 (effective accelerationism) - the belief that AI alignment is not something we need to worry about, as our benevolent and highly competent corporate/government overlords will do exactly as much of it as necessary, when necessary

 - the belief that sufficiently-advanced AI could kill us all, and therefore we need to devote a large amount of resources towards AI alignment research. May also include the belief that there should be a moratorium on cutting-edge AI gain-of-function research ("capabilities research" for short). Leading proponent of this view: Eliezer Yudkowsky

Side node: These are not actually the only two positions on existential risks to humanity from AIs, there are also 

, which holds that if unfeeling AIs replaced humanity, other things being equal, this would be an improvement in the very long term because it would mean less suffering, and 

, which holds that it is good for humanity to be obsoleted and entirely replaced by "superior" life forms such as AIs. So it's possible that an e/acc could switch to one of the two fringe positions instead of switching to Notkilleveryoneism - but the fringe positions are rare at the moment, and don't seem particularly compatible with e/acc.