Is keeping AI "in the box" during training enough? 2021-07-06T15:17:45.014Z
Wisdom of the Crowd: not always so wise 2012-07-01T20:55:36.230Z
Where Fermi Fails: What is hard to estimate? 2012-06-05T03:15:32.232Z
[POLL] Wisdom of the Crowd experiment 2012-01-05T22:13:31.511Z


Comment by tgb on Is keeping AI "in the box" during training enough? · 2021-07-07T16:43:09.853Z · LW · GW

The key question is your second paragraph, which I still don't really buy. Taking an action like "attempt to break out of the box" is penalized if it is done during training (and doesn't work), so the very optimization process will be to find systems that do not do that. It might know that it could, but why would it? Doing so in no way helps it, in the same way that outputting anything unrelated to the prompt would be selected against.

Comment by tgb on Open thread, August 14 - August 20, 2017 · 2017-08-20T16:10:30.404Z · LW · GW

The full answer is: they cannot return if there are only finitely many balls, but they can if there are infinitely many.

Let's first assume that there are finitely many balls. As Thomas pointed out, we can assume that the center of mass is fixed. Let's consider R defined to be the distance from the center of mass to the furthest ball and call that furthest ball B (which ball that is might change over time). R might be decreasing at the start - we might start with B going towards the center of mass. But if R decreased forever then we would know that they never return to their starting location (since R would be different)! So at some point it must become at least as large as it was at the start. At that point either the derivative of R is 0 or it is positive. In either case, R must increase forever onwards - which again shows it can't return to its original starting point. Why is it always increasing from that point onwards? Well, the only way for the ball B to turn around and start heading back towards the center is if there is another ball further away than it to collide with it. But that can't be, since B is the furthest out ball! (Edit: I see now that this is essentially equivalent to cousin_it's argument.)

For infinitely many balls, you can construct a situation where they return to their original position! We're going to put a bunch of balls on a line (you don't even need the whole plane). In the interval [0,1], there'll be two balls with initial velocity heading in towards each other at unit speed, with one ball at the left edge of the interval and one ball at the right. Then do the same thing for each interval [k,k+1]. When you let them go, each pair in each interval will collide and then be heading outwards with unit interval. Then they'll collide at the boundary with the next interval with the ball from the next interval. That sets them back at the starting position. I.e. all balls collide first with their neighbor on one side, then their neighbor on the other side, setting them back to their starting position.

Comment by tgb on Why election models didn't predict Trump's victory — A primer on how polls and election models work · 2017-02-03T00:10:48.506Z · LW · GW

No problem!

Comment by tgb on Why election models didn't predict Trump's victory — A primer on how polls and election models work · 2017-01-28T23:35:33.742Z · LW · GW

A cursory glance through Fivethirtyeight's collected poll data shows a survey with over 84,000 voters (CCES/YouGov) giving Clinton a +4 percentage point lead, with 538 adjusting that to +2. Google and SurveyMonkey routinely had surveys of 20,000+ individuals, with one SurveyMonkey one having 70,000 with Clinton +5 (+4 adjusted). There was no clear reason to prefer your poll (whichever that one was) over these.

And it should go without saying that Clinton did end up at +2 nationally.

Comment by tgb on Triple or nothing paradox · 2017-01-07T15:40:00.494Z · LW · GW

I don't like any of the proposed solutions to that when I glanced through the SEP article on it. They're all insightful but are sidestepping the hypothetical. Here's my take:

Compute the expected utility not of a choice BET/NO_BET but of a decision rule that tells you whether to bet. In this case, the OP proposed the rule "Always BET" which has expected utility of 0 and is bested by the rule "BET only once" which is in turn bested by the rule "BET twice if possible" and so on. The 'paradox' then is that there is a sequence of rules whose expected earnings are diverging to infinity. But then this is similar to the puzzle "Name a number; you get that much wealth." Which number do you name?

(Actually I think the proposed rule is not "Always BET" but "Always make the choice for which maximizes expected utility conditional to choosing NO_BET on the next choice". The fact that this strategy is flawed seems reasonable: you're computing the expectation assuming you choose NO_BET next but don't actually choose NO_BET next. Don't count your eggs before they hatch.)

Comment by tgb on Open thread, Oct. 24 - Oct. 30, 2016 · 2016-10-29T20:18:23.042Z · LW · GW

Vaporising a comet takes significant energy. Heating up a comet to vaporization point takes significant energy. Dissipating the vaporized comet (still the same total mass and momentum as when it was in a solid state) takes significant energy. I really find this simplistic a treatment to be not useful. Still an interesting thought-experiment and a little scary.

Comment by tgb on How my social skills went from horrible to mediocre · 2015-05-20T19:17:41.973Z · LW · GW

I'm curious why you that I'm not part of your target audience – feel free to elaborate.

I'm not sure we understand each other here, but I'm assuming you want to know why I do not consider myself part of your target audience. I don't have a concrete answer here, it's just that I read this and thought it didn't apply to me. I had some of the same difficulties as you, but not in a way that I feel your advice would have applied or still does apply. I can think of a friend for whom some of your advice would probably apply, though, and imagine you are targeting him and not me.

I might be oblivious, but I don't see where I called myself smarter than the typical LW reader

This is what I take from: " I almost never went to Less Wrong meetups, because I had already thought about most of what people discussed, so that it was more efficient for me to learn on my own." and similar comments. I can see how you would take it as saying that you simply have already thought about that ahead of time and so you're not claiming to be smarter. (I've tried to explain that circumstance to people before myself.) But most people will take that to mean that you think you're too smart for them. Other comments suggest that this "most people" does not transfer to "most LW readers," though, so maybe I'm misplaced. This is also something that I thought could have been left out entirely. If you had to include this I would have suggested mollifying it a bit: make the problem seem to be yourself rather than others. Something like, "I struggled to find common ground to talk about even during LW meetups." This loses some specificity, but I'd play around with that kind of phrasing where you claim the fault is your own.

Another example would be something like describing yourself as "The guy who has deep insights but who doesn't get anything done, because he he's socially dysfunctional so nobody listens to him". This is a pretty big humble brag. If I wanted to say that to a typical person I might have said, "The guy who doesn't get anything done, no matter what insight he has, because he's socially dysfunctional so nobody listens to him." It's more cautious and definitely doesn't claim "deep insight" which is a phrase I'd reserve for describing someone else. You leave it up to the reader exactly how insightful you are implying yourself to be. It also changes the focus to your difficulty rather than the strength (which is demoted to an aside). I'm no writer though, so take this specific suggestion with a grain of salt.

Similarly for claims about deep insights from machine learning. Make the focus the difficulty you faced, not the deep insight you had. Maybe say, "I struggled even more after picking up machine learning jargon and modes of thought which I couldn't well articulate, even to my close friends."

Others have pointed out that you're also very humble throughout. I agree with them, too, and admire your ability to spell out your own failings. But people read "humble brag" mixed statements as primarily bragging. To you, it might seem really really significant that you were struggling, but that's not the focus people will read.

For the doctor analogy, I agree that that's what you're trying to say and I think you partly succeeded at that on one level. But on the other level, people will be turned off when you express expertise in areas where you do not have an obvious qualification. A doctor has a diploma to point to, and people are okay with that. A self-proclaimed student of medicine who had spent 15 years learning privately would be treated quite differently form the doctor. It's not a fair world! Had you been more specific I also might not have taken it like that, instead it seemed to be a blanket statement, like how an adult might say that all conversation with a child is tedious since the child just hasn't had any exposure to interesting ideas. Regardless of how factually true that is, the child could feel slighted.

This is all my take. Lumifer's response seems reasonable, too.

What would I have eliminated to make it shorter? It's a matter of taste, I suppose. I might have removed most of the part about how you grew up. I felt it could be summarized in a few sentences. But looking over this a second time, I think I may have clumped a lot of the things I thought came off as "arrogant" under the tag of "needs to be removed" and then interpreted that to mean that the article was too long. I'm sure it could be tightened up, but other than that growing up section there doesn't seem to be anything major. So take that complaint of mine lightly.

Ugh, I need to take my own advice and not write so much. Easier said than done.

Comment by tgb on How my social skills went from horrible to mediocre · 2015-05-20T02:00:55.109Z · LW · GW

For example, if a student tells me that I'm the worst teacher he or she has ever had, it makes me feel bad because I feel like I'm not contributing value, but I'm not at all upset with the student: my attitude is that the student is conveying valuable information to me, and that I should be appreciative.

I'm tempted to take that as a Crocker's rule invocation. But I have realized that you wrote this for people-like-you, that is, after all, pretty much its explicit purpose. As such, I'm not sure I have an criticism that I can't definitively think is helpful.

Nonetheless, I want to point out two general things about this will make this hard post to read for most people. First is the length, and even in this you note that you spend too much time explaining something that you've worked on. I think the length was partially unnecessary and not just a reflection of me not being your target audience (I assume). The second is that you come across as exceedingly arrogant. I think you are attempting to explain your background so that we understand the situation. But you explicitly call yourself smarter than the typical reader of the site that you are posting this on. Ouch! But again, perhaps this is just a reflection of you having a very narrow target audience and for them this could read like a "ah, finally someone gets it!"

I hope that you take this to be useful, particularly for when you write for a wider audience. For what its worth, my mental post it note has you labelled as a user that I should pay attention to. I say that since I kind of suspect that you already know everything I just mentioned and aren't bad at overcoming these in other situations, but thought this worth saying explicitly given the context of trying to improve.

Comment by tgb on Harry Potter and the Methods of Rationality discussion thread, February 2015, chapter 104 · 2015-02-16T21:34:09.347Z · LW · GW

We've gotten much less senses of doom with this encounter with Quirrell than any in the past as far as I can tell. Only Sprout's magic caused apprehension, Q not at all.

Was Quirrell in control of the sense of doom all along? Has something changed?

Comment by tgb on I'm the new moderator · 2015-01-16T13:15:07.311Z · LW · GW

I've been around for more than a couple years and had no idea there was a difference between Main and Promoted. I guess I check "new" for Main by using the side bar on the landing page but didn't pick up on the distinction.

Comment by tgb on 2015 Repository Reruns - Boring Advice Repository · 2015-01-11T18:41:46.500Z · LW · GW

I've been using Mint for a while and, honestly, I'm very unimpressed. You click on the "Trends" tab and what does it show you? A pie chart! Because that's obviously how you analyze your trends...

Actually I just went back there for the first time in a while and it's not as bad as I remembered if you're willing to muck around a bit. It does seem to hide the useful things under a few clicks and force you to do comparisons manually through filtering out certain categories instead of showing you them all at once, things like that. But it's still a lot better than going through a number of accounts. Also, you can type for suggestions when classifying things which is something I had missed before and was a major complaint, having to wander through their list looking for something. Still, though, I'd be open to suggestions for alternatives.

Comment by tgb on Has LessWrong Ever Backfired On You? · 2014-12-16T19:08:03.634Z · LW · GW

I one time asked for advice here and the responses felt overly demeaning and presumptuous and largely ignored trying to help in favor of lambasting me for being in the situation at all. It was not a response I had been expecting and it made me feel bad and less likely to ask somewhat personal questions in the future. I don't think anyone replying was intending to cause me any harm and it wasn't a big deal in any sense of the word. But I felt disappointed with the outcome and the community.

I'm sure anyone sufficiently interested could find this out of my post history, but the details aren't particularly interesting. To a third party it probably won't seem like much at all, but at the time to me it wasn't a good feeling.

Comment by tgb on The "best" mathematically-informed topics? · 2014-11-14T21:04:43.769Z · LW · GW

The book Music and Mathematics is a bit of a jumble of essays from various people. It's not a coherent whole or anything, BUT, if you are the least bit interested in this subject I would strongly you recommend reading the first essay (i.e. chapter). A lot of "music theory" mumbles around circles of fifths and chord progressing and things with vague pretenses of being math, but that chapter includes the only fundamental mathematical idea in the entire book.

Conveniently that chapter, and some others, are here.

Comment by tgb on First(?) Rationalist elected to state government · 2014-11-07T03:17:38.744Z · LW · GW

Pretty cool! Also note that the NH House of Representatives is a massive body of 400 representatives, which is somewhat absurd considering how small NH is. Now we'll have to see if she can bring anything new to a land already full of unusual politics.

Comment by tgb on November 2014 Media Thread · 2014-11-07T02:38:07.109Z · LW · GW

Boyhood is one of the better movies I've seen recently.

Comment by tgb on What is the difference between rationality and intelligence? · 2014-08-14T01:10:38.560Z · LW · GW

Intelligence is INT while rationality is WIS.

Comment by tgb on Bragging Thread, July 2014 · 2014-07-14T16:05:41.895Z · LW · GW

Neat! Can you give a really short description of why this is useful or of the most interesting techincal aspects?

Actually, can you just tell me what's going on in the second movie where the grid appears to stop growing closer?

Comment by tgb on [QUESTION]: What are your views on climate change, and how did you form them? · 2014-07-10T11:38:43.871Z · LW · GW

It's likely that a disproportionate account of optimization of human welfare has occurred in the last few centuries. Moreover people are mobile and the variation in temperatures over the surface of the earth is greater than over a few thousand years. So humans are likely to have optimized their location to approximately optimize their welfare.

Comment by tgb on Come up with better Turing Tests · 2014-06-11T15:01:20.733Z · LW · GW

Yup, I really missed that. Whoops.

Comment by tgb on Come up with better Turing Tests · 2014-06-10T14:38:56.566Z · LW · GW

So long as the bots are easy to distinguish from humans, it'll be easy for competitions to produce false positives: all it takes is for the judges to want to see the bot win, at least kind of. If you want a real challenge, you'd better reward the judges significantly for correctly distinguishing human from AI.

Comment by tgb on Open thread, 9-15 June 2014 · 2014-06-10T14:26:12.567Z · LW · GW

Another reason: ransomware!

Comment by tgb on [News] Turing Test passed · 2014-06-10T14:19:58.388Z · LW · GW

I don't think that was a judge conversation. That was just someone using the online chat program:

"I logged on to what I think is the Goostman program. Here’s the transcript of our conversation: (Eugene is supposed to be around 13 years old.)"

Not only that, but it's an old version from a year ago. (Not that I think the real judges' conversation would be significantly better.)

Comment by tgb on Links! · 2014-06-05T15:35:21.459Z · LW · GW

The number of times I've read an article about something like this that gives it away in the title or opening before giving the reader a chance to experience it for themselves... thanks for not explaining.

Comment by tgb on June 2014 Media Thread · 2014-06-01T21:59:59.046Z · LW · GW

Thanks for filling up my Pocket queue!

Comment by tgb on Open Thread, May 26 - June 1, 2014 · 2014-05-28T18:32:18.037Z · LW · GW

While a good point, the OP's link says that:

  • There is only a moderate correlation between income and taking of test prep
  • Under-performing minorities are more likely to take test prep than whites

In other words, quite a few people taking test prep are ones likely to be going to poor, under-performing school systems. Either test prep companies are incompetent or our school system is doing a lot better than I had expected, even on the low end!

Comment by tgb on Open Thread, May 26 - June 1, 2014 · 2014-05-27T15:53:00.403Z · LW · GW

I like your points, but it does seem awfully surprising that there would not be an improvement on, for example, the reading section which has a number of questions that are little more than vocabulary tests. Vocabulary is easy to study and if you don't know it you have little chance of figuring it out.

Or for the math section, someone who hasn't taken geometry before will do very poorly on any geometry questions on the test. (I think this is an unobjectionable claim.) So it would be surprising if people who studied geometry at all suddenly get all the possible benefit of studying - studying doesn't seem like it should be a binary thing where you hop from no knowledge/poor performance to full knowledge/as-good-as-you-could-get performance.

Note that the two articles cited by the OP's link are not randomized controlled trials and are both actually based off the same survey data.

I, too, join the OP in confusion and mild skepticism of the research.

Comment by tgb on Open Thread, May 19 - 25, 2014 · 2014-05-23T15:31:22.915Z · LW · GW

I think it's clear that you know more about what economists mean than I do, but when the typical person hears that a depression is ending, they imagine people being happier than they were before. I'm not really claiming that anyone thinks that crashing consumer spending + mass shortages = better living standards, just that the average Joe in the US hears about the depression ending and not about those negative things.

Anyway, not sure what point I'm trying to make since I think you already know what I'm saying.

Comment by tgb on Open Thread, May 19 - 25, 2014 · 2014-05-21T20:54:59.816Z · LW · GW

Krugman doesn't quiiiite come out and say it, but he sure seems to want the reader to infer that living standards rose: And in that article, he quotes and quote of Rick Perry's book saying that the recovery happened because of WW2 (due to forcing FDR to "unleash private enterprise", oddly).

So maybe no one actually makes that argument, but boy it's common for people (economists and politicians!) to imply it. (Look at the contortions Perry goes through to not have to refute it!) It's always nice to notice the confusion a cached thought should have made all along.

Comment by tgb on Open Thread, May 19 - 25, 2014 · 2014-05-21T20:43:30.104Z · LW · GW

Do you know of a typical measure (or component) of living standard that would have been measured for the US across both the great depression and WW2? The standard story I have heard informally is that WWII efforts did actually increase standards of living. I'm not surprised to learn that that's false, but given the level of consensus in the group-think I've encountered, I'd be interested in seeing some hard numbers. Plus, I'm interested in seeing whether there was a drop in living standards.

Comment by tgb on Open Thread, May 19 - 25, 2014 · 2014-05-21T20:01:17.474Z · LW · GW

I like that this explanation gives a good reason for why this kind of spending could only work to fix a depression or similar situation versus always inflating standards of living. Thanks.

Comment by tgb on Open Thread, May 19 - 25, 2014 · 2014-05-19T15:33:11.149Z · LW · GW

This just struck me: people always credit WWII as being the thing that got the US out of the great depression. We've all seen the graph (like the one at the top of this paper) where standard of living drops precipitously during the great depression then more than recovers during WWII.

How in the world did that work? Why is it that suddenly pouring huge resources out of the country into a massive utility-sink that didn't exist until the start of the war rapidly brought up the standard of living? This makes no sense to me.

The only plausible explanation I can think up is that they somehow borrowed from the future using the necessities of war as justification. I feel like that would involve a dip in the growth rate after WWII - and there is one, but it just dips back down to the trend-line not below like I would expect if they genuinely borrowed enough from the future to offset such a large downturn as the great depression. The only other thing seems to be externalities.

However this goes, this seems to be a huge argument in favor of big-government spending (if we get this much utility from the government building things that literally explode themselves without providing non-military utility, then in a time of peace, we should be able to get even more by having the government build things like high-tech infrastructure, places of beauty, peaceful scientific research, large-scale engineering projects, etc.). So should we be spending 20-40% of our GDP on peace-time government mega-projects? It's either that or this piece of common knowledge is wrong (and we all know how reliable common knowledge is!).

Or I'm wrong, of course. So what is it?

(Bonus question: why didn't WWI see a similar boost in living standards?)

Comment by tgb on Open Thread April 8 - April 14 2014 · 2014-04-13T15:42:19.294Z · LW · GW

It checks for you, by the way, and will block an attempt and notify you if it looks suspicious. This happened to me earlier this month. Interestingly, that happened 4 days after the vulnerable OpenSSL version was released and my Gmail password is basically the only the one which I do not reuse anywhere and I don't know how anyone could have gotten it... Still more likely to have been a keylogger or something.

Comment by tgb on Open thread, 24-30 March 2014 · 2014-03-27T13:12:57.826Z · LW · GW

It absolutely could be all of the above! But see I write questions like this fairly frequently: I notice something surprising and don't have a good explanation for it. I then write down the question and pose a couple possible explanations which makes me think of more possible explanations. Frequently I realized that taken together the possibilities I thought of are enough to explain what I was surprised at and I don't even ask the question. Other times, like this one, I still feel like I'm missing something. So I ask the question.

In this case it looks like the biggest thing I missed was how much these sales' values depend upon the moment-to-moment stock prices of the parties involved and so that not rounding them hardly even makes sense, as you and bramflakes point out. Thanks!

Comment by tgb on Open thread, 24-30 March 2014 · 2014-03-27T02:11:50.675Z · LW · GW

Facebook bought Oculus Rift for $2 billion. What makes this, and so many other large deals, such clean numbers? Are the press rounding the details? Are the companies only releasing approximate or estimate numbers? Can the value of a company like Oculus really not be estimated to the nearest 10%? Or do these whole numbers just serve as nice Schelling points on which to hinge a bargain? Or am I forgetting lots of ugly-numbered deals?

(WhatsApp purchase was 2 significant figures, and this list on Wikipedia does show mostly 2-3 significant figures though some figures are probably converted from other currencies.)

Comment by tgb on Open Thread: March 4 - 10 · 2014-03-06T03:57:35.083Z · LW · GW

Ah... thanks.

Comment by tgb on March 2014 Media Thread · 2014-03-04T13:01:47.201Z · LW · GW

Great - added to my reading list!

Comment by tgb on Open Thread: March 4 - 10 · 2014-03-04T12:58:32.632Z · LW · GW

Let me just state some obvious ones:

  • Exercise
  • Saving money
  • Eating well
  • Anki
  • Anything in the longevity guide
  • Practicing a new skill (eg: musical instrument, cooking, a sport or hobby)
  • Learn how to make bulleted lists in this system, why isn't this working?
  • Overcoming an annoying habit like biting nails
  • Meditating regularly
Comment by tgb on March 2014 Media Thread · 2014-03-02T22:08:41.741Z · LW · GW

I read "Quantum Computing Since Democritus" by Scott Aaronson and loved it but can only hesitantly recommend it. I think I got significantly more utility out of it than most would as I happened to hit the "sweet spot" for being sufficiently equipped to not be hindered by its (clearly marked and understandable) omissions while still having enough gaping holes in my knowledge that I got a lot out of it. For the record, my background is in math with one course in quantum mechanics, one in intro course in quantum computation and no training in complexity classes.

I have never had as much fun with a book this technical before. Please take that as a challenge and recommend me some competitors!

It covers a huge range of material in a very light and enjoyable manner - I frequently found myself laughing out loud. His exercises were, for me, impressively at the level of being just hard enough that they look ridiculously challenging (in one case, apparently impossible!) at first read over but still having an approachable solution after five minutes of thought. He goes over many things that are directly interesting to this community (the anthropic principle, self-identification assumption, Newcomb's problem, even time travel) and even though I have already read up on those a significant amount found a lot of good stuff in those sections where he does actually manage to relate it to quantum computation.

This is not a traditional book on quantum computing and its title is perhaps misleading. It really is about computational complexity. You will not learn Shor's algorithm for factoring numbers and Aaronson gives only a slight overview of, for example, what quantum gates do. Prior knowledge of this would be helpful but you could still get something out of it. Don't expect to learn this or you will be disappointed. The computational complexity material is however first rate.

It's based off of freely available lecture notes from his website under the same name. I enjoyed this book enough that I like owning a physical copy, but you could probably read it online without much loss. The main difference is some updates for results that came out since the course was originally held (which are valuable). In some sense, I suspect that the online lecture notes are margnially better - for example the book doesn't have colored diagrams even though the text in at least one case refers to colors on the diagram.

Comment by tgb on Is IQ what we actually need to know? · 2014-03-02T20:21:13.240Z · LW · GW

It actually was from a larger conversation. Nancy brought it up at the last Philadelphia meet up. I don't think any relevant details are missing and we didn't come to any thoughts that haven't already been brought up in the comments here. Was some part unclear?

Comment by tgb on "Smarter than us" is out! · 2014-02-26T23:37:49.656Z · LW · GW

What, if anything, should regular LessWrong readers expect to get from reading this book?

Comment by tgb on Open Thread for February 18-24 2014 · 2014-02-22T13:53:50.610Z · LW · GW

They almost always end with the actual murderer being accused and then immediately getting angry and confessing thereby giving them the only actual hard evidence that could ever be used in a conviction. It's convenient that way. (See, for an extreme example, one of the episodes of the recent season 3 of Sherlock.)

Comment by tgb on Mental Subvocalization --"Saying" Words In Your Mind As You Read · 2014-02-18T13:33:16.880Z · LW · GW

Yes, but my comprehension and speed decreases. (Or it feels that way, I haven't actually tested this. This also doesn't feel like the right test to me - I can 'squeeze' counting in between reading, so it's not necessarily the case that one isn't blocking the other, it might just be that I'm alternating between them quickly.)

Edit: And I didn't mean anything deep when I used the Chinese Room as an example! It was just meant to illustrate the enormous gulf in my mind between 'comprehending through subvocalization' versus 'comprehending despite not subvocalizing.' I tried several other metaphors before settling on that. I would be surprised to hear that there was a significant correlation between subvocalization and acceptance of the Chinese Room results and did not mean to imply that, though I suppose it is an interesting question regardless.

Maybe this metaphor makes my meaning clearer: asking me to read without subvocalizing would be like asking me to look at a painting without ever experiencing the qualia of seeing it. If I shut down that qualia somehow (and I don't think could), then it's still possible that some portions of my mind are becoming aware of what was on that picture and maybe you could discover that through some clever experiments. But that part of my brain that would have learned that doesn't feel like it could be 'me' in the same way that it doesn't feel like it could be 'me' that understands some read text without subvocalization.

Comment by tgb on Mental Subvocalization --"Saying" Words In Your Mind As You Read · 2014-02-15T14:52:01.677Z · LW · GW

As an always-subvocalizer, I'm not even sure what it would mean to read something without subvocalizing it. For me, that would be like the Chinese room thought experiment - maybe something is understanding that sentence, but it's not 'me.'

(Edit: also my reading speed is ~240wpm when I actively try to read fast, depending, obviously, upon the material. I suspect my typical is more like 200wpm.)

Comment by tgb on [Open Thread] Links (2014-02-14) · 2014-02-15T13:52:58.852Z · LW · GW

If you like puzzle games, try out this free one: "Jelly No Puzzle". Easily some of the most interesting puzzles I've ever encountered in a game.

Comment by tgb on MEETUP: February 9, Philadelphia · 2014-02-09T13:34:00.774Z · LW · GW

And it's at 1PM, according to the email list.

Comment by tgb on How can I spend money to improve my life? · 2014-02-03T14:04:09.468Z · LW · GW

I brainstormed about this for a while and reflected on what past purchases I had made that I liked and hadn't liked. YMMV but here's what I concluded:

  • Incremental upgrades to existing items is rarely worth it unless the existing option is terrible; a nice mouse is only slightly nicer than whatever you're currently using, most of the time you don't really use the extra speed from a faster computer, etc.
  • Conversely, get things that open up new options for you; e.g. a smartphone lets you do computer things in places where you wouldn't before so is better than a new laptop in many case
  • "buy experiences" seems to work; buy a musical instrument and take lessons or other things, learn to bake (and then give away your baked goods to friends)
  • Weather-specific clothes can make some situations much easier: if you're cold, buy something warm. It's not actually hard to stay warm and being prepared means you don't have to worry about what going out and doing things based on the weather.
  • Spending money to remove minor inconveniences can be worth it; if it bothers you every day then change it
  • I buy way more video games than I finish anymore, but still value completing them so I prefer to buy short indie games now
  • I consistently over-estimate fancy new electronics and should correspondingly reduce my expectations for things like, say, the Oculus Rift
Comment by tgb on Rationality Quotes January 2014 · 2014-01-07T16:37:13.183Z · LW · GW

Surely you should do "0) consider why someone thinks it's wrong and whether there is merit to their view" before either 1) or 2)? Or is there some more context to this quote that makes this objection less relevant?

Comment by tgb on Open thread for December 9 - 16, 2013 · 2013-12-15T16:39:20.205Z · LW · GW

As usual, has some information related to this.

Comment by tgb on Open thread for December 9 - 16, 2013 · 2013-12-10T03:44:46.419Z · LW · GW

If you haven't yet taken EY's suggestion in the author's notes to read Worm yet, do so. It's original fiction, but you probably don't mind.

Edit: also this might belong in the media thread?

Comment by tgb on Personal examples of semantic stopsigns · 2013-12-08T16:03:47.023Z · LW · GW

I don't know. I read and writ a lot of math (where "obvious" is probably most likely to be overused) and I do try to double-think every time I write it down. But when reading, it means to me "this statement SHOULD be obvious, if it's not, then you're missing something and you should probably reread what just happened." If it doesn't say that and just states the result I am less likely to give it a second thought, perhaps ironically. If they don't say "obvious" then maybe they're quoting some theorem I don't know and that often isn't very relevant unless I need to know every detail of this. If they do say "obvious" then they've at least judged that theorem (or definition or whatever) to be sufficiently basic as to be worth my while learning. Not only that, but if I think something seems obvious and the author thought it was sufficiently obvious to call it obvious, then I'm probably not missing anything subtle!