Notes from the work
Build notes, practical systems, and a closer look at what changed. Written to help you make a decision, try something useful, or avoid learning the annoying part twice.
Showing 51 to 60 of 66 articles in Automation and Agents
Keep reading
Automation and Agents
Strands Decider 2B selects supplied options locally. Use five questions and three worked cases to decide which agent choice merits a contained trial.
Read the articleAutomation and Agents
Anthropic says Claude led 26% of its measured AI R&D work in August 2026. The figure depends on task ratings and staff-time weights; here is how to read it.
Read the articleAutomation and Agents
Use Anthropic’s scoped oversight measurements to ask what an agent can change, when people review problems, and what evidence supports wider permission.
Read the articleAutomation and Agents
A practical pilot for measuring Grok 4.7 on your coding work, with a fixed acceptance rubric, reviewer effort, and cost per accepted task.
Read the articleAutomation and Agents
An HTTP 202 can mean an Agents API browser approval was accepted without proving the task finished. Keep origin access, sign-in, action review, same-session recovery and system-of-record verification distinct.
Read the articleAutomation and Agents
OpenAI's managed Agents API browser adds another route for web tasks. This guide helps you decide when its execution boundary fits, what to compare, and how to test one workflow without overstating the announcement.
Read the articleAutomation and Agents
A reusable agent configuration is easier to share. The production question is how to change it without losing the ability to explain which setup handled a request.
Read the articleAutomation and Agents
Before widening a shared-channel coding-agent preview, test the context boundary, trigger permissions, app identity, repository scope, budget and human review path.
Read the articleAutomation and Agents
A chat mention is an intake decision. Use this practical test to turn a bounded thread into traceable GitHub work without handing an agent more context than the task needs.
Read the articleAutomation and Agents
Start an OpenAI Dot with one recurring, reviewable responsibility. Define its sources, stop conditions, and approval boundaries, then pilot it on known work.
Read the article
Follow the work
New builds, honest lessons, and practical ways to make more room for the work that needs you.
Free to read. Unsubscribe at any time. Privacy
