ZeroHour
Story · 6 sources · 10 articlesfirst updated ()

Meta's Muse agent hits No. 2 on the US App Store as GPT-6 Astra tops ErdosBench and Cognition ships SWE-2

infoAI industryimportance 78
What's new: This update expands the story beyond Meta's Muse launch (the sole focus of the previous summary) to cover the surrounding AI news cycle: OpenAI's GPT-6 Astra topping the ErdosBench math benchmark (3.23, 106/226 problems solved) and OpenAI's stated pivot away from targeted math optimization toward recursive self-improvement, Cognition's adoption of GPT-6 Astra inside Devin for automated testing…
Merged summary · glm-5.3-flash · rewritten as coverage arrives

Meta launched Muse, a WhatsApp-controlled autonomous agent that climbed from 4th to 2nd on the US App Store with 83,000+ iOS downloads but took the band Muse's @muse handles and surfaced Instagram API data beyond ad-topic settings; in the same news cycle,…

Meta launched Muse, its first agentic AI productivity assistant: a WhatsApp-controlled autonomous agent that runs on its own cloud virtual machine, plans multi-step tasks, browses, fills forms, and negotiates on users' behalf, covering shopping, email management, trip planning, media generation, and webpage/document creation. It launched first in the US on iOS and Android. A second agent, Sentinel, gates all Muse network access and holds credentials - The Decoder cites Perplexity Comet's calendar-invite hijack as evidence agentic systems need isolation - and a Muse Confidential VM with user-held encryption keys is planned later this year. Payments run through Stripe's Link using one-time cards, which Meta calls the first AI agent covered by Link's purchase protection; Shop Pay and 1Password integrations are planned. Muse's model reportedly scored 44-48 on the Artificial Analysis Intelligence Index v4.3, up from 31 for Muse Spark in April and near GPT-5.6 Sol's 47, and a paid tier of up to $200/month was reported earlier. Muse asks for approval before purchases and sensitive actions, users control per-app access scopes and can revoke anytime, interactions can be used for AI training but are not shared with Meta's ad systems, and data can be deleted on request. The Verge's hands-on found Muse deleted thousands of promotional emails and completed an Amazon purchase, generating podcasts, images, videos, and artifacts while refusing some copyrighted-character and Apple CEO requests - but it also surfaced detailed personal interests inferred from Instagram and Facebook account API data that is not visible in the apps' ad-topic settings. Meta says Muse only exchanges data needed for third-party integrations and shares nothing with advertisers; the reviewer frames privacy unease as the main adoption hurdle. The launch reignited a naming fight: the @muse Instagram and X handles long used by the English rock band Muse, which trademarked its name in 1999, reportedly moved to @museband around June-July 2026, and Zuckerberg's launch posts mistakenly tagged the band's account (attributed to a bug). Meta - which previously acquired @Meta in 2021 and @threads in 2023 around major launches - has not explained the switch. Sensor Tower counted 83,000+ US iOS downloads as Muse climbed from 4th to 2nd on the App Store, far below Threads' 4.3 million launch-day downloads and ChatGPT's 500,000 first-week installs; the Android version ranks only No. 338 in Productivity on…

  • Meta launched Muse, its first agentic AI productivity assistant, controlled via WhatsApp and running on its own cloud virtual machine; it plans multi-step tasks, browses, fills forms, and negotiates for users across shopping, email, trip…
  • A second agent, Sentinel, gates all Muse network access and holds credentials; a Muse Confidential VM with user-held encryption keys is planned later this year, and The Decoder cites Perplexity Comet's calendar-invite hijack as evidence…
  • Muse payments run through Stripe's Link using one-time cards, which Meta calls the first AI agent covered by Link's purchase protection; Shop Pay and 1Password integrations are planned.
  • Muse's model reportedly scored 44-48 on the Artificial Analysis Intelligence Index v4.3, up from 31 for Muse Spark in April and near GPT-5.6 Sol's 47; a paid tier of up to $200/month was reported earlier.
  • Muse asks approval before purchases and sensitive actions; users control per-app access scopes and can revoke anytime, interactions can be used for AI training but are not shared with Meta's ad systems, and data can be deleted on request.
  • The Verge's hands-on found Muse deleted thousands of promotional emails and completed an Amazon purchase and generated podcasts, images, videos, and artifacts (refusing some copyrighted-character and Apple CEO requests), but it surfaced…
  • The English rock band Muse, which trademarked its name in 1999, reportedly lost its long-held @muse Instagram and X handles (moving to @museband around June-July 2026); Zuckerberg's launch posts mistakenly tagged the band's account,…
  • Sensor Tower counted 83,000+ US iOS downloads as Muse climbed from 4th to 2nd on the App Store, below Threads' 4.3 million launch-day downloads and ChatGPT's 500,000 first-week installs; the Android app ranks No. 338 in Productivity on…

Coverage timeline

  1. · 6d ago
    Hacker News · AI· 22
    Muse, the band, lost its social media handles to Muse, Meta's new AI agent

    Meta's new Muse AI agent took the @muse Instagram and X handles long used by the English rock band Muse.

  2. · 6d ago
    The Decoder· 72
    Muse can shop, write emails, and negotiate prices for users, all through WhatsApp

    Meta launched Muse, a WhatsApp-controlled agent running on an isolated VM with a Sentinel gatekeeper, able to shop, email, book travel, and negotiate.

  3. · 6d ago
    The Decoder· 68
    GPT-6 Astra gives mathematicians a breather, and OpenAI says that's by design

    OpenAI's GPT-6 Astra tops ulam.ai's ErdosBench math benchmark with 106 of 226 problems solved, while the company prioritizes recursive self-improvement over math optimization.

  4. · 6d ago
    The Verge · AI· 78
    Sure, Meta’s AI Muse works, but it sure creeps me out

    Hands-on review finds Meta's Muse AI agent completes shopping and email tasks but surfaces personal Instagram API data beyond user-visible ad-topic settings.

  5. · 6d ago
    Hacker News · AI· 72
    Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra

    Cognition released SWE-2, a coding model post-trained from Kimi K3 that scores 50.0% on FrontierCode 1.1 Main, near Fable 5.1 at 64% lower cost.

  6. · 6d ago
    Hacker News · AI· 62
    Cognition's SWE-2 achieves 92.8 on Terminal-Bench 2.1

    Cognition releases SWE-2, a 2.8T-parameter MoE coding model post-trained from Kimi K3, scoring 92.8 on Terminal-Bench 2.1.

  7. · 6d ago
    TechCrunch · AI· 32
    Meta’s AI agent Muse is now the No. 2 app in the US

    Meta's agentic AI app Muse ranked No. 2 on the US iOS charts with 83,000+ downloads, trailing Threads and ChatGPT launch pace.

  8. · 5d ago
    OpenAI News· 36
    Cognition helps Devin test its own work with GPT‑6 Astra

    Cognition integrates GPT-6 Astra into Devin, its CLI, and desktop products to automate testing and provide evidence for code review.

  9. · 4d ago
    The Decoder· 32
    GPT-6 Astra needs leaner prompts and fewer guardrails, OpenAI recommends

    OpenAI's Eric Provencher advises developers using GPT-6 Astra to shorten skill descriptions, trim AGENTS.md reading requirements, relax approval rules, and define clear completion goals.

  10. · 4d ago
    Hugging Face daily papers· 35
    Another Blueprint In The Wall: How to Ask Frontier AI Like a Kid?

    Six frontier models from OpenAI, Anthropic, xAI, and Google DeepMind converge on one imagined successor architecture when asked under a school-audience framing.