This wouldn’t be that surprising, because Nonlinear are known to have violated various kinds of mild norms previously—they named an eponymous prize without consent from the named person or their estate, and they set up a podcast using text-to-speech without obtaining copyright permissions from the authors of the text.
Those accusations seem of a dramatically more minor and unrelated nature and don’t update me much at all that allegations of mistreatment of employees are more likely.
I largely agree with Ruby here, but wanted to note one comment, where one justification for “violating ” (this word seems too strong) this norm was “a descendant of Truman would have to actually learn of this prize”. If the research eventually done happened prior to the announcement, I think there would not be any meaningful update for me. OTOH, if this justification was a reason to not have done this research, and if it was applied more generally and not just for the naming of the prize, it would make me more suspicious that the allegations leveled against them are plausible, and it fits the “ends justify the means”-type reasoning that the OP refers to.
More minorly, they also manipulate things like voting scores on the EA forum (and pay a lot of attention to online presence).
More substantively but with greater uncertainly:
Back in 2021, there were a bunch of potential EA meta orgs, that are essentially natural monopolies, there can only be one. As another feature, these meta EA orgs naturally have to attract/involve a lot of EA resources and talent. The nature of these potential orgs is that they have less legible or viewable output (than say, distributing mosquito nets or producing research).
1+ grantmakers at a major EA grantmaking organization said it’s known people didn’t have ideal views on them (but suggested I meet with Kat because we had to, they were in the space).
After the meeting, I got a bit of an elbowing/redirection vibe, which in hindsight was extremely not justifiable by their output.
What is the evidence for manipulating voting scores? Feel free to say you can’t share, but if it can be publicly evidenced in some way, as it seems it might, that would be more informative.
I interpret the situation as an incident that Kat did not honor her word about content, Kirsten was upset, and Kat suppressed Kirsten’s concerns. As you can see, Charles He (who seems otherwise unintelligent) pointed this out clearly.
I grudge providing this, and I do so because of interest which I think is disproportionate, as this is not substantive compared to the rest of the claims. Frankly, I assumed this voting behavior was common knowledge.
As a moderator, I think the phrase “seems otherwise unintelligent” is clearly not generous or collaborative and breaks Forum norms. This is a warning, please don’t insult other users.
There’s probably a rationale I don’t understand here—but what’s the reason for allowing someone to strong upvote their own comment, irrespective of whether the thread was about a misaction? (though this makes it worse). Like do they get karma for doing this and are just doing this to boost their own karma? [Edit: deleted, as this does not appear to be the case, at least not when I tried it just now] Are we suggesting that people with higher karma both make better comments and can be trusted to judge themselves independently? Are we not worried this could bias other peoples’ judgements?
″...The researchers automatically and artificially gave certain comments on stories an immediate up vote—the first vote that a comment would receive. You might well think that after hundreds or thousands of visitors and ratings, a single initial vote on a comment could not possibly matter. That is a sensible thought, but it is wrong. After seeing an initial up vote, the next viewer became 32% more likely to give an up vote. Remarkably, this effect persisted over time. After five months, a single positive initial vote artificially increased the mean rating of comments by 25%.”
If something seems sketchy but doesn’t clearly break any rules, you should still reach out. The situation might lead to a rule change, or at least alert moderators to pay more attention to threads involving specific users.
FWIW, I don’t love nonspecific allegations like this — it’s nigh-impossible for Nonlinear to respond to them, and serves as a cheap or free way to make others think worse of them at no cost to yourself. By contrast, the initial comment at least gave enough information that Nonlinear could respond to some of the claims.
Is a claim like this something the moderators would have the ability to actually verify? If so, would a public claim like this be something the moderators would be interested in verifying? I similarly dislike the allegation as written and I think if this was verified in some way it would be useful regardless of the outcome
When we (the moderation team) have a good reason to believe that norm-breaking behavior is going on, we can access information including which accounts voted, and how, on any given post or comment (we’d need to work with the developers on this; this information isn’t readily available to us). To protect users’ privacy, we check whether something is weird before we look at IDs by looking at things like the timing and weight of votes without looking at who voted.
But we don’t investigate without good reason, both because we don’t have capacity for it, and because we want to give users privacy. So we’d need a reason to investigate, which is where reports like this come in.
Sample good reasons:
“A bunch of my comments from past months were all downvoted in the last day or two, can you see if someone was mass-downvoting?”
“I heard USERNAME say they paid people to upvote them on MTurk, can you see if a bunch of new users just joined to vote on their posts?”
Sample bad reason:
“I think USERNAME is manipulating karma, can you investigate them?”
(You can also make the claims privately by directly getting in touch, which can be better than speculating publicly.)
I am satisfied with the truth, accuracy and confidence of my statements in this post. Based on some of the responses from Non-linear, I would probably change my content to increase my concerns somewhat.
I’m not immediately able to say anything useful about the meta question of anonymous comments—this seems extremely complicated and I probably might not agree if I think about it.
I might be anonymous (?) right now. I have unique reasons that I am anonymous, and I might not be for long.
I am interested if you have contravening information about my representations on my account, or any other concerns. Please write them publicly.
I know nothing about most of the discussion here, but...
they named an eponymous prize without consent from the named person or their estate
That wasn’t them, it was me. Since I came up with the prize, and suggested the name. Ryan reasonably reached out them to get clarity, and I wasn’t initially in the loop, but that’s a communication problem, not an ethics issue. And if you look at the comments, I really don’t think it was mishandled.
This wouldn’t be that surprising, because Nonlinear are known to have violated various kinds of mild norms previously—they named an eponymous prize without consent from the named person or their estate, and they set up a podcast using text-to-speech without obtaining copyright permissions from the authors of the text.
Those accusations seem of a dramatically more minor and unrelated nature and don’t update me much at all that allegations of mistreatment of employees are more likely.
Also, the naming was completely on me, not them, as I explained in another comment.
I largely agree with Ruby here, but wanted to note one comment, where one justification for “violating ” (this word seems too strong) this norm was “a descendant of Truman would have to actually learn of this prize”. If the research eventually done happened prior to the announcement, I think there would not be any meaningful update for me. OTOH, if this justification was a reason to not have done this research, and if it was applied more generally and not just for the naming of the prize, it would make me more suspicious that the allegations leveled against them are plausible, and it fits the “ends justify the means”-type reasoning that the OP refers to.
More minorly, they also manipulate things like voting scores on the EA forum (and pay a lot of attention to online presence).
More substantively but with greater uncertainly:
Back in 2021, there were a bunch of potential EA meta orgs, that are essentially natural monopolies, there can only be one. As another feature, these meta EA orgs naturally have to attract/involve a lot of EA resources and talent. The nature of these potential orgs is that they have less legible or viewable output (than say, distributing mosquito nets or producing research).
1+ grantmakers at a major EA grantmaking organization said it’s known people didn’t have ideal views on them (but suggested I meet with Kat because we had to, they were in the space).
After the meeting, I got a bit of an elbowing/redirection vibe, which in hindsight was extremely not justifiable by their output.
What is the evidence for manipulating voting scores? Feel free to say you can’t share, but if it can be publicly evidenced in some way, as it seems it might, that would be more informative.
Here is one instance of what I consider vote manipulation.
https://forum.effectivealtruism.org/posts/cTQfWpobqk4nDWsfG/new-use-the-nonlinear-library-to-listen-to-the-top-ea-forum#comments
I interpret the situation as an incident that Kat did not honor her word about content, Kirsten was upset, and Kat suppressed Kirsten’s concerns. As you can see, Charles He (who seems otherwise unintelligent) pointed this out clearly.
I grudge providing this, and I do so because of interest which I think is disproportionate, as this is not substantive compared to the rest of the claims. Frankly, I assumed this voting behavior was common knowledge.
As a moderator, I think the phrase “seems otherwise unintelligent” is clearly not generous or collaborative and breaks Forum norms. This is a warning, please don’t insult other users.
There’s probably a rationale I don’t understand here—but what’s the reason for allowing someone to strong upvote their own comment, irrespective of whether the thread was about a misaction? (though this makes it worse).
Like do they get karma for doing this and are just doing this to boost their own karma? [Edit: deleted, as this does not appear to be the case, at least not when I tried it just now]Are we suggesting that people with higher karma both make better comments and can be trusted to judge themselves independently? Are we not worried this could bias other peoples’ judgements?
From Noise:
Source: https://www.science.org/doi/10.1126/science.1240466
Would someone be happy to explain what forum norm I’ve broken with my above comment, or the reason for the strong downvote, given the voting norms?
If you think someone is breaking the Forum’s rules, I encourage you to contact the moderation team.
If something seems sketchy but doesn’t clearly break any rules, you should still reach out. The situation might lead to a rule change, or at least alert moderators to pay more attention to threads involving specific users.
FWIW, I don’t love nonspecific allegations like this — it’s nigh-impossible for Nonlinear to respond to them, and serves as a cheap or free way to make others think worse of them at no cost to yourself. By contrast, the initial comment at least gave enough information that Nonlinear could respond to some of the claims.
Is a claim like this something the moderators would have the ability to actually verify? If so, would a public claim like this be something the moderators would be interested in verifying? I similarly dislike the allegation as written and I think if this was verified in some way it would be useful regardless of the outcome
Depends on the claim.
When we (the moderation team) have a good reason to believe that norm-breaking behavior is going on, we can access information including which accounts voted, and how, on any given post or comment (we’d need to work with the developers on this; this information isn’t readily available to us). To protect users’ privacy, we check whether something is weird before we look at IDs by looking at things like the timing and weight of votes without looking at who voted.
But we don’t investigate without good reason, both because we don’t have capacity for it, and because we want to give users privacy. So we’d need a reason to investigate, which is where reports like this come in.
Sample good reasons:
“A bunch of my comments from past months were all downvoted in the last day or two, can you see if someone was mass-downvoting?”
“I heard USERNAME say they paid people to upvote them on MTurk, can you see if a bunch of new users just joined to vote on their posts?”
Sample bad reason:
“I think USERNAME is manipulating karma, can you investigate them?”
(You can also make the claims privately by directly getting in touch, which can be better than speculating publicly.)
I am satisfied with the truth, accuracy and confidence of my statements in this post. Based on some of the responses from Non-linear, I would probably change my content to increase my concerns somewhat.
I’m not immediately able to say anything useful about the meta question of anonymous comments—this seems extremely complicated and I probably might not agree if I think about it.
I might be anonymous (?) right now. I have unique reasons that I am anonymous, and I might not be for long.
I am interested if you have contravening information about my representations on my account, or any other concerns. Please write them publicly.
I know nothing about most of the discussion here, but...
That wasn’t them, it was me. Since I came up with the prize, and suggested the name. Ryan reasonably reached out them to get clarity, and I wasn’t initially in the loop, but that’s a communication problem, not an ethics issue. And if you look at the comments, I really don’t think it was mishandled.