#nukethemoon What others do with subagents, I usually handle through separate windows

#nukethemoon What others do with subagents, I usually handle through separate windows: if I want to research something, I might do it in another terminal pane and then paste the relevant findings into the main one. This reflects a broader reality that the fundamentals of software engineering—problem decomposition, algorithmic thinking, and system design—remain essential for anyone who wants to lead. It also provides a much-needed reality check against hyperbolic claims of 10x or 100x productivity gains from AI, offering reassurance to engineers experiencing imposter syndrome in the face of aggressive AI marketing. Intelligent agents need to operate interactively in the world both to accomplish tasks and to learn, allowing us to gain speed without abandoning safety. Rather than spending excessive effort trying to formulate the perfect prompt to keep an agent working on a long-running task, there are often simpler, lazier workarounds. The sensible approach is to use Clanker where it genuinely helps—such as scripts, boilerplate, and repetitive work—while maintaining confidence in human abilities for complex problem-solving, system design, and the creative aspects of engineering that make the work satisfying. Most people assume that the idea or prototype is the difficult part, when in reality it may represent only one percent of the problem: the idea of going to the Moon is trivial, while actually doing it is extraordinarily hard. Clanker tools are valuable additions to our toolkit, but the core skills of software engineering—understanding problems, designing solutions, and building maintainable systems—remain fundamentally human endeavors.