been running 3 agents in parallel for a client project this week
one on backend. one on frontend. one writing tests. sounded like a dream setup when i planned it
but here is what nobody tells u. the agents generate code fast but then u spend hours reviewing all of it. 3 agents means 3 codebases to verify. 3 sets of edge cases to check. 3 different approaches that u have to approve or reject
my brain was fried by wednesday. not from coding. from deciding. every 10 minutes its like should this function go here or there. is this approach better or that one. yes no fix this. repeat all day
read something about this. they called it decision fatigue. the bottleneck moved from writing code to judging code. and judging has no natural stopping point. u can always review one more file
honestly miss the days when i just wrote the code myself and knew every line. now i manage 3 agents and understand maybe 60 percent of what they produce
anyone else feeling this or is it just me being a boomer about it
Top comments (0)