Cutting Token Cost in Parallel Claude Code Runs
Caching, batching, and model routing to keep parallel Claude Code agents cheap and fast on desktop — with a 6-step plan and cost breakdown.
8 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 143 of 631
Agentic AI & LLMs
Agentic AI, LLM engineering, and the models behind modern automation — multi-agent systems, LLM evaluation and comparisons, RAG, fine-tuning, AI infrastructure, security, and production AI engineering.
From the blog
Latest articles
Debugging Parallel Claude Code Agents on Desktop
Find and fix loops, wrong tool calls, and hallucinated args in parallel Claude Code agents on desktop — concrete tactics, hooks, and a debug workflow.
8 min read
Prompt and Context Design for Parallel Claude Agents
Design context for parallel Claude Code agents: what to include, what to cut, why isolation beats sharing, and how to order prompts for focus.
8 min read
Wiring MCP Servers Into Parallel Claude Code Agents
Wire MCP tools into parallel Claude Code agents safely: shared auth, tight schemas, idempotency keys, structured errors, and write serialization.
8 min read
Reusable Patterns for Parallel Claude Code Agents Today
Code-level patterns for parallel Claude Code agents: result contracts, scope tokens, read-only briefings, planner orchestrators, and map-reduce.
8 min read
Build Parallel Agents in Claude Code: A Walkthrough
A hands-on walkthrough of building a parallel-agent workflow in Claude Code on desktop: define subagents, fan out, collect JSON results, retry, and verify.
9 min read
Claude Code Desktop Architecture for Parallel Agents
How Claude Code on desktop runs parallel agents: orchestrator loop, isolated subagent contexts, the tool bus, locks, and shared-state merge.
8 min read
Anthropic and Moody's Data Partnership: Why Grounding Matters in Finance
Anthropic and Moody's announced a data partnership in May 2026 that grounds Claude in audited financial reference data. Why grounding reduces hallucination and what it unlocks.
7 min read
Why Anthropic Restricted Mythos: The Dual-Use Calculus in 2026
Anthropic chose not to release Mythos publicly. Inside the dual-use cybersecurity calculus, what restricted release means for enterprises, and the ripple effects.
7 min read