This is a special post for quick takes by edge_retainer. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.
is there a term for ai related works ratio of increasing safety to capabilities?
is there a list of the types of ai work that have the best ratio?
e.g. you would expect governance to probably have the best ratio
It seems to me that the hardest part of getting ai right will be coordination and governance, not technical alignment. Under this view, I think the EA/LW forums are net negative. They are really bad at creating coordination environments and are actively crowding out similar but more coordination/governance environments from existing.
1Can you think of any examples of successful social movements whose main medium of communication was something akin to a forum? Maybe something involving riots?
To be more object level, doing things requires making not perfect decisions but making them with conviction. A panopticon where peoples explicit goal is to be scout mindset (at best) or pedantically criticize at worst is not going to foment do-acracy. It also requires sub groups/committees, incentivizing doing, etc. I can flesh this out more if you aren't getting the spirit of what I mean.
re crowding out, I think these forums are on net substitutes not complements for other forms of organization. I don't have specific evidence here and am not that confident though.
I think I get the spirit. I thought, from the way you phrased it, that you were making an empirical claim (eg: ‘I saw [x better thing] not happen because LessWrong something something’). But it sounds like you’re making a more conceptual argument (which I think has legs, but for different reasons than you do, and I think we disagree on the effect size).
Thanks for responding.
requires sub groups/committees, incentivizing doing, etc.
So, more like the Bolsheviks, then? The Sons of Liberty?
Mothers against Drunk Driving?