While the topic of "safe reasoning" may seem more or less preliminary before a good implementation of reasoning, it remains a theoretical discipline with its ow

While the topic of "safe reasoning" may seem more or less preliminary before a good implementation of reasoning, it remains a theoretical discipline with its own importance and should be studied alongside the rest, also largely irregardless if its stage. > We have no idea what the fundamental architecture for superintelligence would even begin to look like Ambiguous expression. Not implemented technically does not mean we would not know what to implement.