This comment surprised me “Advocacy is riskier than the average grant”.
Yes, it might be riskier than the average GHD or Animal welfare grant, but I would have guessed that advocacy would be less risky than than technical AI safety grants. You’ve illustrated some ways that advocacy may have caused real world changes. I doubt many technical AI safety can concretely point to a way that they may have made the world even a little safer from AI.
Is MIRI now not largely an advocacy organisation as well now, emerging from previous technical work?
Two things here. When I use risky, I’m using it in the sense of “this action could cause net-negative consequences in the world” rather than risky in the “risky bet” sort of sense where risky means something like “high odds this doesn’t actually work out”. In the first sense I think most (but not all) technical work seems to be less risky than advocacy work, but I totally agree with you that it’s not clear in the second sense.
The other thing is that I’m using advocacy in this post to mean a more narrow version of the word, what’s also know as direct lobbying or “arguing for a cause or policy and trying to influence decisionmakers towards that policy”. MIRI’s current work is certainly advocacy in the broad sense, but (to my knowledge) they’re not engaging much in advocacy in this more narrow sense that I’m focused on here.
Oh sorry—yeah that was my laziness I didn’t even read your explanation properly. I’m used to the “risky bet” terminology in global health. And yes I understand your narrative framing better now.
This comment surprised me “Advocacy is riskier than the average grant”.
Yes, it might be riskier than the average GHD or Animal welfare grant, but I would have guessed that advocacy would be less risky than than technical AI safety grants. You’ve illustrated some ways that advocacy may have caused real world changes. I doubt many technical AI safety can concretely point to a way that they may have made the world even a little safer from AI.
Is MIRI now not largely an advocacy organisation as well now, emerging from previous technical work?
Two things here. When I use risky, I’m using it in the sense of “this action could cause net-negative consequences in the world” rather than risky in the “risky bet” sort of sense where risky means something like “high odds this doesn’t actually work out”. In the first sense I think most (but not all) technical work seems to be less risky than advocacy work, but I totally agree with you that it’s not clear in the second sense.
The other thing is that I’m using advocacy in this post to mean a more narrow version of the word, what’s also know as direct lobbying or “arguing for a cause or policy and trying to influence decisionmakers towards that policy”. MIRI’s current work is certainly advocacy in the broad sense, but (to my knowledge) they’re not engaging much in advocacy in this more narrow sense that I’m focused on here.
Oh sorry—yeah that was my laziness I didn’t even read your explanation properly. I’m used to the “risky bet” terminology in global health. And yes I understand your narrative framing better now.