I really liked this post in that it seems to me to have tried quite seriously to engage with a bunch of other people’s research, in a way that I feel like is quite rare in the field, and something I would like to see more of.
One of the key challenges I see for the rationality/AI-Alignment/EA community is the difficulty of somehow building institutions that are not premised on the quality or tractability of their own work. My current best guess is that the field of AI Alignment has made very little progress in the last few years, which is really not what you might think when you observe the enormous amount of talent, funding and prestige flooding into the space, and the relatively constant refrain of “now that we have cutting edge systems to play around with we are making progress at an unprecedented rate”.
It is quite plausible to me that technical AI Alignment research is not a particularly valuable thing to be doing right now. I don’t think I have seen much progress, and the dynamics of the field seem to be enshrining an expert class that seems almost ontologically committed to believing that the things they are working on must be good and tractable, because their salary and social standing relies on believing that.
This and a few other similar posts last year are the kind of post that helped me come to understand the considerations around this crucial question better, and where I am grateful that Nate, despite having spent a lot of his life on solving the technical AI Alignment problem, is willing to question the tractability of the whole field. This specific post is more oriented around other people’s work, though other posts by Nate and Eliezer are also facing the degree to which their past work didn’t make the relevant progress they were hoping for.
I really liked this post in that it seems to me to have tried quite seriously to engage with a bunch of other people’s research, in a way that I feel like is quite rare in the field, and something I would like to see more of.
One of the key challenges I see for the rationality/AI-Alignment/EA community is the difficulty of somehow building institutions that are not premised on the quality or tractability of their own work. My current best guess is that the field of AI Alignment has made very little progress in the last few years, which is really not what you might think when you observe the enormous amount of talent, funding and prestige flooding into the space, and the relatively constant refrain of “now that we have cutting edge systems to play around with we are making progress at an unprecedented rate”.
It is quite plausible to me that technical AI Alignment research is not a particularly valuable thing to be doing right now. I don’t think I have seen much progress, and the dynamics of the field seem to be enshrining an expert class that seems almost ontologically committed to believing that the things they are working on must be good and tractable, because their salary and social standing relies on believing that.
This and a few other similar posts last year are the kind of post that helped me come to understand the considerations around this crucial question better, and where I am grateful that Nate, despite having spent a lot of his life on solving the technical AI Alignment problem, is willing to question the tractability of the whole field. This specific post is more oriented around other people’s work, though other posts by Nate and Eliezer are also facing the degree to which their past work didn’t make the relevant progress they were hoping for.