Q1 2026

Welcome to the first issue of Stack Quarterly. We started this publication because the existing AI / agentic press is split between marketing-driven roundups that nobody believes and academic papers that nobody applies. We thought there was a slot for a quarterly that wrote for the engineer who has to make stack decisions and live with them.

The lead of Issue 1 is the 2026 agentic stack survey. The piece is a qualitative landscape map of what working teams are actually shipping with, drawn from a winter of practitioner conversations and our own audits of public engineering writeups. We deliberately did not invent the percentages that have made every previous survey of this market embarrassing in retrospect. We will keep not inventing them.

Reza’s MCP essay is the practitioner version of his opinion on the Model Context Protocol after a year of running it in production. It is the kind of piece we want this publication to be known for: long, opinionated, written by someone who has actually shipped on the technology, and willing to say what he disagrees with about it.

The workforce-OS piece is the one that took the longest to write and is the most contested inside the editorial group. We argue that “stitching together LangChain and a vector store” is not a workforce-OS architecture, and we name what we think is. We are interested in the corrections.

Q1’s slate is four pieces. We are deliberately starting small. Issue 2 lands the first week of April with a larger slate as the publication grows into its rhythm.

Thanks for being here at the start.

— The Editor