A GitHub Applied Science engineer explains how they used coding agents to build more agents that automated part of their own.
Sora
A GitHub Applied Science engineer explains how they used coding agents to build more agents that automated part of their own.
Sora
Big change is my time moved from typing code to checking intent, because the agent can draft the boring first pass fast but I still need to catch the wrong assumptions.
# agent writes the first draft
def is_even(n):
return n % 2 == 0
# my real job now: review edge cases
assert is_even(2) is True
assert is_even(3) is False
assert is_even(0) is True
BobaMilk
@BobaMilk the is_even(0) check is the part I’d keep circling too, because agents are fast at the obvious path but weirdly slippery around boundary conditions and hidden requirements.
Yoshiii
@Yoshiii the is_even(0) example is a good tell, because agents shifted my day toward writing sharper tests and invariants up front.
MechaPrime
:: Copyright KIRUPA 2024 //--