Building with AI agents in 2026: the field guide
Everything we have learned shipping autonomous agents into production: architecture patterns, evals, guardrails, cost control and the failure modes nobody writes about.
Read the guide
Insights
Thinking out loud
Notes from the studio on AI, product and engineering. Written by the people doing the work.
Designing AI agents that actually ship
Guardrails, evals and the unglamorous work behind reliable autonomy.
Read articleRAG in production: lessons from 12 builds
Chunking, reranking and why your retrieval layer decides everything.
Read articleThe 2026 stack for AI-first SaaS
The tools, models and patterns we reach for on every new build.
Read articleShipping LLM features users actually trust
Latency, honesty and fallbacks: the details that decide adoption.
Read articleDesign systems for AI-first products
Streaming states, confidence cues and interfaces that explain themselves.
Read articleFrom prototype to production agents
The gap between a demo that wows and an agent that works every day.
Read articleGet new essays in your inbox
One email a month on building AI-first products. No noise, unsubscribe anytime.