Code with Claude: The 5 biggest updates explained

Key takeaways

Cloud Code now supports routines for scheduling tasks, automating workflows based on time or webhooks.
Outcomes enable defining success criteria for AI agents using rubrics, which agents will then iterate to meet.
Multi-agent orchestration allows programmatic setup of teams of specialized agents working together against the same problem.
Dreams introduce a new way to review and write agent session memories, aiding long-term agent behavior.
Usage limits have increased, making it easier to experiment and develop with Anthropic’s platform.
Overall, Anthropic is deploying practical, immediately usable features and shaping the future of agentic AI systems.

Summary

Introduction and Event Highlights

Clairvo opens by sharing her excitement about attending 'Code with Claude', Anthropic's early developer event, and hints at the new features announced. She aims to provide a quick, under-10-minute overview of the key updates and her thoughts on their potential impact for builders.

Claude Code Routines: Automating Scheduled Tasks

The first major update is the introduction of routines in Cloud Code, allowing users to trigger actions on a schedule or via webhooks. Claire demonstrates setting up a weekly newsletter draft, emphasizing how this helps automate tasks that previously needed manual initiation. She explains the trigger types (cron, HTTP, GitHub webhook) and discusses potential use cases like PR monitoring.

Outcomes and Managed Agents: Defining Success for AI Agents

Claire delves into the new 'Outcomes' feature in Anthropic's cloud API, akin to the goal product in Codex, which uses a rubric to evaluate agent success. She explains how users can define success criteria via markdown files and have agents iterate up to 20 times to meet the rubric. The multi-agent framework allows creating orchestrator and specialist sub-agents, supporting complex AI task forces.

Dreams: Enhancing Agent Memory Systems

The 'Dreams' feature is introduced as a primitive for managing agent memory. It enables reviewing past sessions and writing important memories to disk, similar to how humans consolidate memories through dreaming. Claire notes this is currently in research preview and discusses its implications for agent behavior and long-term memory management, including the potential for agent forgetting mechanisms.

Increased Usage Limits and Future Potentials

Anthropic has doubled Cloud Code's usage limits across major tiers, removing peak hour restrictions and increasing API model rate limits, facilitating more extensive experimentation and development. Claire reflects on the broad range of announced features, their practical utility, and how Anthropic aims to position itself as a leading agent platform for builders.

Show notes

Claire breaks down the biggest announcements from Anthropic’s “Code with Claude” event and what they actually mean for builders shipping AI products today. From scheduled AI routines to outcome-based agents, multi-agent orchestration, and new memory systems, Claire walks through the features she’s most excited to use immediately—and how they could reshape the future of agentic software.

What you’ll learn:

How Claude Code routines let you automate recurring workflows on schedules or webhooks
What “Outcomes” are and how rubric-based agent grading works
How multi-agent orchestration enables specialized AI teams with different roles and tools
Why Anthropic’s new “Dreams” memory system matters for long-term agent behavior
Why increased Claude Code usage limits are a bigger deal than they sound
How Claire thinks about building practical agentic products today

—

Resources:

• Code with Claude: https://claude.com/code-with-claude

• Claude Code Routines Docs: https://code.claude.com/docs/en/routines

• Define Outcomes Docs: https://platform.claude.com/docs/en/managed-agents/define-outcomes

• Dreams Docs: https://platform.claude.com/docs/en/managed-agents/dreams

• Multi-Agent Docs: https://platform.claude.com/docs/en/managed-agents/multi-agent

• Managed Agent Webhooks Docs: https://platform.claude.com/docs/en/managed-agents/webhooks#supported-event-types

• Codex (OpenAI): https://openai.com/codex

• GitHub: https://github.com

—

Where to find Claire Vo: