#architecture

Six-hour agent runs are now real. The harness — checkpoints, durable state, recovery — matters more than the model. A field guide to the long-running pattern.

Anthropic just shipped 10 financial services agent templates. The interesting part isn't the templates — it's the three-layer architecture quietly becoming the standard for enterprise agents.

Deep dive into voice agent architectures. Why cascaded models give you control and fused models trade complexity for naturalness. What we're learning from shipping production agents at scale.

Caching, routing, distillation, and per-task model selection. The four moves that take a $0.40/task agent to $0.04/task without anyone noticing the difference.

The three-role pattern that powered Manus, then LangChain Deep Agents, then half the production agents shipping in early 2026. Why it works, when it doesn't, and how to actually build one.

Three kinds of memory, four-tier consolidation pipelines, hybrid retrieval that beat pure vector search, and the unsolved staleness problem. The 2026 state of the art for agents that learn over time.

Google's Agent Development Kit isn't another agent loop. It's an event-driven runtime, a multi-language SDK, and a deploy target that lines up with Cloud Run and GKE. Here's how the pieces fit.

The orchestrator-worker pattern that took over agent design in late 2025 is the same pattern that took over backend design in 2014. The wins are real. So are the failure modes.

Six serious frameworks, four orchestration styles, and one tired question I keep getting asked. Here's the decision tree I actually use.

LangGraph turns the agent loop into an explicit state graph — nodes, edges, reducers, checkpointers. The mental model, the type system, and the patterns that make it production-grade.

LangChain is no longer the agent framework — LangGraph is. But its primitives (runnables, retrievers, callbacks) became the connective tissue underneath. Here's what to keep and what to drop in 2026.

What actually goes inside a production AI agent — the reasoning loop, the tool layer, memory, control flow, and the surface where models meet your business systems.