Upvote, disagree: Raising productively useful awareness looks like exactly the post you made. Insufficiently detailed awareness of a problem that just looks like “hey everyone, panic about a thing!” is useless, yeah. And folks who make posts here have warned about that kind of raising awareness before as well.
(Even as a lot of folks make posts to the ai safety fieldbuilding tag. Some of them are good raising awareness, a lot of them are annoying and unhelpful, if you ask me. The ones that are like “here’s yet another way to argue that there’s a problem!” often get comments like yours, and those comments often get upvoted. Beware confirmation bias when browsing those; the key point I’m making here is that there isn’t any sort of total consensus, not that you’re wrong that some people push for a sort of bland and useless “let’s raise awareness”.)
Interesting recent posts that are making the point you are or related ones:
If you ask me (which you didn’t): There’s real reason to be concerned about the trajectory of AI. There’s real reason to invite more people to help. And yet you’re quite right; just yelling “hey, help with this problem!!” is not a strategy that is a good idea to make reputable. Science is hard and requires evidence. Especially for extraordinary claims.
Also, I think plenty of evidence exists that it’s a larger than 5% risk. And the ai safety fieldbuilding tag does have posts that go over it. I’d suggest opening a bunch of them and reading them fast, closing the ones that don’t seem to make useful points; I’m sure you will, on net, mostly think most of the arguments suck. If you don’t think you can sift good from bad arguments and then still take the insights of the good arguments home, like, shrug, I guess don’t let you brain be hacked by reading those posts, but I think there are some good points in both directions floating around.
Upvote, disagree: Raising productively useful awareness looks like exactly the post you made. Insufficiently detailed awareness of a problem that just looks like “hey everyone, panic about a thing!” is useless, yeah. And folks who make posts here have warned about that kind of raising awareness before as well.
(Even as a lot of folks make posts to the ai safety fieldbuilding tag. Some of them are good raising awareness, a lot of them are annoying and unhelpful, if you ask me. The ones that are like “here’s yet another way to argue that there’s a problem!” often get comments like yours, and those comments often get upvoted. Beware confirmation bias when browsing those; the key point I’m making here is that there isn’t any sort of total consensus, not that you’re wrong that some people push for a sort of bland and useless “let’s raise awareness”.)
Interesting recent posts that are making the point you are or related ones:
first, they post https://www.lesswrong.com/posts/FqSQ7xsDAGfXzTND6/stop-posting-prompt-injections-on-twitter-and-calling-it but then they followed it up with https://www.lesswrong.com/posts/guGNszinGLfm58cuJ/on-second-thought-prompt-injections-are-probably-examples-of
https://www.lesswrong.com/posts/nExb2ndQF5MziGBhe/should-we-cry-wolf
link specifically to a previous comment of mine, but I really mean the whole post—https://www.lesswrong.com/posts/AvQR2CqxjaNFK7J22/how-seriously-should-we-take-the-hypothesis-that-lw-is-just?commentId=JmhiskFRhYJc3vyXk
If you ask me (which you didn’t): There’s real reason to be concerned about the trajectory of AI. There’s real reason to invite more people to help. And yet you’re quite right; just yelling “hey, help with this problem!!” is not a strategy that is a good idea to make reputable. Science is hard and requires evidence. Especially for extraordinary claims.
Also, I think plenty of evidence exists that it’s a larger than 5% risk. And the ai safety fieldbuilding tag does have posts that go over it. I’d suggest opening a bunch of them and reading them fast, closing the ones that don’t seem to make useful points; I’m sure you will, on net, mostly think most of the arguments suck. If you don’t think you can sift good from bad arguments and then still take the insights of the good arguments home, like, shrug, I guess don’t let you brain be hacked by reading those posts, but I think there are some good points in both directions floating around.