Anthropic's answer to Dots and Muse is already inside Claude

OpenAI launched Dots at DevDay this week. Each Dot is an always-on agent with its own cloud computer and browser, wired into the 4,000-plus apps connected to ChatGPT: in OpenAI's example, a Dot spots a bug alert in Slack and digs in. Meta shipped Muse before that, and xAI launched Grok Bot in August.

Anthropic's answer arrived two weeks ago as a Claude update, without a new name or mascot. The update folded Cowork, which has run scheduled jobs since July even after users close their laptops, into the main Claude app.

Anthropic shipped to the people it already had

The merged Claude rolled out to Pro and Max subscribers, with Team and Free plans later. Teams have had Claude Tag in Slack since June, a proactive teammate with an audit trail. Cat Wu said an internal Claude version accounts for about 65% of product teams' code changes. The New Stack argues it was the right call: Anthropic needs the people already building with Claude to hand it real, recurring work and keep coming back; repeat use matters more than a flashy launch.

Anthropic, though, has not put Claude where Muse lives: Muse works inside Meta-owned WhatsApp; Claude still asks you to open its app or tag it in Slack.

Muse is a hit; the others start at the paid end

Apptopia estimated 359,000 daily U.S. iOS users in Muse's first 12 days, against 231,000 for ChatGPT's iOS app at the same point. Sensor Tower counted 3.4 million-plus downloads by September 24; over 95% of Muse users also use Facebook. The app has a free tier, plans at $20 and $100 a month, and connectors to Shopify and Stripe. Scale has a downside: Amazon began blocking Muse two weeks after launch.

OpenAI and xAI started at the paid end instead. Dots requires a ChatGPT Pro plan at $100 to $500 a month, or a Business Premium or Enterprise account; Sam Altman called it premium because each Dot needs so much compute. Grok Bot reached 418,000 weekly users by mid-September, per Bloomberg, and xAI launched Team Bots this week.

Too early for a winner

The wave is young. Peter Steinberger's project Clawdbot became Moltbot, then OpenClaw, after a trademark request from Anthropic; it turned into a security headache, and its author joined OpenAI in February. This week the OpenClaw Foundation released a pre-1.0 enterprise version. Meta's Nat Friedman said Muse was heavily inspired by OpenClaw, and users found a nearly identical SOUL.md file inside it.

The real test is trust

An always-on agent holds credentials and acts while nobody is watching. xAI's own documentation tells users not to treat Grok Bots as a security boundary, and the OpenClaw Foundation says most IT departments ban agent platforms outright. Anthropic's defaults lean cautious: Claude asks before it acts unless told otherwise, and Tag keeps its own audit trail. If builders keep finding useful work to hand off, the race will be decided by which agent they trust with the next job.