been reading this whole AI slop debate that blew up this week
boris cherny (the claude code guy) replied to a dev who emailed him about it. said throwaway code can be black box but production code from AI needs a higher bar than human written stuff
been feeling this hard. when i ship client work with AI agents i still review every line. lint rules tests security checks automated reviews the whole pipeline
model writes fast but someone gotta hold the quality bar. thats still us
my job lowkey feels like managing a junior dev who types at lightspeed but needs supervision on everything lol
do u guys full trust your AI code or verify everything
Top comments (0)