Quote:
Originally Posted by ovillo
I'm not inherently anti-AI, but there is a difference between a tool that allows for illegal uses (like Calibre allowing users to share copyrighted media) and a tool that is built upon illegal activities (like an LLM that uses copyrighted media as training data). So I see no hypocrisy in drawing a line between them.
|
That's actually a good argument.
Personally I don't much care. LLMs are not the only invention we use that was built upon illegal activity. Some medical solutions, for example, were built upon the experimentation that Nazi Germany and the Japanese conducted on prisoners during WWII. Often very sadistic experimentation. And yet these solutions have saved many lives later.