Peter Krupa on Nostr: with neural net machine translation, often the user isn't fluent in the target ...
with neural net machine translation, often the user isn't fluent in the target language. "looks good!" but how would they know?
with image generators, you get an image of an elegant old man riding a horse. "looks original!" but how do you know it's not a copy of something the model memorized?
with coding assistants, they can produce things that work, but that are insecure, dangerous, or inaccessible. "looks good!" but if you don't work in that language, how do you know?
with image generators, you get an image of an elegant old man riding a horse. "looks original!" but how do you know it's not a copy of something the model memorized?
with coding assistants, they can produce things that work, but that are insecure, dangerous, or inaccessible. "looks good!" but if you don't work in that language, how do you know?