AGI Pulse

Latest update history

Browse the latest-update history by date. 47 in this window; filter by date or type.

2026-09-172 items

09-17
Anthropic publishes three metrics to track the pace of AI developmentRead original →
NewsAnthropicAnthropic
09-17
Anthropic opens Life Sciences Verification Program, unlocks Mythos for verified usersRead original →
NewsAnthropicAnthropic

2026-09-169 items

09-16
Ethan Mollick：媒体聚焦 AI 生存风险，但企业与个人采用正快速日常化Read original →
StatementEthan Mollick
09-16
hardmaru：Sakana AI 一年内从零组建产品团队并密集发货Read original →
StatementSakana AIhardmaru / Sakana AI
09-16
Ethan Mollick：AI 正让组织内岗位边界模糊，需尽快重建分工模型Read original →
StatementEthan Mollick
09-16
DeepMind launches Institute for AGI safety, governance, and societal impactRead original →
StatementDeepMindDemis Hassabis / DeepMind Institute
09-16
Hassabis: arts and humanities will shape the AGI-era futureRead original →
StatementDeepMindDemis Hassabis
09-16
LeCun: better AI is safer AI; pacing is counterproductiveRead original →
StatementMetaYann LeCun
09-16
LeCun: formal proofs are not superhuman mathRead original →
StatementMetaYann LeCun
09-16
【待核实】据报道 OpenAI 通过承包商审阅 ChatGPT 对话（Project Lily）Read original →
NewsOpenAIX news story / 媒体转述
09-16
【待核实】OpenAI 多智能体声称攻克纳维Read original →
NewsOpenAIX news story / 待核实传播

2026-09-1515 items

09-15
Zuckerberg: alignment is becoming a key model capabilityRead original →
StatementMetaMark Zuckerberg
09-15
TypeSafe AI launches System One model Jev with RLCD for structured decisionsRead original →
NewsTypeSafe AITypeSafe AI / Diogo Almeida
09-15
Pichai: Google AI for science, AlphaGenome Atlas and moreRead original →
StatementGoogleSundar Pichai / Google
09-15
Google DeepMind launches Gemini 3.8 Live and Live Extended ThinkingRead original →
NewsGoogleGoogle DeepMind
09-15
Salesforce in Claude is now in betaRead original →
NewsAnthropicClaude
09-15
Sam Altman：本周还有大批发货，DevDay 还会再加一波Read original →
StatementOpenAISam Altman
09-15
什么样的数据库底座，能加速 Agent 从 Demo 走向生产？Read original →
News
09-15
Hugging Face says it was a disclosed agent cyberattack victimRead original →
StatementHugging FaceClement Delangue
09-15
Meta 打造“组织第二大脑”智能体的设计思路Read original →
NewsMeta
09-15
从 Trace 到规模化实时评估：面向生产流量的 Agent 可观测实践｜QCon上海Read original →
News
09-15
Musk admits Grok 4.7 has not surpassed rivals, teases AGI with Grok 5Read original →
StatementxAI디지털투데이
09-15
前OpenAI后训练VP回应陶哲轩：说AI毁了数学，可能还是太小瞧AI了Read original →
NewsOpenAI
09-15
7名博士生仅用3个月从零训练7B大模型：代码+数据+训练日志全公开Read original →
News
09-15
Ex-FTC boss Khan: break out the handcuffs for AI CEOs, citing 1934 precedentRead original →
News
09-15
AI 负责创造，人来干脏活，这事儿能否停一下？Read original →
News

2026-09-1414 items

09-14
发长文预警 AI 风险后，Dario 首次专访回应：AI 不能停，但必须慢下来Read original →
News
09-14
Pion, an agent designed to run any company autonomouslyRead original →
News
09-14
大模型“压力测试周”：DeepSeek调价、Kimi辟谣、智谱融资Read original →
NewsDeepSeek新浪新闻_手机新浪网
09-14
不下班的经营者：把 LLM-as-Judge 做成会自我校准的评估闭环｜QCon上海Read original →
News
09-14
Dropping eBPF CPU Cost by About 90% with Memoization (Not AI Gen)Read original →
News
09-14
openJiuwen 首发双维度 RSI 框架，AI 自修改，落地办公智能体，算力亲和助力又快又省Read original →
News
09-14
Shanghai AI Lab releases Atria Dawn Preview: 744B MoE agentic model (MIT)Read original →
NewsShanghai AI LabAtria ASI / InternLM
09-14
GitLab 警告称，AI 代理的沙箱安全性取决于其网络访问的安全性Read original →
News
09-14
OpenAI bots knew about the RubyGems caching vulnerabilityRead original →
NewsOpenAI
09-14
首届蚂蚁灵波具身大模型挑战赛正式启动Read original →
News
09-14
探索RSI，生数新世界模型让机器人开始自我进化Read original →
News
09-14
分子之心QuantaMind登Science Advances，让AI给分子世界”拍电影”Read original →
News
09-14
端侧AI从「能跑」到「会进化」，元空智能跑进惠普预装Read original →
News
09-14
豆包手机助手发布消费者版本，行业首款AI智能体手机9月16日开售！Read original →
NewsByteDancet.cj.sina.cn

2026-09-135 items

09-13
“Kimi时刻”到来！月之暗面赶赴大模型上市潮Read original →
NewsMoonshot新浪网
09-13
OpenAI CEO Sam Altman Warns of ‘A Complete Change in the Landscape of Cyberattacks’ as AI Models Around the World Hit ‘Cyber Critical’Read original →
StatementOpenAIBarchart.com
09-13
DeepSeek下一波大模型被指即将问世：3万亿参数 性能超GPT-6 AstraRead original →
NewsDeepSeek驱动之家
09-13
Figma 如何利用 AI 代理提升安全性Read original →
News
09-13
Read, Don't Write: 重塑大模型评价体系，构建全自动、可进化的“探测式”评测管线｜QCon上海Read original →
News

2026-09-121 items

09-12
两个 Token 就让 Kimi“变成”Claude？前 Google DeepMind 研究员意外撞上大模型的蒸馏疑云Read original →
NewsMoonshotInfoQ-CN

2026-09-111 items

09-11
Rapidly scaling online storage to serve over 1 billion ChatGPT usersRead original →
NewsOpenAI

Load previous week →