3 ms·
Reasonable humans understand the request at hand. LLMs just output something that looks like it will satisfy the user. It's a happy accident when the output is
by heavyset_go 11mo ago
Reasonable humans understand the request at hand. LLMs just output something that looks like it will satisfy the user. It's a happy accident when the output is useful.
- buu700 11mo agoSure, but that doesn't prove anything about the properties of the output. Change a few words, and this could be an argument against the possibility of what we now refer to as LLMs (which do, of course, exist).