I Tested OpenAI's Dots vs. Meta's Muse. What You Need to Know.

Summary
Nate Herk from the AI Automation Society provides a hands-on comparison between OpenAI’s Dots and Meta’s Muse personal AI agents. He evaluates both tools across their user interfaces, model performance benchmarks, third-party messaging integrations (Slack vs. WhatsApp), and subscription pricing tiers, concluding that while Dots runs on a more capable frontier reasoning model (GPT-6 Astra), Muse offers a significantly more polished, intuitive consumer experience.

What is shown

• Feature Comparison Matrix: A breakdown comparing Dots and Muse across background agent work, core models, local vs. cloud computer environments, deliverables, and communication channels [01:37].
• Muse Artifacts & Memory System: Demonstration of Muse’s interface, including the artifacts library (generating interactive web dashboards and PDFs) and inspectable/editable SOUL.md and MEMORY.md markdown files [02:30, 03:17].
• Communication Integrations: Live demonstrations of interacting with Dot "Herk" inside Slack DMs and team channels [04:44], contrasted with chatting live with Muse "Charlie" via WhatsApp [05:31].
• Pricing & Quota Overview: A plan table comparing Muse Free ($0), Power ($20/mo), and Maximum ($100/mo) against ChatGPT Pro tiers ($100, $200, and $500/mo) [07:00].
• Benchmark Comparisons: A side-by-side metric table comparing GPT-6 Astra against Muse Spark 1.3 on the Artificial Analysis Intelligence Index, knowledge-work ratings, coding benchmarks, and generation speeds [07:29].
• Muse Productivity Features: Walkthrough of Muse's proactive 30-minute Heartbeat scheduler, goal tracker, curated interest feed, and AI podcast generation workflow [08:48, 09:46, 10:35].
• Dots Interface & Codex Integration Bugs: Overview of Dot's voice-call feature, remote browser/terminal desktop, and clunky thread routing issues when coordinating with project folders in Codex [10:57, 12:12].

Claims & numbers

• The presenter notes Meta’s Muse quickly reached #1 on the Apple App Store after its launch [00:15].
• The presenter confirms Dots is powered by GPT-6 Astra, while Muse runs on Muse Spark 1.3 [01:49].
• Monthly pricing tiers displayed:
  • Muse Free: $0 for up to 100M tokens/week [07:00].
  • Muse Power: $20/month for 500M tokens/week [07:00].
  • Muse Maximum: $100/month for 3B tokens/week [07:00].
  • ChatGPT Pro: $100/month (Pro 100), $200/month (Pro 200), and $500/month (Pro 500, includes Astra Ultrafast) [07:00].
• Benchmark comparisons between Astra and Spark 1.3 shown by the presenter:
  • Overall Intelligence Index: Astra 53 vs. Spark 1.3 48 [07:37].
  • Knowledge-work rating (AA-Briefcase): Astra 1,569 vs. Spark 1.3 1,586 [07:42].
  • Professional-work rating (GDPval-AA): Astra 1,542 vs. Spark 1.3 1,672 [07:42].
  • SaaS automation score: Astra 68% vs. Spark 1.3 58% [07:54].
  • Terminal/coding benchmark: Astra 59% vs. Spark 1.3 33% [07:56].
  • Generation speed: Astra 51 tokens/sec vs. Spark 1.3 184 tokens/sec [07:58].
• The presenter highlights that Muse's background "Heartbeat" runs automatically every 30 minutes to check connected Google Docs and emails [08:52].

Notable quotes

• "Dots fits your current work environment... whereas Muse is much easier to evaluate for free... It just has a better experience." [06:30]
• "If I was building a bunch of little automations and scripts that I wanted to run... I'd probably choose Dots. But... for general knowledge work... I would probably be pretty comfortable using Spark." [08:01]
• "Dots does not even know how it works... so how do I, you know, it just, I start to lose trust over what kind of knowledge it actually has." [12:38]

Assessment
This is an authentic hands-on product comparison and review based on live testing of both systems. The screencasts show real UI interactions, live message exchanges in Slack and WhatsApp, and authentic workflow friction without obvious deceptive editing or synthetic staging.

Described by gemini-3.8-flash on 2026-10-01 from the video's audio and frames.