= 27d567bc2d48d9f40e9ccc20ea370b15)
Eliezer_Yudkowsky comments on Debunking komponisto on Amanda Knox (long) - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (116)
So, Rolf, do you understand how this post failed? Hint: Others are not to blame.
As always, feel free to share your opinion on the matter.
In the title, you named an opponent. You lost most of us right there, because debating against a person and searching for truth are incompatible mindsets. Since you tried to turn it into a status competition, we can't treat anything you say on the subject as trustworthy; you're too likely to deceive yourself and pass misconceptions on to us.
Seconded. If this post had merely failed to add a significant new angle to the debate, I would have simply ignored it. I down-voted it because it comes off as a sort of political attack against another member, and I would really rather not see that flavor of discourse here.
And then responding to everyone by saying "but I want to talk about C1 and you're not talking about C1" gave it the appearance of downright trolling. If I didn't have previous familiarity with you, that's what I'd have assumed, actually.
You might say, "But they didn't directly address C1!"
But no one cared, at that point, what you wanted them to address. You didn't have enough credit built up with them to steer the conversation - and objecting that it wasn't going where you wanted it to go just pissed them off further. If you'd engaged with them on their points, you could have built up credibility. As it is, the post was just dropped into the void.
Example problem: You stepped into a giant past discussion and didn't refer to it. If, for each point, you had either pointed to and refuted previous comments about those points, or else said, "And I read through the comments and found no reference to this point", you would have been picking up the conversation where it left off. As it is, the reaction is more like, "Oh, same points being rehashed again and ignoring the previous conversation we had about it." This reaction was sufficiently severe that no one bothered to talk about your points - so far as they were concerned, it had probably been already discussed and refuted in the past conversation, since you didn't bother to refer back to it.
My own impression was that you hadn't read the ~700 comments on the previous two posts.
It seems to me that komponisto's The Amanda Knox Test similarly made a case for "not guilty" without reference to the all of the arguments for "guilty", and had other substantial flaws. But it's at 29, compared to Rolf's -10.
As someone who has no interest in the case (I haven't made any comment related to the case, nor read any of the discussions), judging from just the posts, I think Rolf's main mistake is that he forgot that one needs to be much more careful when arguing against a majority belief.
That is not similar to stepping into a giant past discussion without referring to it. Kompo started a giant discussion. He would have been downvoted mercilessly had there been a preexisting bloated corpse of a thread on the topic.
With respect, I suggest reading the discussions, not just the posts, before making any negative judgment of komponisto.
I don't want to read that much text about a topic I have little interest in. Can you explain how my judgment would be changed if I did read the discussions? Also, I think a post should stand on its own content. If one needs to have read hundreds of comments before it can be fully understood/appreciated, then at least some kind of disclaimer to that effect should be attached.
We don't even expect people to follow the posts here without having read hundreds of other posts; "stand on its own" seems a little strong.
Ok, that was too strong, but I think I still have a reasonable point here.
Once you've read the "hundreds of other posts", you can follow most subsequent posts, and even then we encourage people to link unfamiliar concepts to the wiki or past posts. But in this case apparently you need to have read hundreds of comments just to follow one particular post.
This appears to be the common denominator regarding people who think that Rolf has brought up game-changing information in his post.
I don't think that he did. In fact, I can infer, without having read the discussions, that the information in Rolf's posts are likely contained in the discussions. So the only thing I think he did was to put them in a top-level post without making a sufficiently convincing case for his position.
Exactly. Therefore, it was voted down. Is this not working as intended?
Wei, I hope you will excuse me for being a little blunt here, but it's pretty clear that both you and Anna Salamon have assigned far too much weight to the fact that Rolf Nelson is a big-shot SIAI donor who says the right words and hangs out with the right people, while (to you) I'm just some guy on LW with an obscure pseudonym. I mean, seriously, if two major posts on LW by a 1000+ karma user that attracted ~700 comments did less to get you interested in the topic than Rolf's negative-scoring rant...well, that's something you might want to think about.
If by "topic" you mean whether Amanda Knox is a murderer, I still have zero interest in that (and actually I find it puzzling why so many people do have a strong interest in it). How did you infer that Rolf's post got me interested in that, or are you referring to something else?
Also, which part of the sentence you quoted do you disagree with? That "one needs to be much more careful when arguing against a majority belief" or that this is Rolf's main mistake?
You became interested enough to make comments like the above, comparing Rolf's post to mine and concluding that mine wasn't much better. I hope you'll understand why, given your status here, this felt like a huge insult to me.
If you think my post has "substantial flaws", I would like to know what those are. More to the point, I would have liked to have known what they were back in December, when the discussion was taking place. Maybe I could have addressed those flaws, or even -- if they truly were substantial -- revised the post. Evidently, however, you didn't think they were important enough to be worth mentioning at the time. Now, however, that Rolf Nelson's response to my post has been voted down to the lower depths of karma hell, now you suddenly seem to care (more) about the quality of my post.
Do you see where I'm coming from here? It seems to me that you have a high opinion of Rolf and not much opinion one way or the other of me, and so you just found it superficially implausible that such a karma differential could possibly be warranted. So you glanced at the posts, noticed (I guess) the fact that I didn't go into the details of the prosecution case, and said, "What's the big deal?" without actually thinking too much about the case itself. Well, the numbers should tell you something. Groupthink may account for some karma anomalies, but it's unlikely to produce a 40-point difference all on its own -- not here, anyway.
As for why people would care about this topic, well I really don't think it's all that hard to understand. For me, personally, the fact that Knox and Sollecito are, you know, sentient creatures (and actually, from what I can tell, pretty swell human beings all around) has a lot to do with it. But even putting that aside, it's not like there was any dearth of rationality issues and meta-issues at play here. This is an empirical question about reality that happens to be a rationality no-brainer and yet is the subject of intense public controversy (and not of a standard "political" type, either). I have found it to be a beautiful way of testing people's understanding of how the world works outside of their cached comfort zones. The cognitive biases involved are numerous; issues of calibration and confidence were also raised. In short, one needn't have been particularly interested in the fate of Amanda Knox and Raffaele Sollecito to have been interested in our discussions here -- though I happen to be so interested, and I don't think I need to apologize for that.
I interpreted that as, "The only thing wrong with Rolf's post was the tone." Maybe that's uncharitable, but what did you mean by "careful"?
Yes, that is the main flaw that I noticed. I think you can't make a case for .99 probability of innocence without going through the prosecution case in detail, unless you assume that the reader is already familiar with the prosecution case and knows the counterarguments for all of the prosecution's evidence and arguments, and you didn't make that assumption clear. Someone reading it without knowing much about the case (like me) would have just thought it unconvincing.
The reason I didn't point it out at the time was that I just didn't think about it. I read the post, wasn't very interested, and moved on.
Well, I was curious why Rolf's post was voted down so much, since I knew him from before. That seems reasonable to me.
I meant he needed to spend more effort to look for and eliminate the flaws in his post. I think when you write a post that contradicts some people's beliefs, those who are being contradicted will tend to search the post more carefully for flaws. So he should have known that his post would be scrutinized more and acted accordingly.
The problem with using it as a rationality learning tool is that the amount of time I'd have to put in to learn the details of the case, compared to what I might get out of it, just seem too high for me to make the effort. Your post is already quite long, and to make it more convincing would have required a lot more text.
You feel strongly for Knox and Sollecito, which is fine, but it's hard for me to understand why, given that there are billions of people on this planet, and a significant fraction of them suffer worse fates than they. Why these two, and not all the others? (I assume you don't feel equally strongly about all the others, but I guess I could be wrong. Let me know if I am.)
The post was prominently labeled as a followup to its predecessor; and as the latter consisted of instructions for a survey and thus amounted to little more than a prompt for discussion, it should have been obvious (or so I would have thought) that its substantive content lay in the comments.
In other words, I thought it was perfectly clear from context that if you weren't interested enough in the exercise to read at least the comments on the first post, if not the actual links provided, then you simply weren't in the target audience for the second post.
What sort of additional wording do you think would have helped make this clearer?
You may have missed a key point here; see comments by mattnewport and Eliezer. This is not, it turns out, a difficult case that hinges on the details; instead it's decided overwhelmingly just by the priors. The thesis of my post was that a skilled rationalist shouldn't need more than a few minutes of study in order to arrive at a high probability of innocence. Now, whether you could actually get to a probability of 0.99 or 0.999 that way, as opposed to merely 0.9, is an interesting, but separate, question.
I'm not sure why you would make such an assumption. It's not as if my total level of global outrage is fixed, and hearing about Knox and Sollecito unfairly moved some sympathy in their direction at the expense of others. It just so happened that this case came to my attention, and provoked an emotional response appropriate to my state of knowledge. The result was a net increase in my concern about the madness of this world and in my desire to alleviate suffering; you may be interested to know that additional resources were allocated to the general goal of improving humanity's future as a direct result of my feelings about this case.
Perhaps something along the lines of "This post assumes that the reader has read the comments section of the previous post."
The above might be true if I didn't know that Knox and Sollecito were convicted by an Italian court, but once I take that into account, it seems impossible to get to 0.9 probability of innocence that quickly. Unless I've seen nearly all of the evidence and arguments that the court has seen, I think there's a probability higher than 0.1 that the court knows something significant that I don't.
I assume that a human being has a finite capacity for outrage. To feel that strongly about Knox and Sollecito implies that you cannot feel equally strongly about all other individuals who have suffered equally terrible fates. You say that the case increased your net concern about the madness of the world, but surely you must have already known that injustice like this (assuming you're right Knox and Sollecito) happens every day to many people all over the world? Is it just that the abstract knowledge didn't engage your emotions, like seeing these two specific individuals did?
You have no evidence of Wei's thoughts on the matter.
Downvoted. The comment I replied to contains Wei's thoughts on the matter.
This response doesn't really suggest that you're taking your own responsibilities seriously in this matter.
The specifics here are interesting, but there's a more general point. If your interaction with a community isn't productive, there is simply no point in bemoaning the failings of the community and then carrying on as before. No matter what failings the community might or might not have, the only actions you directly decide are your own, and the only options that make sense are to give up and leave, or to figure out what action you can take to bring about more productive interaction.
You might think that the focus should be on what we should do to improve the interaction. If that's so, your thoughts should still lead you to consider what you can do to bring about that change in us; if you think the problem is a systematic failure in us, there are few places where a genuine effort to persuade us of that would be more welcome.
Whatever you do, don't bang your head against the wall. Either give up on us, or think about what you can do to maximize the productiveness of your interaction with us.
I'll admit I'm a bit confused as well. I red komponisto's previous posts and agreed with him. rolf_nelson's posts seem to raise important points about the issue. I honestly don't see what justifies the severe downmodding he's getting. Could someone fill me in?
Rolf raises a number of points that were already widely discussed (and widely dismissed) in the comments on komponisto's posts. He has not added any new information to the discussion that was not already contained in the voluminous commentary surrounding those two posts.
More importantly, he seems to have missed/ignored a fairly central point from komponisto's posts as relates to rationality (which is what we are generally here to discuss) which is that all of the physical and circumstantial evidence against Knox and Sollecito is irrelevant within a wide range of plausible weights assigned to it in light of the prior probabilities surrounding a case of this type. The reason this case is interesting as a test of rationality is that you can reach a likely-correct conclusion without examining any of the evidence discussed here in detail. You would need orders-of-magnitude more compelling evidence to outweigh the very small priors that should have been assigned to the joint guilt of Knox and Sollecito given the compelling evidence against Guede.
Click.
That much I understand. And I did read a lot of the previous discussions. But just to take a random example, rolf_nelson brought up how DNA of random people couldn't have easily gotten onto the bra clasp, that there had to be some more direct contact. And, RS's DNA is clearly on it.
Where was that specific point already addressed in previous discussion?
Here for one.
This CNN story covers some of the doubts around the handling of the bra and the validity of the DNA results:
Here's a source for the 'three unidentified individuals' DNA' claim:
Emphasis mine. There is sufficient doubt around the DNA evidence to render it largely irrelevant in light of the prior probabilities for this kind of crime.
Okay, thanks -- I wasn't sure if the rolf_nelson's claim about the three other people was the consensus at this point. I assumed he said it because he had found a good refutation of that point [1]. Confusion resolved. Disappointment with rolf_nelson's posts on this matter understood.
[1]Note: This is not a common practice of mine, but due to the factors at play for this specific discussion.
ETA: Changed phrasing to be less personal.
Can we say "disappointment with rolf_nelson's post"? I feel more comfortable with that.
I wasn't endorsing the general dislike of rolf_nelson, just saying why I understand the hostility now. And the issue is about more than just his top level post, but the broader exchange between him and komponisto.
(And why do you insist on the underscore?)
ETA: Come on, it's a joke, people. I criticized wedrifid for simply following my usage (plus making it the possessive form), even though I use that exact form all the time. I was jokingly making it look like it's wedrifid who insists on the practice of adding underscores. Get it? Ha ha? No? Okay then...
Who are you and what have you done with Silas??
LOL! :-)
ETA: I notice that many people have since picked up the practice of referring to other users with their exact username, and replacing spaces therein with underscores. I'm not going to say it was because of me ... but I had been doing that since way back into the OB days, when no one else was ...
I'm confused. I just added 's post in this instance. More generally I tend to refer to a user by either their username or their first name. Maybe a last name instead if the reference was towards a more formal academic contribution.
whoosh! ;-)
Updated phrasing to a happy medium.