3 ms·
It follows just fine, I think, given that the possibility space is not limited to "create beneficial AGI" and "don't create AGI". It also includes "create unal
by comp_throw7 3y ago
It follows just fine, I think, given that the possibility space is not limited to "create beneficial AGI" and "don't create AGI". It also includes "create unaligned AGI", which is obviously much worse "don't create AGI"; the board would be remiss in its duties if it didn't try to prevent that from happening.
- concinds 3y agoCould we all stop entertaining these LessWrong hypotheticals? The notion that the board wants to shut down OpenAI to stop non-existent AGI is mad.