Curiosity: Why did you mega-downvote "AI is Software" ?

by AndyWood1 min read5th Jun 201443 comments


Personal Blog

I've never experienced anything like it before on LessWrong. Would you care to re-read the post, and offer your feedback?

38 comments, sorted by Highlighting new comments since Today at 12:12 PM
New Comment
  1. Your phrasing was pseudo-mystical.
  2. You used non-standard meanings for words.
  3. It was a bunch of assertions, without a meaningful way to springboard into more discussion.
  4. My first thought was "why doesn't this person recognize that something so out of the ordinary requires a lot more explication?"
  5. I had to think much more than should have been necessary to even begin to grasp what you could even possibly have meant by saying things like "Software is becoming the new System".

Basically, it came across as a very low-effort post that seemed explicitly designed to create poor quality discussion.

Ordinary for what? Ordinary for this small community? It's a big world out there.

It's a big world out there.

But you're talking to the world in here.

Well, you posted it to this small community, and then wondered why it was downvoted.

I downvoted your post because it contained virtually no meaningful discussion or ideas and it didn't explain or expound upon what it did contain. Frankly, I'd rather not have that kind of thing even appear in Discussion, because it clutters the space when looking for actual content.

Edit: there was apparently more content there earlier that was deleted. I downvoted based on the short version with the deleted content, not on the version with more information.

AndyWood lost all the karma he accumulated over 5 years, with three lines posted to Discussion. There was little discussion, lots of reactionary criticism.

Does this seem in balance to you, Dentin?

AndyWood lost all the karma he accumulated over 5 years, with three lines posted to Discussion. There was little discussion, lots of reactionary criticism.
Does this seem in balance to you, Dentin?

It wasn't 3 lines. The actual discussion post is at -39 right now. If Andy had just realized that he was in a hole and stopped digging, he could have walked away with most of his karma intact. Instead, he made comment after comment trying to defend his post, and the combined downvotes on all those comments is what wiped out 5 years of karma.

Andy needs to talk to Professor Quirrell and ask for a lesson on how to lose.

AndyWood lost all the karma he accumulated over 5 years

He lost all the karma he bothered to accumulate over five years. Karma is easy to come by if you want it.

I didn't down-vote, but was tempted to. The original post seemed content-free. It felt like an attempt to start a dispute about definitions and not a very interesting one.

It had an additional flaw, which is that it presented its idea in isolation, without any context on what the author was thinking, or what sort of response the author wanted. It didn't feel like it raised a question or answered a question, and so it doesn't really contribute to any discussion.

I did not mega-downvote, but I do feel it was a low-quality post.

The problem is that you made a big bold declaration whose meaning was pretty unclear (the post was very short too), anthropomorphized computer software (giving your post a mystical flavor) and seemed to be attacking the people interested in AI on this site without clear supporting reasons.

It was mega-downvoted because it was (a) very confused, and (b) talked about "chauvinism". Either one of these errors would've drawn a handful of downvotes and then been ignored, but the combination is much worse: it signaled an intent to spread Social Justice into a place where it doesn't belong, in a way that doesn't make sense, in a way that would've damaged both Social Justice itself and thinking about AI if it stuck, and we are on guard against that, and we should be.

It was made of not-even-wrong.

BTW, are you knd1? That stream of posts goes beyond not-even-wrong to not-even-voteable-on. I just don't know what I'm looking at there.

Would you care to re-read the post

(The current state of the post is very different from how it was when I originally downvoted it.)

Anyway. I downvoted "AI is Software" or whatever it was actually called at that point for the following reasons (all of which have already been mentioned by others here, I think):

  • Main content was a grandiose claim with no real support offered for it.
    • Another way of looking at it: the central claim was (approximately) a deepity: you could be taken either as saying "let's redefine AI to mean all software" (not an exciting claim, merely a proposal to redefine a term; seems like a pretty stupid redefinition given the huge amount of pre-existing use of the term) or as saying "actually, with a proper understanding, all software is AI even with something closely resembling the prevailing definition" (a grand claim indeed, at least on its face, but also rather obviously false and with no evidence offered to outweigh its implausibility)
  • What others have termed "pseudo-mystical" language with, again, no indication of why it's appropriate.
  • A general sense that the reasoning (in so far as there was any) was muddled, and in particular
    • Little expectation that engaging with the author would lead to enlightenment on either side.
  • Very preachy tone.

I have seen quite a lot of things written in that sort of way, and to date I don't think I have ever known any good to come from trying to engage their authors in a rational discussion.

I don't now remember for sure, but the following may also have been a reason and certainly would have been if I'd run across the article later:

  • Author's extremely poor response to criticism: little actual engagement, much name-calling and taking-offence, and at some point completely rewriting the post without any acknowledgement of having done so.

The comments to this post contain further examples of this.

Context: I would have downvoted it, if it hadn't already been so low. I only saw the shortened version.

The basic issue is that you were disputing definitions. You weren't showing that all software is software that does tasks that historically humans remained better at than computers (i.e. "AI"), you were just redefining the word AI to be a synonym for software. You but didn't seem to address or realize that fact.

Also, I may be remembering wrong, but the points seemed to be based on a personal sense of aesthetics. That sort of argument doesn't work well when others don't agree on what's elegant. Better to focus on why the existing definition is contrived, how it doesn't cut reality at its seams, how it leads to misunderstandings, etc.

Upvoted - I think the original post was not even wrong and the poster doesn't get a lot of the criticisms, but I think the honest question really does not deserve downvoting. He's trying to take on the criticisms and deserves credit for that.

Saying that people disagree with him because they are autists is a strange way to take on the criticism.

Well, some of the criticisms :-)

The primary issue? No matter how many times I read your post, I still don't know what your claim actually is.

I am actually pretty disappointed in the overall response. It's not that I thought the post was great, but the massive downvoting and heckling struck me as the pathological autoimmune response of a sick system.

You didn't exactly handle the criticism with grace and humility either, since you asked.

The post was low-quality.

It can be difficult to know when to "risk" posting. For my first post, I was so cautious that I ran it by Michael Vassar. But a simpler way is to get involved in comment threads including the open thread. You can mention your ideas there and see what feedback you get.

I didn't read that post, but I up voted this post because it seemed honestly curious and engaged in real inquiry.

I didn't downvote it, and was disappointed when I looked at it and it appear to have been downvoted simply for being wrong. My impression at the time was that the author was thinking sloppily, and was insisting on continuing the same pattern of thought despite corrective comments.

It appears to have been removed, so I can't go back and check my assumptions.

In my view, the only reason for downvoting something so far would be if it were completely off topic or the poster were aggressively insulting other users.

Thank you, Gavin

It just occurs to me now that the mass downvote of your post demonstrates a possible flaw with a major reason for downvoting things.

Many people will say something like "I downvoted it because I don't want to see posts like this on lesswrong." Lets call the people who believe something like this about downvotes class A.

If the majority of downvoting is class A, then once a post had been downvoted to -5 or maybe -10, you would expect it to gather no more downvotes. After all, class A people are using downvotes to signal things they don't want to see. At some number of downvotes, the signal is made with high reliability, and if class A people are rational they will not bother looking at the post knowing with high likelihood it is a post they don't want to see. And so there will be no additional class A downvotes past some significant number, perhaps 10 downvotes.

But in fact what we see is more than 200 downvotes. AndyWood is currently down 247 in the recent month, I checked his score perhaps 2 days ago and it was much higher. So despite an overwhelmingly clear signal that this is a post that class A people don't want to see, we continue to have people who will downvote this post opening it and downvoting not only this post but many of AndyWOod's comments in the comment thread of the post.

I don't think one can escape the conclusion that the reason for the massive downvoting is not the "class A" reason. It is not because this is a post people don't want to see. Or at least if it is, the suggestion that downvoting a post signals to people that this is a post they don't want to see and therefore that they will, if rational, choose not to look at, is incorrect.

Perhaps an objection to my conclusion could be raised, "as a responsible Class A citizen, when I see an undesirable post I run towards it to help my fellow class A citizens properly signal its nature." Well this would seem to demonstrate the futility of class A thinking: a heavily downvoted post becomes a post that people WANT to see so they can demonstrate good citizenship by helping signal it as a post people won't want to see.

I think you're confused. The bulk of his 200+ karma loss isn't from the main post; it's from his responses in various threads, where he takes criticism very poorly, is combative, or is non-responsive.

For the main post, there's another reason it's at such a high negative ranking: for the first day or two, the vote counter was not visible. That's when I added my downvote. It could have already been at -35 when I voted, but there's no way I could have known. I don't add a downvote to stuff that's less than -10 already.

For the main post, there's another reason it's at such a high negative ranking: for the first day or two, the vote counter was not visible. That's when I added my downvote. It could have already been at -35 when I voted, but there's no way I could have known. I don't add a downvote to stuff that's less than -10 already.

It was at -20 when I first saw it, with its karma showing clearly, and is now at -39. So perhaps you do not vote things down that are already at -10, but at least 19 people do.

In general I downvote things as I read them with no regard as to how other people have already voted.