> What makes this argument particularly poisonous, with or without AI, is that it is kinda true.
"End-to-end does everything" also glosses over the specificity of results. It'll tell you something is wrong but with much less detail as to what or how.
Even if we assume the future belongs to a narrative-chaos engine doing fixes, I would prefer it fixed just the actual problem instead of re-architecting everything until big broad tests stop failing.
2 comments
[ 0.25 ms ] story [ 15.7 ms ] threadI'd love to hear other people opinions on this piece, it's my first long form article so I'd appriciate some feedback!
Thanks!
"End-to-end does everything" also glosses over the specificity of results. It'll tell you something is wrong but with much less detail as to what or how.
Even if we assume the future belongs to a narrative-chaos engine doing fixes, I would prefer it fixed just the actual problem instead of re-architecting everything until big broad tests stop failing.