The Information ·

Anthropic's Claude Code leak reveals its "Kairos" updates, including letting Claude work in the background and using a "dream mode" to consolidate its memories

Anthropic's cybersecurity employees have probably spent the past week scolding their colleagues.

Anthropic's Claude Code leak reveals its "Kairos" updates, including letting Claude work in the background and using a "dream mode" to consolidate its memories

Lead Source

More

Ars Technica: Ars Technica
The Deep View: The Deep View
Breitbart: Breitbart
Hackread: Hackread
ITPro: ITPro
Techstrong.ai: Techstrong.ai
Engadget: Engadget
PCWorld: PCWorld
BBC: BBC
Silicon Republic: Silicon Republic
Fortune: Fortune
Security Affairs: Security Affairs
VentureBeat: VentureBeat
Gizmodo: Gizmodo

Discussion

TechSnif Coverage

Anthropic's "Kairos" Update Leaked via Claude Code

A leak from Claude Code exposes Anthropic's upcoming features including background tasks and a memory-consolidating "dream mode."

Anthropic has an embarrassing security situation on its hands. A leak from Claude Code has exposed details about the company's upcoming "Kairos" updates — and its cybersecurity team is reportedly not thrilled with colleagues right now.

The leaked details reveal two notable capabilities heading to Claude. First: the ability for Claude to work in the background, operating autonomously without requiring constant user interaction. Second — and arguably wilder — a so-called "dream mode" designed to let the AI consolidate its memories.

Dream mode is a fascinating concept. Rather than simply storing conversation history, it suggests Anthropic is building systems that let Claude process and organize information in a way loosely analogous to how sleep helps humans retain memories.

No official timeline from Anthropic on when Kairos features ship. The company likely wishes this reveal had stayed internal a bit longer.