Every time you describe what you want an LLM to build, something is lost. Not because the LLM is dumb—because English is vague by design and code is precise by necessity. That gap is where AI coding breaks down at scale.
Context switching used to kill productivity and make me feel drained. 2pm would arrive for me to muster the energy to build context and start on deep work for the day. Now I'm managing a fleet of agents across 10-15 projects a day, learning faster than ever, and I'm exhausted.
In college they taught us waterfall was dead and agile was the future. Now I practice waterfall daily. Research and planning phases measured in hours instead of months. The highest leverage activity isn't writing code anymore—it's planning and context management.
When I started programming, learning my environment WAS the thing. Somewhere along the line, improving my dev environment felt like toil- wrestling with configs and bash scripts while juggling deadlines. Sometimes it was fun, often it was not. AI has completely changed that. My dev environment has become the playground for my imagination.