Loop Engineering
A practical way to turn one-shot agent prompts into feedback loops that plan, build, verify, and improve.
WRITING
Long-form work on AI engineering, agent design, interfaces, and the decisions behind useful software.
A practical way to turn one-shot agent prompts into feedback loops that plan, build, verify, and improve.
Why an explicit product and implementation spec gives an AI coding loop something reliable to execute.
Memory is not remembering everything. It is deciding what remains useful after the current conversation ends.
A practical way to turn one-shot agent prompts into feedback loops that plan, build, verify, and improve.
Why an explicit product and implementation spec gives an AI coding loop something reliable to execute.