Talk:LessWrong/Archive9

From RationalWiki
Jump to navigation Jump to search

This is an archive page, last updated 3 May 2016. Please do not make edits to this page.
Archives for this talk page: , (new)(back)

Alexander Kruel's interviews with AI experts[edit]

http://kruel.co/2012/08/15/qa-with-experts-on-risks-from-ai-5/

Doesn't really seem to share much of the LW/SIAI take on the matter, particularly in the case of Laird.--Baloney Detection (talk) 20:44, 17 August 2012 (UTC)

Ohh, but you see, that's why you should donate! The experts are so reckless and are going to kill us all! Dmytry (talk) 11:27, 18 August 2012 (UTC)

Some thoughts on the 'bayesian' rationality.[edit]

I've put some of my thoughts there: Essay:LessWrong_Bayes_and_Rationality (i hope it is not against the rules). Basically, when you have incomplete inference graph with cycles, doing some theoretically optimal shit that works only on trees is a pretty damn stupid idea, along the lines of believing that you can walk off 8th floor and just keep on going because you just learnt of the awesome concept of inertia and ignored such low and specific details of Earth environment as gravity. — Unsigned, by: Dmytry / talk / contribs 18:05, 6 June 2012 (UTC)

An utterly predictable car crash of a thread[edit]

At least they are unlikely to reproduce. - David Gerard (talk) 10:58, 11 September 2012 (UTC)

Did anything come out of that? I mean how many of the people talking actually go to meetings? --83.84.137.22 (talk) 12:07, 11 September 2012 (UTC)
I can see where this is going... do I need to read on? Scarlet A.pngtheistModerator 12:10, 11 September 2012 (UTC)
I'm not sure about "car crash", actually. Skimmed it, mulled over a few points, and it's a bit of a mixed bag. The little bit about discussing "privilege" from the perspective of heuristics and biases is good, but, of course, I'm biased because I think "privilege" is a heuristics and biases issue and should be treated as such to avoid it being a catch-all term for dismissing certain groups of people. Scarlet A.pngbominationModerator 12:49, 11 September 2012 (UTC)
If you're suggesting people should try to act less creepy, you're just oppressing the socially unskilled, and besides you haven't defined "creepy" yet, and besides it's all just a status game so I can ignore it - David Gerard (talk) 13:27, 11 September 2012 (UTC)
Have you ever actually met anyone with no social skills? I mean, actually met them and not been just repulsed and dismissed them? It's considerably more complicated than just asking them to "try and act less creepy". In that respect,at least, they're onto something. Scarlet A.pngnarchistModerator 14:34, 11 September 2012 (UTC)

Consider that 'creepy' is a useful detector of 'dangerous'. When you see people take seriously their amateur explorations into nature of morality, who believe in a an almost certain doomsday, and are stupid enough to buy into Roko's basilisk, as well as highly prone to taking their own bullshit with utter certainty (e.g. Doomsday Paradox and other anthropic reasoning for 'foolproof' estimate of likehood of AI killing us all)... you can laugh all day how they are so nerdy they would never actually do anything but you know this is just comforting BS and it is common for e.g. shooters to be nerdy and creepy. I wouldn't be particularly surprised if, when the AIs gets a little less pathetic, some people from LW will shoot up an AI conference. I also wouldn't be surprised if someone from "individuals tending to savagery" participates on LW. In this case, creepy is an entirely valid summary of multitude of mental health criteria that dramatically increase risk of dangerous behaviours. If someone fails to intuitively learn social behaviours from interaction with people, who knows what else did they fail to learn? edit: also, I consider the LW discussions described here as a great example of 'creepy'.Dmytry (talk) 07:28, 13 September 2012 (UTC)

What's kind of interesting is that this is a chance for them to look at a social issue from their perspective of eliminating bias. A significant portion of what privilege is, is bias to the point of not realizing that a perspective is not reality for everybody. So to me, this is a great chance for plenty of LWians to say 'aw shit, I found another one, time to try and correct myself to eliminate this' because it reaches out to them in their own area of interest rather than trying to tell them about the topic and then if they don't get it, "TAKE A SOCIOLOGY COURSE THEN YOU'LL GET IT." To which many of the enginners there probably would say "OVER MY DEAD BODY." ±Knightoftldrsig.pngKnightOfTL;DRlavishly loquacious 12:32, 13 September 2012 (UTC)

How the new LW article style works[edit]

Scientists discover the truth behind Colbert’s “truthiness” . It is rather funny actually, how the site that is supposed to obsess over biases uses every trick and bias to help people believe the technoreligion. Dmytry (talk) 13:47, 15 September 2012 (UTC)

Yeah I've been fascinated by how people talking about biases as much as they do seem incapable of discovering it in themselves. Bottom line is that it is always a bad idea to put one single person (Yudkowsky in this case) on a piedestal. Nobody is infallible.--Baloney Detection (talk) 16:53, 25 September 2012 (UTC)

Luke Muehlhauser facts[edit]

From Yvain and Alicorn. - David Gerard (talk) 13:07, 25 September 2012 (UTC)