响力: 41.8万粉丝 · 互动分 4715.5
💡 核心观点: 莫里克预测，一周多后将发布的《The Last Human Game》标志着一个节点：不会再有 GTA 6 这种规模、零 AI 辅助的 AAA 游戏。
可信度: 32/100
事实核查:
- 无法核实: in a little over a week we will see the release of The Last Human Game（发布时间是作者的预测，原文无可核验依据）
- 无法核实: There will never be another AAA game at the scale of GTA 6 with no AI assistance.（属对未来的断言，无数据支撑）
原文内容:
Anyhow, in a little over a week we will see the release of The Last Human Game. There will never be another AAA game at the scale of GTA 6 with no AI assistance. People will look back at an all-human game with a billion dollar budget and wonder how they did it, like the Pyramids
⏰ 13:02 | ❤️ 3685 点赞 | 📝 56词 | 入选: 高赞 3685 | 查看原文 →
Ethan Mollick
｜ @emollick
｜ 影响力: 41.8万粉丝 · 互动分 1558.79
Ethan Mollick
｜ @emollick
｜ 影响力: 41.8万粉丝 · 互动分 1558.79
💡 核心观点: 作者认为游戏界反 AI 的风潮主要伤害独立工作室：在几乎所有编程都会涉及 AI 的环境里，要求游戏不含 AI 代码是不现实的标准；如果玩家在意人类创作，应奖励以人类艺术家和设计师为核心的游戏，而非审查 AI 使用。
可信度: 42/100
事实核查:
- 无法核实: Insisting on no AI code in a world where all coding will effectively involve AI is an impossible standard.（这是作者的判断，原文未给出行业数据支持。）
- 原文支持: Big Walk used AI for "under 0.1%" of its code（引用文本给出低于 0.1% 的具体比例，但未附检测方法。）
原文内容:
The anti-AI backlash in games is mostly hurting indie studios. Insisting on no AI code in a world where all coding will effectively involve AI is an impossible standard. If fans care about human work, reward games centering human artists & designers rather than policing AI use.
引用 @Kotaku: The reveal that Big Walk used AI for "under 0.1%" of its code has caused a big blowup among fans, developers, and game award judges.
Where do people draw the line on genAI in gaming, and how did a 2026 GOTY indie darling become ground zero for the debate?
We broke down the messy discussion:
htt…
⏰ 12:30 | ❤️ 1073 点赞 | 📝 48词 | 入选: 高赞 1073 / 引用讨论 | 查看原文 →
AI 智能体
智能体编排、B2A 流量与企业自托管接入共同显示 agent 正成为基础层。
GREG ISENBERG
｜ @gregisenberg
｜ 影响力: 76.1万粉丝 · 互动分 989.5
GREG ISENBERG
｜ @gregisenberg
｜ 影响力: 76.1万粉丝 · 互动分 989.5
💡 核心观点: 作者认为 Grok Bot 是当前最好的 AI 智能体编排器，并称自己见到的这套配置为最佳：工程环节运行 Claude Code，设计环节使用 Paper。引用内容介绍 Grok Bot 1.0 为可与整个团队协作的 AI 队友，能在 Slack、Teams 共享并受邀参会。
可信度: 30/100
事实核查:
- 无法核实: grok bot is the BEST AI agent orchestrator right now（属主观评价，原文未提供对比数据或评测依据）
- 无法核实: it runs claude code for engineering and paper for design（原文仅陈述该配置组合，未附可核验的操作细节）
原文内容:
grok bot is the BEST AI agent orchestrator right now and THIS is the best setup i’ve EVER seen https://t.co/NjkQuOVZag it runs claude code for engineering and paper for design here is the FULL setup and how to copy it: https://t.co/LXWN13lf3D
引用 @bot: Introducing Grok Bot 1.0. AI teammates that work alongside your whole team.
Share Bots, collaborate with them in Slack and Teams, invite them to meetings, and watch them learn and become more helpful. https://t.co/7569CZRc9G
⏰ 19:08 | ❤️ 636 点赞 | 📝 47词 | 入选: 高赞 636 / 引用讨论 | 查看原文 →
Justine Moore
｜ @venturetwins
｜ 影响力: 24.9万粉丝 · 互动分 1070.0
Justine Moore
｜ @venturetwins
｜ 影响力: 24.9万粉丝 · 互动分 1070.0
💡 核心观点: 作者指出企业过去多年在对抗机器人流量，如今转为争夺这类流量背后的生意；依据是 AI 智能体已占网络流量多数、产出大部分上线代码，并预测其将买下线上销售的大部分商品。后两点为预测。
可信度: 42/100
事实核查:
- 无法核实: AI agents are the majority of Web traffic and generate most code pushed to production（原文仅给出该判断，未附数据、来源或可核验细节）
- 无法核实: Soon, they'll be buying most of what gets sold online（原文用 Soon 表述未来趋势，属预测，无可核验数据支撑）
原文内容:
Companies spent years battling bots. Now they’re fighting for their business. AI agents are the majority of Web traffic and generate most code pushed to production. Soon, they'll be buying most of what gets sold online. New piece with @omooretweets on the B2A economy 👇 https://t.co/AGkbjLrD95
⏰ 16:22 | ❤️ 470 点赞 | 📝 48词 | 入选: 高赞 470 | 查看原文 →
AI 应用
艺术家用 Claude 编写代码生成视频装置，展示创作流程中的工具化使用。
Claude
｜ @claudeai
｜ 影响力: 206.3万粉丝 · 互动分 3863.0
Claude
｜ @claudeai
｜ 影响力: 206.3万粉丝 · 互动分 3863.0
💡 核心观点: Claude 官方账号介绍艺术家森万里子的回顾展“All That Shines”，其中一件新视频装置由她用 Claude 编写代码生成动画。
可信度: 45/100
事实核查:
- 无法核实: she used Claude to write code for a computer-generated animation（属厂商自述的用例，无第三方核验信息）
- 原文支持: presented as a new video installation（与原文表述一致，属展览信息）
原文内容:
Mariko Mori has pushed the boundaries of art for three decades. For her retrospective "All That Shines," she used Claude to write code for a computer-generated animation, presented as a new video installation. https://t.co/MtJ1rlClxt
⏰ 00:05 | ❤️ 2979 点赞 | 📝 37词 | 入选: 高赞 2979 | 查看原文 →
黑客松
收费筛选作弊者的设计误伤建设者，主办方改为免费并全额退款。
Marc Lou
｜ @marclou
｜ 影响力: 42.3万粉丝 · 互动分 2371.0
Marc Lou
｜ @marclou
｜ 影响力: 42.3万粉丝 · 互动分 2371.0
💡 核心观点: Marc Lou 承认将 $100K Hackathon 设为收费 $99 入口是自己的失误：原本想用付费筛掉作弊者、吸引认真开发者，却忽略了 $99 对印度、付不起账单或想赚第一笔钱的人是一笔大钱，等于把真正的建设者挡在门外。因此他决定改为完全免费，已付费者全部退款。
可信度: 72/100
事实核查:
- 原文支持: So I've decided to make the $100K Hackathon completely FREE :)（作者明确宣布赛事改为免费，属原文直述。）
- 无法核实: Everyone who already paid got a full refund.（作者称已全额退款，但未给出退款人数、金额或凭证。）
- 原文支持: So I added a $99 entry fee.（作者自述曾设置 $99 报名费，与后文退款说明一致。）
原文内容:
I made a mistake. When Jack and I decided to organize this $100K Hackathon, my biggest concern was cheating. Unlike a regular hackathon, everyone participates from home. And because the winner is whoever makes the most PROFIT (not revenue), there are so many ways to manipulate the numbers. So I added a $99 entry fee. One thing I’ve learned from building businesses is that people tend to be more committed when they pay for something. I thought it would filter out spam, attract serious builders, and make the competition fairer for everyone. But I completely missed the other side. 3 years ago, I was a living with my apps. Today, I’m financially free. My life changed so much that somewhere along the way, I lost perspective on what $99 means to someone else. For someone in India, someone struggling to pay their bills, or someone trying to make their first dollar online, $99 can be a lot of money. And those are exactly the people I want to help! I’ve spent years sharing what I’ve learned, reviewing random people’s landing pages, and organizing free meetups around the world. Not because there’s money to make, but because I genuinely love seeing people build things and succeed. And somehow, I was so focused on keeping cheaters out that I didn’t realize I was also keeping genuine builders out. That’s my mistake. So I’ve decided to make the $100K Hackathon completely FREE :) Everyone who already paid got a full refund. I’m still worried about cheating. That problem hasn’t disappeared, and Jack and I will have to figure out better ways to verify profits and keep the leaderboard fair. But that’s our problem to solve. Thank you to everyone who shared their thoughts. Sometimes you need people to remind you of things you’ve forgotten. I genuinely hope this hackathon changes someone’s life. ❤️
引用 @marclou: Higgsfield gave me $100,000 to realize my dream:
A hackathon where participants build a real business, not just an app.
So… Introducing The $100K Hackathon:
🏴☠️ https://t.co/nHDJ5Qruhu 🏴☠️
It starts on November 16, and you'll have 30 days to build a startup and make it profitable. The prizes:…
⏰ 10:14 | ❤️ 1404 点赞 | 📝 316词 | 入选: 高赞 1404 / 引用讨论 / 长文 | 查看原文 →
体育分析
布雷迪开球前的手势规律经录像研究识别，并用于超级碗防守部署。
Aakash Gupta
｜ @aakashgupta
｜ 影响力: 33.6万粉丝 · 互动分 2048.5
Aakash Gupta
｜ @aakashgupta
｜ 影响力: 33.6万粉丝 · 互动分 2048.5
💡 核心观点: 作者引用 Osi Umenyiora 的说法：布雷迪开球前指向防守球员时必定开球，不指则不开球，这一规律在研究录像两周后被发现，并用于 Super Bowl XLII，巨人队因此 5 次擒杀布雷迪。作者据此解释先手信息如何瓦解进攻线的时间优势。
可信度: 58/100
事实核查:
- 原文支持: they went out and sacked Brady 5 times（原文给出 5 次擒杀，并另称有 9 次撞击，属具体可对照的比赛数字。）
- 无法核实: He went back through all their games and says it held 100% of the time.（这是转述 Osi 的个人观察，原文未提供逐场统计或样本说明。）
- 原文支持: The 2007 Patriots had three Pro Bowl linemen.（原文提及三名职业碗进攻线球员，但未列出姓名。）
原文内容:
Tom Brady lost a perfect 19-0 season because of one little habit with his finger. Osi Umenyiora, the Giants defensive end, caught it on film two weeks before the Super Bowl. He pulled the whole defensive line into a room, showed them, and they went out and sacked Brady 5 times. Here's the habit. Brady would walk up to the line and point at somebody on defense before he yelled "hut." Osi says that whenever Brady pointed, the ball was snapped on "hut," and whenever he skipped the point, the snap wasn't coming. He went back through all their games and says it held 100% of the time. That sounds small until you think about how a pass rush works. On every play, the offensive linemen know when the ball is getting snapped and the defense doesn't. That's their whole edge. The big guys up front get to move first, and the pass rusher has to sit there and react, already a step behind someone who outweighs him and only has to get in his way. Osi basically took that edge away from one of the best offensive lines in football. The 2007 Patriots had three Pro Bowl linemen. They scored 589 points, the most in NFL history at that point, and Brady got sacked just 21 times in 16 regular season games. The Giants had actually played them in the last week of the regular season and lost 38-35. They got to Brady once that night. Five weeks later, it was the same quarterback and the same line, and the Giants sacked him 5 times and hit him 9 more. Osi says the difference was the two weeks he spent watching tape. And get this. Earlier that same season, the NFL had fined Bill Belichick $500,000, the biggest fine any NFL coach had gotten up to that point. They also hit the team for another $250,000 and a first-round draft pick, all for filming the Jets' defensive signals from the sideline. So New England paid $750,000 and a first-rounder for stealing signals with a camera, and their perfect season ended because a defensive end found Brady's own signal sitting on regular game film, totally legal, for free. Osi didn't even get a sack that night. Before the game, Brady had said his number one job was keeping #72 and #92 off him, meaning Osi and Michael Strahan. Justin Tuck got two sacks, and Strahan, Jay Alford and Kawika Mitchell got one each. Osi went home with 4 tackles, a hit and a fumble recovery. He found the most valuable piece of information in the building and gave it to everybody else. Rookie Jay Alford sacked Brady with 19 seconds left to seal it, 17-14. Four years later, Brady said he still couldn't watch highlights of that game.
引用 @JoshChambers: Osi Umenyiora says he discovered Tom Brady's pre snap cadence and knew when he was going to snap the ball ahead of Super Bowl XLII, and says it allowed the Giants defensive line to dominate the game with 5 sacks and 20 hurries
“But as I was watching the tape, because you have two weeks leading up t…
⏰ 13:49 | ❤️ 1649 点赞 | 📝 459词 | 入选: 高赞 1649 / 引用讨论 / 长文 | 查看原文 →
思维谜题
狮子与毒牛排的答案取决于狮子数量的奇偶与完全理性的多层共识。
Aakash Gupta
｜ @aakashgupta
｜ 影响力: 33.6万粉丝 · 互动分 1901.5
Aakash Gupta
｜ @aakashgupta
｜ 影响力: 33.6万粉丝 · 互动分 1901.5
💡 核心观点: 作者复述一道 Goldman Sachs 面试题：253 只狮子与一块毒牛排，狮子吃后睡着会被其他狮子吃掉，结果随狮子数量奇偶翻转，253 为奇数时有一只狮子会吃。作者强调答案依赖所有狮子完全理性的多层共识。
可信度: 50/100
事实核查:
- 无法核实: There's a riddle Goldman Sachs has asked job candidates（原文称该题曾用于 Goldman Sachs 面试，未提供题目出处或面试记录。）
- 原文支持: all you really need to know is whether 253 is odd. It is.（原文的奇偶推理与 253 为奇数一致，属文本内部可核对。）
原文内容:
There's a riddle Goldman Sachs has asked job candidates where the right answer is to eat the poison, and it only works because there are 253 lions in the room instead of 254. The setup goes like this. There's one poisoned steak, and any lion who eats it falls asleep. A sleeping lion is fair game for the rest of them, and whoever eats him falls asleep too. Every lion is perfectly rational and wants to live. The trick is to forget about 253 lions and start with one. If there's only one lion, he eats the steak, takes a nap, and nobody's around to bother him. With two lions, nobody eats. Whoever goes first passes out, and the other one gets a free meal with nobody left to come after him. With three lions, somebody eats. Once he's asleep, the other two are stuck in the two-lion situation, where taking a bite means you're next, so they leave him alone. He ate the poison and he's the safest lion in the room. Four lions, nobody eats, because the first one to bite leaves three behind, and we just saw that one of those three will happily eat him. Keep counting and the answer flips with every lion you add, so all you really need to know is whether 253 is odd. It is. One lion eats, and the other 252 spend the night staring at a sleeping lion every one of them has figured out they can't touch. What makes this a Wall Street question is how much trust the answer takes. For lion 253 to eat safely, he has to know every other lion is perfectly rational, and know that they all know that, and so on, 252 layers deep. He's betting his life that the lion at the far end of the room did the exact same math. Actual humans tap out after two or three layers, and somebody measured it. Back in 1997, Richard Thaler got the Financial Times to run a contest. Readers picked a number from 0 to 100, and whoever got closest to two-thirds of the average won two business-class tickets worth over $10,000. Follow the logic all the way down and everyone should pick zero. The average came in at 18.91, which made 13 the winner. The two most popular guesses were 33 and 22, which is what you get from thinking one and two steps ahead. Plenty of people did the perfect math, wrote down zero, and lost to people who just had a good read on how far everyone else would bother to think. Keynes said the same thing about the stock market back in 1936, that you make money by guessing what the average investor thinks the average investor is going to do. A Goldman quant told eFinancialCareers the lion question was the easiest of seven riddles she got in a single hour, which tracks, since the math only gets you hired. Once you're on the desk, the money comes from knowing the person on the other side of your trade probably stopped thinking somewhere around 22.
引用 @MercuriusFilius: How would you answer this common Goldman Sachs interview question? https://t.co/6hya34Ri5V
⏰ 16:12 | ❤️ 1454 点赞 | 📝 514词 | 入选: 高赞 1454 / 引用讨论 / 长文 | 查看原文 →
美国政治
马库斯援引报道要求卢特尼克与特朗普辞职并应被起诉。
Gary Marcus
｜ @GaryMarcus
｜ 影响力: 25.7万粉丝 · 互动分 5654.0
Gary Marcus
｜ @GaryMarcus
｜ 影响力: 25.7万粉丝 · 互动分 5654.0
💡 核心观点: 加里·马库斯援引 Axios 报道，称卢特尼克与特朗普“出卖美国”，要求二人辞职并应被起诉，措辞指向比水门事件更严重。
可信度: 38/100
事实核查:
- 无法核实: both Lutnick and Trump sold out the United States（这是作者对引述报道的立场性概括，引述内容仅涉及 Lutnick 对 Megaspeed 未被列入黑名单的处置，原文未提供特朗普相关证据）
- 原文支持: Megaspeed, a company accused of illegally diverting U.S. chips to China（该说法出自引述的 Axios 报道文本，原文标明来源）
- 无法核实: They should both resign.（属作者主张，无事实依据可核）
原文内容:
Way worse than Watergate: it looks like both Lutnick and Trump sold out the United States. They should both resign. In a just world they would be brought up on charges.
引用 @MarioNawfal: 🚨🇺🇸🇨🇳 Sec. Bessent reportedly CONFRONTED Sec. Lutnick inside the WH Situation Room over CHINA.
Bessent demanded to know why Megaspeed, a company accused of illegally diverting U.S. chips to China, had NOT been blacklisted.
Lutnick's response? “I don't know. I'll look into it.” 💀
Officials had rep…
⏰ 17:43 | ❤️ 3532 点赞 | 📝 31词 | 入选: 高赞 3532 / 引用讨论 | 查看原文 →
社会话题
无家可归者威胁母子的帖子引发评论者责任归因争议。
Justine Moore
｜ @venturetwins
｜ 影响力: 24.9万粉丝 · 互动分 3149.5
Justine Moore
｜ @venturetwins
｜ 影响力: 24.9万粉丝 · 互动分 3149.5
💡 核心观点: 摩尔转述 Reddit 帖：一名女性称无家可归者反复威胁她与幼儿，数十名评论者却指责她才是问题所在。
可信度: 28/100
事实核查:
- 无法核实: Woman posts on Reddit that a homeless man is repeatedly threatening her + her toddler at their bus stop（为转述的单一网络帖文，原文未提供链接或佐证）
- 无法核实: Dozens of commenters tell her that SHE is the problem（转述评论区情况，无可核验来源）
原文内容:
Living in SF is so crazy sometimes. Woman posts on Reddit that a homeless man is repeatedly threatening her + her toddler at their bus stop - yelling at them and following them down the sidewalk. Dozens of commenters tell her that SHE is the problem 🫠 https://t.co/MJDKbwTGrq
⏰ 20:08 | ❤️ 2539 点赞 | 📝 48词 | 入选: 高赞 2539 | 查看原文 →
AI 观点
以芒格旧语戏谑 AI 风险，属于低可信度的调侃。
Anthony Pompliano 🌪
｜ @APompliano
｜ 影响力: 244.0万粉丝 · 互动分 2937.5
Anthony Pompliano 🌪
｜ @APompliano
｜ 影响力: 244.0万粉丝 · 互动分 2937.5
💡 核心观点: 庞普里亚诺引用芒格“酒、杠杆与女人是男人败因”的说法，戏称大家以为 AI 会因杠杆出事，结果没料到是“女人”。
可信度: 15/100
事实核查:
- 无法核实: Charlie Munger once said liquor, leverage, and ladies are the downfall of men.（原文未给出该引语的出处，无法核验）
- 无法核实: Everyone thought AI would blow up over leverage, but no one was prepared for it to be women.（属调侃式评论，无事实内容）
原文内容:
Charlie Munger once said liquor, leverage, and ladies are the downfall of men. Everyone thought AI would blow up over leverage, but no one was prepared for it to be women.
⏰ 23:29 | ❤️ 2279 点赞 | 📝 31词 | 入选: 高赞 2279 | 查看原文 →
🔗 引用关系
- @GaryMarcus 引用 @MarioNawfal：加里·马库斯援引 Axios 报道，称卢特尼克与特朗普“出卖美国”，要求二人辞职并应被起诉，措辞指向比水门事件更严重。（原文：🚨🇺🇸🇨🇳 Sec. Bessent reportedly CONFRONTED Sec. Lutnick inside the WH Situation Room over CHINA.
Bessent demanded to know…）
- @demishassabis 引用 @GoogleDeepMind：哈萨比斯宣布 AlphaGenome 2 DNA 序列模型，称其可预测 DNA 变异在特定细胞类型中的影响，助力理解复杂疾病，学术研究可通过 API 使用。（原文：Your heart, brain, and immune system share the same DNA sequence. So how can a single-letter change affect disease in so…）
- @marclou 引用 @marclou：Marc Lou 承认将 $100K Hackathon 设为收费 $99 入口是自己的失误：原本想用付费筛掉作弊者、吸引认真开发者，却忽略了 $99 对印度、付不起账单或想赚第一笔钱的人是一笔大钱，等于把真正的建设者挡在门外。因此他决定改为完全免费，已付费者全部退款。（原文：Higgsfield gave me $100,000 to realize my dream:
A hackathon where participants build a real business, not just an app.…）
- @EXM7777 引用 @EXM7777：作者称 Hermes 桌面端改版已接近完成，并开放早期体验名额；引用推文中他表示要为 Hermes agent 打造最好的 UI/UX，并称会优于 Grok Bot、ChatGPT Dots 等产品。（原文：i'm building the best UI/UX for Hermes agent
it will be better than Grok Bot, ChatGPT Dots or anything else
what's one…）
- @aakashgupta 引用 @JoshChambers：作者引用 Osi Umenyiora 的说法：布雷迪开球前指向防守球员时必定开球，不指则不开球，这一规律在研究录像两周后被发现，并用于 Super Bowl XLII，巨人队因此 5 次擒杀布雷迪。作者据此解释先手信息如何瓦解进攻线的时间优势。（原文：Osi Umenyiora says he discovered Tom Brady's pre snap cadence and knew when he was going to snap the ball ahead of Super…）
- @aakashgupta 引用 @MercuriusFilius：作者复述一道 Goldman Sachs 面试题：253 只狮子与一块毒牛排，狮子吃后睡着会被其他狮子吃掉，结果随狮子数量奇偶翻转，253 为奇数时有一只狮子会吃。作者强调答案依赖所有狮子完全理性的多层共识。（原文：How would you answer this common Goldman Sachs interview question? https://t.co/6hya34Ri5V）
- @kimmonismus 引用 @kuhhbin：作者反对让 Jensen Huang 离开 NVIDIA。他承认黄仁勋似乎犯了错并承担责任，但认为黄从零打造 NVIDIA、本人就是公司的象征，其重要性超过这次失误，公司和行业发展不应因此受损。（原文：Jensen Huang must go
We argue in @TheEconomist that his position is untenable
"Like many powerful leaders, he came to…）
- @emollick 引用 @Kotaku：作者认为游戏界反 AI 的风潮主要伤害独立工作室：在几乎所有编程都会涉及 AI 的环境里，要求游戏不含 AI 代码是不现实的标准；如果玩家在意人类创作，应奖励以人类艺术家和设计师为核心的游戏，而非审查 AI 使用。（原文：The reveal that Big Walk used AI for "under 0.1%" of its code has caused a big blowup among fans, developers, and game a…）
- @levelsio 引用 @BullTheoryio：作者以反问评论纽约时报关于 Jensen Huang 第二次婚外情的报道，猜测这件事会不会成为戳破 AI 泡沫的导火索。（原文：BREAKING: 🇺🇸 Nvidia CEO Jensen Huang had a second affair with VP of Health Care Kimberly Powell, per the New York Times.…）
- @DrJimFan 引用 @DrJimFan：作者展示 GR00T 2 驱动真实灵巧手完成化学实验并自行清理的全程视频，称其为自主、未剪辑、1 倍速。引用内容称该模型为面向机器人的开放 System One 模型，完全在机器人在载算力运行，支持 20 步以上、5 分钟以上的长程灵巧操作与全身控制。（原文：Introducing GR00T 2: our most advanced, open System One model for robotics. It runs entirely on your robot’s onboard com…）
- @gregisenberg 引用 @bot：作者认为 Grok Bot 是当前最好的 AI 智能体编排器，并称自己见到的这套配置为最佳：工程环节运行 Claude Code，设计环节使用 Paper。引用内容介绍 Grok Bot 1.0 为可与整个团队协作的 AI 队友，能在 Slack、Teams 共享并受邀参会。（原文：Introducing Grok Bot 1.0. AI teammates that work alongside your whole team.
Share Bots, collaborate with them in Slack …）
- @kimmonismus 引用 @kimmonismus：Dario Amodei 称癌症、阿尔茨海默病、心脏病和精神疾病可能不到十年就成为历史。作者认为这发生在 Anthropic IPO 前，并且觉得这是真的，不只是 IPO 炒作。（原文：Looking crazy good for Anthropic: Its IPO is already multiple times oversubscribed, according to Bloomberg, with investo…）
🧩 跨推文归纳
共识
- AI 正从对话工具进入具体工作流：Gemini 决策 API、ChatGPT 财务功能与 GR00T 2 长程操作都强调可执行任务。
- AI 智能体生态加速成形：机器人模型、编排工具与 B2A 流量争夺共同指向 agent 成为基础设施。
- 对 Gemini Decision 1 的评价相互印证：官方强调批量校准决策，独立评测确认基准领先但成本偏高。
分歧
- 对 AI 参与游戏制作的判断分歧：一方认为零 AI 辅助的 AAA 游戏将消失，另一方认为反 AI 审查误伤独立工作室。
- 对黄仁勋去留的态度不一：有人认为个人失误不应否定其价值，有人把丑闻视为可能戳破 AI 泡沫的导火索。
数据来源：X 列表「AI 英文」（22 条候选）· 由 twscrape 采集、grok 整理
© 版权声明
文章版权归作者所有，未经允许请勿转载。
相关文章
暂无评论...