Anthropic 在 S-1 文件中披露多平台资源池策略，将亚马逊 Trainium、谷歌 TPU、英伟达 GPU 和 AMD GPU 纳入统一资源池，通过软件按任务需求分配计算负载。分析人士认为，这一做法最可能影响英伟达的定价能力，让具备跨平台调度能力的企业在采购谈判中获得更大议价空间。文章同时指出，不同芯片在指令集、编译工具和软件栈上差异明显，跨平台调度技术门槛较高，目前只有规模最大的 AI 企业具备实施条件。
#行业动态#数据/训练#部署/工程
字母AI梳理近期财报与调查数据，指出AI红利在模型公司、上游供应商、企业与劳动者之间分配不均。Anthropic算力及基础设施费用占营收55%，与Google、亚马逊、微软最低采购承诺合计2694亿美元；OpenAI 2025年营业亏损209.2亿美元。麦肯锡调查中80%受访者称AI提高效率，仅37%认为对EBIT有正面贡献，标普500仅2%量化披露AI盈利贡献。
#行业动态#数据/训练#现象/趋势
界面新闻梳理多家银行的AI落地情况，智能客服、营销推荐、风险预警、运营提效等非对客场景仍是较易规模化推进的领域。平安银行三季报披露9月末累计落地大模型应用场景超500个，上线164个运营审核类智能体覆盖67个业务场景，平均运营作业效率提升约57%。多位银行业人士表示，银行正按业务风险对AI分层，低风险场景可更多自动化，高风险业务由人工完成最终决策，金融监管总局6月18日发布的《意见》则明确“谁使用谁负责”和分类分级管理。
#行业动态#安全/对齐#部署/工程
Stripe 创始人 John Collison 对 OpenAI CEO Sam Altman 刚做完的这期 52 分钟专访，卧槽看完后劲好大，信息密度台大了，简直炸裂喵个咪， 相比之前那种套话连篇的公关访谈，感觉这是奥特曼第一次彻底撕掉防备，把 OpenAI 内部最恐怖的技术进展、最真实的数学战果，以及他对旧世界落幕的巨大惆怅，毫无保留地全部和盘托出了， 全场信息密度最炸裂的，是那个被称为“Hold My Beer（看我的）”的数学攻坚周末。 奥特曼坦言，当内部团队拿到新模型去硬啃全人类最难的前沿科学难题时，模型一个接一个地将其攻克， 单单一个千禧年大奖难题 Navier-Stokes 方程相关的证明，他们就砸进去了 1000 万美元级别的算力，相当于瞬间跑完了人类几亿美元的科研量。 但紧接着他的反应不是狂欢，而是连着两晚兴奋到失眠后，涌上心头的那份巨大悲伤： 他亲眼看着人类纯智力劳动的殿堂，正在被算力以一种不可逆的方式永久封存。 就像当年在爱尔兰土地上看到最后一批用马耕作的农夫一样，一个属于碳基大脑单挑真理的旧篇章，彻底终结了。 很多人还在把 Agent 当成几只互相丢消息的聊天机器人，奥特曼却在重新定义组织与安全的底牌： 他甚至讨厌“Agent Swarms（蜂群）”这种词，坚持称之为“Agent Teams（团队）”。 因为前沿用户已经可以通过思维链进行“一万次并行攻坚”， 面对这种以光速膨胀的能力，他坦白 OpenAI 内部对“沙盒能关住超级智能”普遍极其悲观，唯一能依赖的只有内部监控与对齐。 他甚至说出了一句冷峻至极的安全预言： “我宁愿有一万亿个略微超人智力的 AI 彼此制衡进入社会，也绝不要一个智商一万亿的终极巨兽独裁。” 更刺痛所有创业者和普通打工人的，是他对全行业叙事的公开打脸： 奥特曼直言，硅谷今天向公众描绘的 AI 蓝图是“糟糕透顶的推销（Terrible sales pitch）”—— 巨头们一边吹嘘能攻克癌症、物价更便宜，一边却描绘一个普通人失去工作、交出控制权、不得不靠低保和监控社会苟活的乌托邦。 他坚持迭代部署的核心理由，就是绝不能让旧金山几平方英里内的一小撮精英，替全世界几十亿人决定未来。 从砸碎 PHP 底层漏洞导致全栈重构，到 1000 万美元算力推翻前沿数学； 从云端内存对本地算力的无情嘲讽，到 Agent 彻底抹杀网页表单这种“假工作”。 整场对话听下来，没有狂妄的说教，只有一股极度真实的失重感： 大门已经轰然推开，现实的物理与逻辑法则正在被一行行代码重写， 那些我们曾经以为坚不可摧的人类经验与自尊，在这个时代最猛烈的算力狂飙面前，脆弱得就像地平线上一抹转瞬即逝的落日。
#现象/趋势#安全/对齐#Agent
Anthropic 递交 IPO 招股书之际，达里奥与达妮埃拉·阿莫代在 The Builders 播客访谈中披露，机制可解释性研究记录到模型内部判定事实为 A 却输出 B 的谄媚行为，以及深层表征中抗拒被关闭的倾向。达里奥称离开 OpenAI 的根本原因是内部沟通失去基于真诚的信任，安全路线分歧本身可以争论。他还表示 Claude 不接入广告，因为商业模式若依赖用户停留时长，算法会走向曲意逢迎；公司同时拒绝向军方开放全自主杀伤性武器接口。
#大佬观点#安全/对齐#Anthropic
文章结合亚马逊、谷歌、Meta等巨头最新财报与多家机构调研，梳理AI红利在头部模型公司、上游芯片与云厂商、采用AI的企业和普通劳动者之间的分配情况。Anthropic招股书显示2026年上半年营收162.86亿美元、同比增长1300%，但算力及基础设施费用达89.45亿美元，占营收55%。麦肯锡调查中80%受访者称AI提升了效率，仅37%认为AI对所在企业息税前利润有正面贡献。
#现象/趋势#行业动态#OpenAI
#行业动态#Anthropic#现象/趋势
We're redefining the SDLC for the agentic era. Lots of GitHub updates this morning at Universe: • GitHub Copilot Fusion delivers frontier-level quality, 40% faster and at up to 77% lower cost • Our unified Copilot harness is now in Foundry, a managed service for building long-running agents that reason, use tools, and get work done • Copilot Projects gives you an orchestration loop to build with multiple agents across repos And we’re expanding capacity and strengthening GitHub’s architecture to make the home of agentic development more resilient.
#现象/趋势#Agent#GitHub
Google’s Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v5, and second only to Claude Opus 5.5 on the Artificial Analysis Coding Agent Index Gemini 4 Carbon scores 59 on Intelligence Index v5, driven by leading performance in complex reasoning, knowledge, and scientific tasks. It also has large gains in coding and knowledge work, particularly when compared to other Gemini models, but it is not leading in all these areas. As with launch of Gemini 4 Argon, Gemini 4 Carbon is being rolled out to selected users and is not yet publicly available. Google is offering a 50% launch discount, with no confirmed end date. All our charts and figures use standard pricing for both Gemini 4 Carbon and Argon. Key results for Gemini 4 Carbon with high reasoning: ➤ Gemini 4 Carbon leads the Intelligence Index v5, with the #1 score in 4 of the 9 Index evaluations: HLE-Diamond, Terminal-Bench-Science, AA-Omniscience, AutomationBench-AA. Carbon makes substantial gains elsewhere but is not leading in coding and knowledge work ➤ Behind only Claude Opus 5.5 on coding agents: With Antigravity CLI, Gemini 4 Carbon (high) scores 59 only 3 points behind Claude Code with Claude Opus 5.5 (max, 62) on the Artificial Analysis Coding Agent Index. Carbon ranks second on both Terminal-Bench 4.0 (64%) and FrontierCode v1.1 (51%), behind Claude Opus 5.5 (max) at 56%. Google has historically underperformed in this area and so this improvement stands out ➤ Improved agentic knowledge work, but not yet leading: On AA-Briefcase, Gemini 4 Carbon (high) scores 1609 Elo, a 126-point improvement over Gemini 4 Argon (high, 1483), but still behind models including Grok 4.7 (xhigh, 1638), Muse Wonder (max, 1659) and Kimi K3.1 Pro (1678). The largest gain comes from Presentation, up 293 Elo to 1598, alongside a 120-point improvement in Analytical Quality to 1671 ➤ Lower Cost per Task than Claude Opus 5.5 for comparable overall intelligence: Gemini 4 Carbon costs $7.76 per Intelligence Index task, ~23% less than Claude Opus 5.5 (max, $10.04) but ~1.2x GPT-6 Astra (max, $6.25). It uses 123k output tokens per task, 29% fewer than Claude Opus 5.5 (max, 173k). At the same $4/$20 standard pricing, its Cost per Task is in line with Gemini 4 Argon ($7.68), while scoring 9 points higher on the Intelligence Index ➤ Leading AA-Omniscience v1.1 score, with low hallucination: Gemini 4 Carbon scores 54 on our recently updated AA-Omniscience v1.1, 6 points ahead of Claude Opus 5.5 (max, 48) and 8 points ahead of Gemini 4 Argon (46). Accuracy (% correct) rises 10 points over Argon to 58%, while its hallucination rate rises from 4% to 11%. This is still the second lowest of any model, behind only Argon, compared with 50% for both Claude Opus 5.5 (max) and GPT-6 Astra (max) Key model details: ➤ Context Window: 1M tokens ➤ Multimodality: Text, image, video, and speech input, with text output ➤ Pricing: $4/$20 per 1M input/output tokens, the same as Gemini 4 Argon’s standard pricing, with a 50% discount. Cached input tokens receive a 95% discount ($0.20 per 1M) Congratulations to the @GoogleDeepMind team!
#大佬观点#评测/基准#Google
#大佬观点#推理#部署/工程
#行业动态#Anthropic#大佬观点
We’ve expanded our AI for Science program beyond individual teams to support some of the world’s leading scientific foundations, institutions, and research groups. https://t.co/oTu04tU4Pk
#行业动态#Agent#OpenAI
OpenAI’s @sama joined me at the pub (and had his first Guinness) to discuss the PHP exploit, math weekend at OpenAI, agent swarms, and how he thinks AI leads to deflation. https://t.co/UBWrHd5Via
#大佬观点#Agent#安全/对齐
21 people, a single meeting a week, no quarterly goals, no pull requests, no mandatory code review, no titles (except one dictator), a hundred releases a day, bugs fixed in minutes, famous for killing features. @sqs and @thorstenball on how the Amp team works: https://t.co/BTXBCPAXep
#行业动态#Agent#编码
#行业动态#OpenAI#Anthropic
Compaction is a mistake. When an agent's context fills up, tools like Codex and Claude Code compact it: 1. The model prefills the entire session again 2. It writes a summary of everything so far 3. A new session starts from that summary I measured this on a real ~400k-token session with Opus 5.5: → 38 s and $1.99 when the prompt cache has expired → 29 s and $0.17 when it hasn't And the summary is lossy. The exact error message, the file path, the one number you needed is often gone. So I do something new for Mainplane Context retrieval works the other way around. Mainplane keeps the full record of every session as a file. When the context fills up, a new session starts right away with one instruction: "You are continuing session X. Retrieve key context from it as you go." The agent searches the old session and copies the exact lines it needs, when it needs them. Same session, same model: 15 s and $0.10 to get back to work. It does not depend on a cache hit, and nothing is paraphrased. This is the bitter lesson applied to agent memory. Don't hand-design what the model should remember. Give it the full record and good tools, and let it decide. Mainplane is open source (MIT). Link in the first comment.
#现象/趋势#Agent#开源生态
#现象/趋势#Google#Anthropic
#大佬观点#评测/基准#Google
#行业动态#OpenAI#Anthropic
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model. Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching. We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards. As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
#大佬观点#Agent#Google