AGI Pulse

Latest update history

Browse the latest-update history by date. 46 in this window; filter by date or type.

2026-09-194 items

09-19
Qwen3.8-LiveTranslate: real-time simultaneous interpretation across 60 languagesRead original →
Model releaseAlibabaQwen
09-19
[Editorial] Chinese X is arguing one thing: after chat, agent capability is outrunning trustRead original →
PolicyAGI Pulseagipulse editorial
09-19
[Editorial] Claude Code reads AGENTS.md: the story is agent-instruction convergence, not one fewer fileRead original →
ProductAnthropicagipulse editorial
09-19
[Editorial] What the Jev frenzy is really about: agents lack a cheap decision layerRead original →
Model releaseTypeSafe AIagipulse editorial

2026-09-186 items

09-18
Anthropic partners with Accenture on embedded independent AI evaluationRead original →
PolicyAnthropicAnthropic
09-18
Google AI weekly recap: Gemini 3.8 Live, Dreambeans, CC, Google Pics and AlphaGenome AtlasRead original →
ProductGoogleGoogle AI
09-18
Meta SAM 3.1 lands on Meta Model API for detection and segmentationRead original →
ProductMetaMeta for Developers
09-18
MiniMax open-sources Code CLI under MIT with FrontierHarness SOTA claimRead original →
ProductMiniMaxMiniMax
09-18
MiniMax joins Singtel AI Pass for Singapore SkillsFuture learnersRead original →
ProductMiniMaxMiniMax
09-18
Qwen launches Qwen3.8-Omni-Flash, an agentic omni-modal modelRead original →
Model releaseAlibabaQwen

2026-09-175 items

09-17
Muse personal agent launches on MacRead original →
ProductMetaMuse
09-17
Anthropic cuts cost of biomolecular open-source models; launches protein design contestRead original →
PaperAnthropicAnthropic
09-17
OpenAI launches Astra for Law on GPT-6 AstraRead original →
ProductOpenAIOpenAI
09-17
Claude Projects run parallel cloud threads from one Claude Code chatRead original →
ProductAnthropicClaude
09-17
DeepMind’s AlphaGenome Atlas used to find disease-linked DNA variantsRead original →
PaperGoogle DeepMindGoogle DeepMind

2026-09-1615 items

09-16
Schmidhuber：发布 IDSIA-9-26，梳理 RSI 自 1987 以来路径；软件 RSI 已实用，完整 RSI 需自改进硬件Read original →
TheoryJürgen Schmidhuber
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
Meta：Muse Code 原生支持 Windows（无需 WSL），含 x64/ARM64 与零管理员安装Read original →
ProductMetaMeta for Developers
09-16
OpenAI：公布模型错位（misalignment）披露框架，并同步发布六例报告Read original →
PolicyOpenAIOpenAI
09-16
Anthropic：Claude Cowork 与聊天合并为统一 Claude，并上线 Docs/Slides/DesignRead original →
ProductAnthropicAnthropic / Claude
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
OpenAI introduces Sponsored Agents and new ChatGPT Ads tools with HubSpot and ShopifyRead original →
IndustryOpenAIOpenAI
09-16
OpenAI：ChatGPT Work / Codex 管理后台上线用量—成本Read original →
ProductOpenAIOpenAI
09-16
Mistral and Mozilla bring private, multilingual AI to Firefox Smart WindowRead original →
IndustryMistralMistral AI / Mozilla
09-16
OpenAI 经济研究：工人如何用 AI 解锁新工作方式Read original →
IndustryOpenAIOpenAI
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

2026-09-1510 items

09-15
Zuckerberg: alignment is becoming a key model capabilityRead original →
StatementMetaMark Zuckerberg
09-15
Brockman: ChatGPT Work can act on company data and dashboardsRead original →
ProductOpenAIGreg Brockman
09-15
Pichai: Google AI for science, AlphaGenome Atlas and moreRead original →
StatementGoogleSundar Pichai / Google
09-15
Odyssey unveils Odyssey-3 foundation world modelRead original →
Model releaseOdysseyOdyssey
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
Hugging Face says it was a disclosed agent cyberattack victimRead original →
StatementHugging FaceClement Delangue
09-15
Anthropic：Claude for Small Business 新增工作流与 27 项集成，覆盖获客到结账Read original →
ProductAnthropicAnthropic / Claude
09-15
LeCun: LLMs alone are not a path to human-level AIRead original →
TheoryMetaYann LeCun
09-15
Musk admits Grok 4.7 has not surpassed rivals, teases AGI with Grok 5Read original →
StatementxAI디지털투데이
09-15
Google Faces Red Flags in AI Development as Doubts Grow Over Delayed High-Performance Model ReleaseRead original →
ProductGoogleasiae.co.kr

2026-09-145 items

09-14
Arm 推出 AI Portal：当 AI 应用从“模型可用”走向“平台可用”Read original →
Product
09-14
OpenAI may have violated California’s AI safety law with latest model releases, AI watchdog saysRead original →
PolicyOpenAIFortune
09-14
国内首个团队智能体“豆包工作伙伴”发布_财经_24H快讯Read original →
ProductByteDancebihainews.com
09-14
Why DeepSeek-V4.1-Flash Is Such an Exciting Open Model ReleaseRead original →
Model releaseDeepSeekKDnuggets
09-14
“豆包手机助手”消费者版本正式发布 - OSCHINA - 开源 × AI · 开发者生态社区Read original →
ProductByteDanceOSCHINA

2026-09-131 items

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

Load previous week →