Replying to Avatar utxo the webmaster 🧑‍💻

The thing about open source AI models is you can't exactly verify the data it was claimed to be trained on

It could be mass deployed by state actors to manipulate us under the guise of open source, omitting many parts of the reinforcement process and what data holds more weight etc

Basically just be careful of psyops ok anon
32
322905e2... 11mo ago

Seems we need new language to talk about supposedly open AI models. "Open source" doesn't even really make sense here. When people use that term they actually mean the output of the source code being open, not the inputs or even the code itself.

Reply to this note

Please Login to reply.

Discussion

No replies yet.