It's interesting how you've articulatd the subtle shift from gatekeeping to just getting status updates. Your point about supervision requiring understanding really resonates. I'm curious, how do you envision debugging or validating these increasingly large and opaque AI-made changes? Is it just blind trust eventually?
Hi there, On the practical front, I'm using design docs and test driven development to manage the work. Longer term, I suspect it'll work like how large organizations work. You'll have processes for these new kinds of workers. We're all learning how to incorporate them into our workflows....
It's interesting how you've articulatd the subtle shift from gatekeeping to just getting status updates. Your point about supervision requiring understanding really resonates. I'm curious, how do you envision debugging or validating these increasingly large and opaque AI-made changes? Is it just blind trust eventually?
Hi there, On the practical front, I'm using design docs and test driven development to manage the work. Longer term, I suspect it'll work like how large organizations work. You'll have processes for these new kinds of workers. We're all learning how to incorporate them into our workflows....