Anthropic has shipped eight major features in eight weeks, culminating in "computer use," allowing Claude to directly control a user's mouse, keyboard, and apps by viewing the screen like a human.
This suite of features (Claude Cowork, Code Review, Remote Control, Computer Use) collectively forms a potential AI-native operating system, termed "Claude OS," which may compete with macOS and Windows.
Anthropic's product velocity has outpaced OpenAI, which acquired OpenClaw (a similar agentic system) only four weeks ago; Anthropic built its own solution, banning OpenClaw users from integrating with Claude for safety and control.
The enterprise focus and security architecture of Anthropic are seen as key advantages for corporate adoption, contrasting with OpenClaw's perceived security concerns for governments and enterprises.
Market sentiment for an Anthropic IPO is extremely bullish: a venture fund (VCX) holding private Anthropic shares traded at a 15x return in five days, representing an 8x premium to its net asset value.
Prediction markets (Polymarket) imply an 82% chance Anthropic's IPO market cap closes above $600B and estimate OpenAI's at ~$952B, reflecting high investor expectations for both companies.
The "computer use" feature is currently slow, clunky, and macOS-only, with plans to expand to Windows; speakers advise caution and recommend testing with small tasks before granting full computer access.
A potential Anthropic app store (marketplace) for computer-use plugins could leverage its 19 million daily Claude Code users, creating a distribution advantage akin to Apple's early App Store.
Speakers compare the AI OS landscape: OpenClaw as the open-source "Linux," Anthropic as "Microsoft" (or "Apple") with enterprise/marketplace focus, and OpenAI as "Apple" with a hardware angle.
Long-term, the speakers envision a future where AI agents like Claude become faster and more capable than humans at computer tasks, fundamentally changing human-computer interaction.