Mastodon Feed
Boosted by pixxl:
ajroach42@retro.social ("Andrew (making toys)") wrote:
While this is not the only or the worst thing about LLMs, it is also a problem with LLMs.
They are machines designed to predict what is statistically likely to follow what they have been prompted with, based on what they were trained on. There's no magic, there's no intelligence. It's just spicy autocomplete.
Just another way tech corporations are trying to push us towards mediocrity.