/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

Anthropic updates Claude Managed Agents with “dreaming”, a scheduled process that reviews recent work and updates memory, available in research preview

In April, Anthropic launched the public beta of Managed Agents, its platform for running AI agents on its infrastructure.

The New Stack Frederic Lardinois

Context & Ripple Effects

Anthropic put Managed Agents into public beta in April as a hosted platform for developers building and deploying agents at scale. The new research-preview feature arrives as the company broadens Claude’s agent tooling, following agent teams in Claude Code and earlier APIs for execution, connectors, files, and prompt caching.

The coverage points to a consistent product arc: moving Claude from isolated model interactions toward systems that can coordinate work, retain useful context, and operate over longer-running workflows.

First-order effects

  • Managed Agents users can test a scheduled process that reviews recent sessions and updates an agent’s memory, reducing the need to manually preserve relevant context between work periods.
  • Anthropic differentiates its managed agent platform by making memory maintenance a platform capability rather than leaving all continuity logic to each developer.

Second-order effects

  • Developers evaluating hosted agent platforms will increasingly compare how reliably they manage state and continuity, not just model quality, tool access, or orchestration.
  • A memory-updating process can make agent behavior more useful across repeated tasks, while also raising the practical importance of controls over what information is retained and revised.

Third-order effects

  • If such features prove dependable, agent platforms may compete as managed runtime environments for persistent work rather than as APIs for one-off model calls.
  • The combination of parallel agent teams and maintained memory points toward more autonomous, multi-step software workflows, making observability and governance central platform requirements.

The trend: AI vendors are packaging the infrastructure for persistent, coordinated agents—memory, execution, tools, and orchestration—into managed platforms for production deployment.

Discussion

  • @claudedevs @claudedevs on x
    Agents write to memory in a session. Dreaming is a background process that reflects over many sessions to curate memories: it can edit them based on patterns, add new skills, or remove stale ones. https://platform.claude.com/ ...
  • @claudedevs @claudedevs on x
    In Claude Managed Agents, we've added multiagent orchestration, an outcomes loop for rubric-driven self-improvement, dreaming for self-learning, & webhooks. [video]
  • @unusual_whales @unusual_whales on x
    BREAKING: Anthropic has released a feature called “dreaming” which allows AI agents to self-improve