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 41 to 50 of 130 articles
Keep reading
AI in Practice
A ChatGPT Site can host a supported plugin, but the useful question is what its tools should be allowed to do. Here is a practical contract-and-test process before you share.
Read the articleAutomation and Agents
A Team Bot crosses several access boundaries. Map each one, test shared chats separately, and keep credentials and high-impact actions narrowly scoped.
Read the articleAutomation and Agents
A shared AI coworker is only useful when the team has a clear job, shared reference material, a checkable result, and someone who will keep the context current.
Read the articleSystems and Workflows
A screenshot request is incomplete until its destination is explicit. Give agents an approved private path, restrict public actions, and test the fallback before an agent reaches for a personal repository.
Read the articleAutomation and Agents
A repository search is not a complete audit. Start with the people who could publish, inventory public surfaces using metadata, and route suspected screenshots through a controlled response process.
Read the articleAutomation and Agents
Plan a bounded Auto Router test with eligible models, privacy checks, spend limits, quality sampling, session affinity, stop rules, and rollback.
Read the articleAI in Practice
Cloudflare reports lower costs with trade-offs in task success. Read the benchmark limits and decide whether your workload merits a pilot.
Read the articleSystems and Workflows
Choose an image workflow by where people need to review and revise, not by the model name alone. Then compare Flare and Sunburst on identical work.
Read the articleAI in Practice
ChatGPT Images 2.5 may make reference-led editing more reliable. A repeatable source, edit, inspect and approve loop still matters more than trusting a promising output.
Read the articleSystems and Workflows
DeepSeek's release plan changed. Verify what each model identifier resolves to, test the behaviors your agent relies on, and keep rollback simple.
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
