Quarterly deep dives on the tools real teams actually ship with.

Stack Quarterly is a publication for the engineers who have to live with their stack choices for the next four quarters. Each issue is a small, opinionated bundle of long-form pieces: how working teams are putting AI and agentic systems into production, which abstractions actually survive contact with real workloads, and which “default” tools deserve to keep their slot in the chart.

We are practitioner-first. The writing is dry, the code samples run, and the claims that can be checked are checked. We are skeptical of stack-survey statistics that nobody can reproduce, of benchmark numbers without a methodology section, and of “AI for X” pitches that ship a single API call as a category. When we do not know a number, we say so. When a tool is overrated, we say that, too.

We publish four issues a year and run rolling pieces between issues. The archive runs deep on agentic orchestration patterns, MCP in production, vector DB and inference infrastructure, observability for LLM applications, and the working tooling choices of teams shipping on the new infrastructure.