My idea-generation rate has generally been faster than my idea-execution rate. AI coding equalises or even inverts this relation.
Curiosity/creativity/agency and Claude Code are complements, not substitutes. Increasing use of one increases use of the others and viceversa.
Meatproxification is a real risk but not inevitable, a human-machine relation doesn't need to be reduced to executing LLM output in meatspace. The relation can actually enrich the human (not sure the machine cares either way)
Organisations that limit the agency of their employees are rate-limiting themselves and might get outcompeted by ones that don't.