Skip to content

#部署/工程

3 today

Today · Nov 17Tuesday · 3 items

Hacker News front page

Just 400 US Data Centers Use More Electricity Than Michigan

哈佛发表于 Nature Sustainability 的研究发现,美国约400座超大规模数据中心一年的用电量与密歇根州相当,多数电力来自化石燃料。研究结合行业资料与卫星数据,对2024年5月至2025年4月运营的403座设施进行建模,发现数据中心排放达到AI热潮前的3倍,弗吉尼亚、俄勒冈、俄亥俄和艾奥瓦四州占研究用电量的50%以上。样本并非全部用于AI,研究也未纳入数据中心现场自建电厂的供电。

Ars Technica · AI

US moves to tighten chip controls, targeting remote Chinese use of third-country AI compute

The US Commerce Department is preparing tighter rules on advanced chips, aimed at stopping China and other countries from renting restricted chips in third-country data centers to train and run AI models. The rule takes effect 90 days after publication in the Federal Register. Firms in high-diversion-risk countries must get licenses to move large volumes of advanced chips and run customer checks, due diligence and monthly transaction reports. Compute dedicated to commercial recommendation systems is exempt, and enforcement remains disputed inside the government.

Why it matters: The Commerce Department is pulling remote compute in third-country data centers into chip controls, so firms serving Chinese customers face licenses and ongoing due diligence.

Yesterday · Nov 16Monday

TechCrunch · AI

US moves to tighten AI chip export rules, targeting Chinese firms' remote access to overseas compute

The Trump administration is preparing new AI chip export controls that would require overseas data centers serving companies in restricted countries such as China to obtain licenses for the export, re-export or transfer of the chips and servers involved. The rule covers data centers or networks above 12 million TPP of total processing performance, including indirect access gained through intermediaries. Exporters and operators must review customer ownership and activity. Existing operators have 90 days after the rule takes effect to get licensed; commercial recommendation systems are excluded.

Why it matters: The US plans to pull overseas data centers serving restricted-country firms into export licensing, extending compliance to customer ownership checks and indirect access channels.

Nov 14Saturday

TechCrunch · AI

Jensen Huang's exit leaves investors and founders worried about a funding gap for AI startups

Jensen Huang has stepped down as Nvidia CEO. Investors and founders interviewed fear the company will pull back on investing, hitting AI startup funding, valuations and deals. PitchBook data shows Nvidia took part in 105 funding rounds this year and was often the largest investor in them. Those interviewed expect interim CEO Colette Kress to take a more conservative investment approach.

Why it matters: Investors expect Nvidia to tighten startup investing after Huang's departure, putting funding and valuation pressure on AI infrastructure companies that rely on its money and backing.

Computing Life · Share · Yage

微软邮箱故障为何拖累办公软件与三大 AI 助手超过 26 小时?

微软云端邮箱故障持续超过 26 小时,波及办公软件及 Copilot、Claude、ChatGPT 获取企业资料的功能。网络配置变更触发潜伏缺陷,安全核验程序过载与客户端自动重连使恢复受阻,团队通过扩容、限流、修补组件和逐步恢复连接处理故障。邮件、日历及权限服务的共同依赖让多家 AI 助手连接器同时受影响,微软通报称故障与 AI 智能体无关。

Ars Technica · AI

FAA expands AI advice to manage air traffic congestion across entire US airspace

美国联邦航空管理局(FAA)将 SMART AI 空管建议系统扩展至全国全部 22 个空管中心,以减少延误和燃料消耗。FAA 称,这套耗资 $875 million 的系统自 9 月 21 日在华盛顿特区上空启用以来,已帮助 180 个航班改道,平均每班节省 10 minutes 和近 80 gallons 燃料。

Nov 13Friday

Computing Life · Share · Yage

Why nearly half of leaked configs were committed to GitHub by Claude Code itself

Claude Code's "don't ask again" option writes commands containing plaintext credentials into .claude/settings.local.json. Apiiro found that 48.8% of traceable leaking repos had that config uploaded by commits signed with Claude.

Why it matters: Claude Code saves approved commands, passwords included, to a local config, so project-level ignore rules and env-var injection cut the credential exposure when that config gets committed.

OpenAI News

Gondola doubles revenue as its team builds more with Codex

Gondola 使用 Codex 后,团队月度工程产出提高约 25%,公司营收同比增长超过 2 倍,未增加薪资成本或显著开销。业务团队用 Codex Cloud 制作可运行原型,再交由工程师完善上线;首席工程师按合并代码变更衡量的月度产出提高约 50%,工程团队人数未变。团队还用 ChatGPT 准备提案、管理日程和汇总跨系统信息,重要工作在对外发送或用于业务决策前仍由人工审核。

Ars Technica · AI

FAA expands AI advice to manage air traffic congestion across entire US airspace

美国联邦航空管理局(FAA)将 SMART AI 空管建议系统扩展至全国全部 22 个空管中心的空域,用于减少拥堵、延误和燃油消耗。FAA 称,这套耗资 $875 million 的系统自 9 月 21 日在华盛顿特区上空部署以来,已帮助 180 个航班调整航线,平均每班节省 10 分钟和近 80 加仑燃油。

TechCrunch · AI

SoftBank’s billion-dollar loss may have put a cap on AI’s natural gas ambitions

软银周二宣布季度亏损 $7.1 billion,其中超过一半来自发电设备相关损失,其押注 AI 天然气供电的策略遭遇挫折。软银据报为超过 13 gigawatts 的燃气轮机支付了不可退还的订金,订购量比所需高出近 50%,如今将其中大量订金视为无价值。

Nov 12Thursday

TechCrunch · AI

Meta is losing its longtime head of infrastructure

Meta 确认,长期基础设施负责人兼工程联合负责人 Santosh Janardhan 将卸任,由今年加入公司的 Dave Brown 接任。Janardhan 将留任完成交接并转为 Meta Compute 顾问,预计明年离开;Meta Compute 负责扩大公司的数据中心投资并制定相关战略。

TechCrunch · AI

CoreWeave turns to customer prepayments, shorter contracts to fund AI growth

CoreWeave正增加客户预付款、按特定合同抵押借款及短期交易,以支持AI基础设施扩张并减少对无担保债务的依赖。公司第三季度收入为$3.66 billion,同比增长168%,净亏损为$709 million,季度利息支出为$846 million,预计未来3个月获得约$9 billion客户预付款和保证金。

Nov 11Wednesday

Simon Willison

Anthropic IPO roadshow video details multi-chip strategy and compute efficiency

Anthropic began its IPO roadshow yesterday, and the roadshow video covers how Claude is trained and deployed across multiple chips, plus methods for getting more out of compute.

Why it matters: The roadshow video compares a multi-chip adaptation that took dozens of engineers months with the same job done by Mythos in three days.

The Decoder

Anthropic says its AI moved Claude to AMD GPUs in 3 days, undercutting CUDA's moat

Anthropic says its own model, Mythos, got Claude running on AMD GPUs in 3 days with almost no human intervention. The model ported or wrote nearly 100% of the kernels, measured on prompt processing on AMD MI355X. Adapting a new accelerator by conventional means takes 6 to 12 months.

Why it matters: Anthropic claims Mythos cut new-hardware adaptation from 6 to 12 months to 3 days, comparing model-written kernels with conventional engineering.

The Decoder

Claroc finds Grok-built apps exposing patient data, plaintext passwords and SpaceXSI sales pipeline

Security firm Claroc checked 37,571 Grok-generated apps that are still online. Of those, 35,578 could be opened without logging in, exposing patient records, plaintext passwords and SpaceXSI internal sales data. Claroc traces the leaks to default sharing settings and broken permission checks.

Why it matters: Claroc traced the exposure to default sharing settings and permission-check bugs, and outside accounts could read, change and delete team data after logging in.

Hacker News front page

Tinfoil (YC P25) is hiring an applied AI security engineer in San Francisco

Tinfoil 正在旧金山招聘应用 AI 安全工程师,该岗位需作为战略客户的主要技术顾问,把客户的威胁模型与隐私需求转化为基于 Tinfoil 的部署架构。要求 2 年以上软件工程、前向部署工程、安全工程或技术产品经理经验,并具备 Python 或 TypeScript 生产级编程能力及 LLM 智能体、工作流或 LLM 产品的实操经验。

OpenAI News

Singapore Airlines elevates travel & work with OpenAI technology

新加坡航空通过 OpenAI API 和模型升级客服聊天机器人 Kris,客户满意度(CSAT)翻倍,客服代表平均处理时长缩短约 10%。基于 OpenAI API 的 Flight Recommender 让用户用自然语言描述行程,过去一年月活用户增长近 3 倍。OpenAI 模型还用于内部 AI 平台 Jarvis 及软件开发流程。

TechCrunch · AI

Google broadens its AI hardware portfolio with two new specialized chips

Google 发布第八代 TPU 的两款新变体 TPU 8tx 和 TPU 8i Max,分别面向机密推理与大内存负载。TPU 8tx 与 AMD 合作设计,基于 TPU 8t,将安全防护做到硬件层,并配备可在训练同时运行推理的定制推理网络,支持即时蒸馏小模型。TPU 8i Max 则将两颗 TPU 8i 合入单一封装,内存容量、算力与带宽翻倍。

Nov 10Tuesday

Hacker News front page

Fireworks AI security incident may affect under 1% of Notion workspaces

Notion's investigation confirmed that a Fireworks AI security incident may affect under 1% of Notion workspaces, with some Notion AI chat records possibly exposed. The records carry no user or workspace labels and contain no tool-call results, and other workspace data is unaffected. Notion has disconnected from Fireworks and disabled Insights.

Why it matters: The exposed chat records lack user and workspace labels, but the text itself may hold sensitive credentials, so affected customers should treat it as exposed data.

Hacker News front page

ClickHouse Acquires Xata

ClickHouse 收购 Xata,计划将即时数据库分支、闲置计算缩容至零及智能体服务层整合进 ClickHouse Managed Postgres。Xata 表示,即使数据库包含数 TB 数据,创建独立分支也只需约 3 秒,可供每个拉取请求或智能体运行隔离测试。Xata 客户短期内可照常使用,平台继续接受注册,待功能达到一致后再迁移至 ClickHouse Managed Postgres。

Hacker News front page

Google launches TPU 8tx and TPU 8i Max for confidential inference and ultra-long context

Google expanded its eighth-generation TPU lineup with TPU 8tx for secure confidential inference and TPU 8i Max for ultra-long context. TPU 8tx uses the same chip as TPU 8t and adds hardware-level protection for data and model weights, cutting latency to as little as half of TPU 8t. It supports fine-tuning on private data, knowledge distillation and reinforcement learning.

Why it matters: Google built specialized TPU systems by reusing an existing chip and changing security, networking and memory settings, shortening hardware adaptation from years to months.

Nov 9Monday

The Decoder

Google ships four TPU variants in one year as it ramps up competition with Nvidia

Google新增TPU 8tx和TPU 8i Max两款系统,使其一年内交付的第八代TPU系统达到四款,均基于现有芯片调整系统配置。TPU 8tx面向机密计算,Google称数据和模型权重在处理期间仍受到保护与隔离;TPU 8i Max将两颗TPU 8i芯片放在同一封装内,在相同机架规格下将算力、内存容量和内存带宽翻倍。

Hacker News front page

Apple capped engineers at a $200 daily AI budget

Apple 据报道将部分工程师的每日 AI 支出上限从 $300 下调至 $200。Stork AI Daily 将这一调整解读为企业规模使用生成式 AI 的成本压力,并主张通过缓存、小模型路由和提示词优化减少支出;预算调整的信息来源标为 Bloomberg。

Nov 7Saturday

TechCrunch · AI

Crusoe sells troubled Wyoming data center project to Google

Crusoe 确认已将怀俄明州东南部一个未完工的数据中心项目出售给 Google,由后者接手开发。据报道,Google 曾担忧 Crusoe 管理下的项目成本与施工进度;收购价未披露,Google 宣布计划投入 $17 billion 开发该园区。

Hacker News front page

Google has acquired Crusoe's data center site in southeast Wyoming

Google 已收购 Crusoe 位于怀俄明州东南部的数据中心选址,交易完成标志着 Crusoe 在该项目上的前期开发工作结束,包括选址、政府审批和电力策略。Crusoe 称其已签约超过 6 GW、管线超过 40 GW 算力容量,并正在开发 Google 位于得州 Armstrong County 的在建数据中心园区。

Nov 6Friday

The Verge · AI

Democrats win control of Congress, putting AI regulation and data center buildout under pressure

Democrats won control of the House and Senate on Tuesday, leaving the AI industry facing tighter regulation, slower data center construction and pressure for safety legislation. Of 44 local ballot measures to pause data center construction, 43 passed. Texas governor-elect Gina Hinojosa promised to pause all data center approvals until the state legislature passes laws protecting residents. Experts surveyed expect federal attempts to block state-level AI limits to get harder, and tech accountability and AI safety legislation to keep advancing.

Why it matters: Voters passed 43 of 44 local data center pause measures, so AI companies' compute expansion plans face more direct local approval constraints.

Nov 5Thursday

The Decoder

After midterm defeats, OpenAI offers community pledges to win support for AI infrastructure

OpenAI issued a voluntary community pledge after US voters rejected new data centers in 40 of 41 local referendums. It covers future company-owned sites: OpenAI pays all power and grid-upgrade costs, cuts use at peak times, and publishes build plans, noise studies and water data. Whether partner-built facilities such as Oracle's are covered is unclear. The announcement came two days before the midterms, and OpenAI says its own projects may face delays.

Why it matters: OpenAI's pledges to cover power costs and publish water data apply only to new sites it owns and operates, leaving partner-held infrastructure unclear.

MIT Technology Review · AI

Texas AI data center boom may stall after election result

Democrats won in Texas, and governor-elect Gina Hinojosa backs a full pause on data center construction. That could slow existing AI projects and block new ones, though local permitting power and building unions' positions will shape how far a state-level limit goes.

Why it matters: A winning Texas Democrat wants a data center construction pause, and local permitting plus union stances will decide how much of it lands.

Nov 4Wednesday

Hacker News front page

North Korean fake job interview led to $6.3B Binance heist

朝鲜黑客通过虚假招聘面试诱使 Binance 工程师在工作电脑下载恶意代码,最终实施上周的 $6.3B 加密货币盗窃。Binance 的事件报告称,攻击者于 October 20 利用受感染电脑和内部 AI 工具访问钱包工程师设备,随后在 October 29 使用窃取的凭证部署容器并执行盗币脚本。

Hacker News front page

AI stocks slide as Texas governor result raises data center risks

支持暂停建设 AI 数据中心的民主党候选人 Gina Hinojosa 被预测赢得得州州长选举,数据中心政策风险推动 AI 相关股票隔夜下跌。Nasdaq-100 期货下跌0.8%,Robinhood 平台上的 CoreWeave、Nebius Group 和 Oracle 分别下跌7%、4%和4%。

TechCrunch · AI

Data centers are turning to old jet engines as the wait for power grows

部分 AI 数据中心正将退役飞机的喷气发动机改装为天然气发电设备,以绕过电网接入和新涡轮机供应的多年等待。据《纽约时报》报道,xAI 已在密西西比州安装八台改装机组,为孟菲斯的 Colossus 2 数据中心供电;分析师称这类机组可让数据中心在 24 个月内运行,而北弗吉尼亚接入 100 megawatts 电网容量约需七年。

Ars Technica · AI

Ukraine’s drone strikes spur hasty Russian plans for underground data centers

乌克兰无人机袭击损坏俄罗斯人工智能超级计算设施并中断数百万用户的云服务,促使俄罗斯科技企业加固数据中心并筹划将其转入地下。国有电信公司 Rostelcom 已与矿业公司 Norilsk 探讨迁移部分数据中心,另有地区官员提议在露天煤矿内建设数据中心、在上方种植公共绿地作为掩护。克里姆林宫要求数据中心业主为国家紧急需求预留算力,并已授权在必要情况下使用境外数据中心。