Sample content · easy to replace
news · November 1, 2026
This Week in Agents (sample strip)
Sample weekly roundup format for AI and agent news. Replace with reported items; no invented launches or stats.
Publication · Directory · Research
A calm, sourced site for curious kids, older and non-technical readers, builders, and researchers — covering news, tools, papers, shopping, task performance, harnesses, and education.
Four doors in
Ages of wonder
Friendly explanations, pictures in words, and experiments you can try with a grown-up.
Plain language
Large, readable writing for anyone who wants to understand agents without the jargon.
Practitioners
Harnesses, tools, evals, and how working systems are put together.
Frontier
Paper breakdowns, original notes, and benchmarks written for people who live in the literature.
This Week in Agents
Sample content · easy to replace
news · November 1, 2026
Sample weekly roundup format for AI and agent news. Replace with reported items; no invented launches or stats.
Sample content · easy to replace
guide · October 22, 2026
Sample guide to the loop around a language model: tools, memory, permissions, and evaluation — without pretending one stack is the winner.
Sample content · easy to replace
explainer · October 20, 2026
Sample explainer of what people mean by AI agents, how they differ from a chatbot, and which questions to ask next.
Directory
Flagship research
Sample content · easy to replace
original report
Sample original report on how this publication will read agent leaderboards: task match, contamination, cost, and human review — with pointers to public evals, not invented scores.
Sample content · easy to replace
paper breakdown
Sample editorial breakdown of Yao et al. 2022, a widely cited pattern for interleaving free-text thoughts with tool actions.
Coverage
A calm weekly strip of what actually moved in AI and agents.
Launches, product changes, and the directory of tools worth tracking.
Papers, reports, and what they change about how agents are built.
How agents search, compare, and buy — and where they still fail.
Computer-use, office work, coding, and the gap between demos and reliability.
The loops, tools, memory, and guardrails that turn a model into an agent.
Tutors, classroom tools, and how to learn with (and about) agents.
First-principles explainers: models, tools, planning, memory, and evaluation.
Newsletter
Form posts to an API route with a provider TODO. No keys in this repo. We will not invent subscriber counts.