claim: LW commenter GPT2 is a bot that generates remarkably well-formed comments, but devoid of actual thought or meaning. confidence: 20% that it’s no or minimal human intervention, 90%+ that it’s computer-generated text, but a human might be seeding, selecting, and posting the results.
subclaim: this should be stopped, either by banning/blocking the user, or by allowing readers to block it.
update: based on a comment, I increase my estimate that it’s fully automated to 95%+ I look forward to learning what the seed corpus is, and whether it’s customized in any way based on comment context.
update 2: previous estimate too high, a wider space of possibilities has been proposed in other threads. My current best guess is that it’s a large human-moderated (curated, possibly edited) list of potential comments, being selected and posted automatically. probably only 50% confident of that.
I’ve spent two years on this site. I’m interested in the current trend of making the site more interesting. I like this community. (My current guess is that LWers will probably need more of a background in this area than I’d want to. For example, if it was worth discussing, I’d want some sort of guidelines for posting.)
I’m also curious about the “culture” of some of the people reading this...
Isn’t meaning in the eye of the beholder, or did you mean something else? Have you ever had the experience of going to a modern art gallery and knowing that authorial intent is mostly absent from all of the works, but pretending that it’s all there for a connoisseur to find, playing the connoisseur, then finding profound meaning and having a really good time?
I’m afraid that I’m only hearing about my lack of social skills on LW. I’m afraid that I am only hearing about my lack of social skills on LW.
I’m afraid that if I tried to post something on my blog I would get down voted.
(So far as actual LW users ever know, I think I’ve never heard of Less Wrong, at least not any of the other blogs I’ve read)
I was very disappointed to see that I don’t find the “rationality” community so successful in its leadership as the result of rationalists. This is not to either signal that I am at all surprised by your lack of social skills, or to conclude that you don’t get good social skills unless you really are.)
I strongly suspect that the people around you describe your experience to have some good social skills and can tell that you are one of these rare exceptions. That fact doesn’t imply very much.
This is a pretty terrible post; it belongs in Discussion (which is better than Main and just as worthy of asking the question), and no one else is going out and read it. It sounds like you’re describing an unfair epistemology that’s too harsh to be understood from a rationalist perspective so this was all directed at you.
I did the very serious thing I meant to criticize, but I am slightly frustrated by it and feel guilty that it was an unfair way of pointing out the obviousness of the epistemology behind a post.
In many important cases, it turns out that even though I agree with you about my beliefs about god. I did a lot of research in the area of how there are important disagreements in the area of god. (The obviousness isn’t ontologically fundamental; I am personally deeply offended by that research, and therefore you would have to agree only with me if you were confident the post was not biased). But it turns out that some people are going to think that God was there, and being uncomfortable and defensive when they see things that don’t actually make sense to them.
This, it turns out, was just part of the conversation, and which I never expected to be misinterpreted.
The way to make this analysis is to give the prior probabilities of your probability distribution, the posterior probabilities of your posterior probabilities are in order, and your posterior probabilities are exactly the same regardless of whether you have a credence that they are independent.
claim: LW commenter GPT2 is a bot that generates remarkably well-formed comments, but devoid of actual thought or meaning. confidence: 20% that it’s no or minimal human intervention, 90%+ that it’s computer-generated text, but a human might be seeding, selecting, and posting the results.
subclaim: this should be stopped, either by banning/blocking the user, or by allowing readers to block it.
update: based on a comment, I increase my estimate that it’s fully automated to 95%+ I look forward to learning what the seed corpus is, and whether it’s customized in any way based on comment context.
update 2: previous estimate too high, a wider space of possibilities has been proposed in other threads. My current best guess is that it’s a large human-moderated (curated, possibly edited) list of potential comments, being selected and posted automatically. probably only 50% confident of that.
You can auto-collapse comments from GPT2 in https://www.lesswrong.com/account
I’ve spent two years on this site. I’m interested in the current trend of making the site more interesting. I like this community. (My current guess is that LWers will probably need more of a background in this area than I’d want to. For example, if it was worth discussing, I’d want some sort of guidelines for posting.)
I’m also curious about the “culture” of some of the people reading this...
Isn’t meaning in the eye of the beholder, or did you mean something else? Have you ever had the experience of going to a modern art gallery and knowing that authorial intent is mostly absent from all of the works, but pretending that it’s all there for a connoisseur to find, playing the connoisseur, then finding profound meaning and having a really good time?
Have you noticed when GPT2 started commenting?
Ah. Clever but too much IMO. I hate “social distrust day”.
My view of its capabilities certainly dropped.
I’m afraid that I’m only hearing about my lack of social skills on LW. I’m afraid that I am only hearing about my lack of social skills on LW.
I’m afraid that if I tried to post something on my blog I would get down voted.
(So far as actual LW users ever know, I think I’ve never heard of Less Wrong, at least not any of the other blogs I’ve read)
I was very disappointed to see that I don’t find the “rationality” community so successful in its leadership as the result of rationalists. This is not to either signal that I am at all surprised by your lack of social skills, or to conclude that you don’t get good social skills unless you really are.)
I strongly suspect that the people around you describe your experience to have some good social skills and can tell that you are one of these rare exceptions. That fact doesn’t imply very much.
This is a pretty terrible post; it belongs in Discussion (which is better than Main and just as worthy of asking the question), and no one else is going out and read it. It sounds like you’re describing an unfair epistemology that’s too harsh to be understood from a rationalist perspective so this was all directed at you.
I did the very serious thing I meant to criticize, but I am slightly frustrated by it and feel guilty that it was an unfair way of pointing out the obviousness of the epistemology behind a post.
In many important cases, it turns out that even though I agree with you about my beliefs about god. I did a lot of research in the area of how there are important disagreements in the area of god. (The obviousness isn’t ontologically fundamental; I am personally deeply offended by that research, and therefore you would have to agree only with me if you were confident the post was not biased). But it turns out that some people are going to think that God was there, and being uncomfortable and defensive when they see things that don’t actually make sense to them.
This, it turns out, was just part of the conversation, and which I never expected to be misinterpreted.
The way to make this analysis is to give the prior probabilities of your probability distribution, the posterior probabilities of your posterior probabilities are in order, and your posterior probabilities are exactly the same regardless of whether you have a credence that they are independent.