"Tweet id","Tweet permalink","Tweet text","time","impressions","engagements","engagement rate","retweets","replies","likes","user profile clicks","url clicks","hashtag clicks","detail expands","permalink clicks","app opens","app installs","follows","email tweet","dial phone","media views","media engagements","promoted impressions","promoted engagements","promoted engagement rate","promoted retweets","promoted replies","promoted likes","promoted user profile clicks","promoted url clicks","promoted hashtag clicks","promoted detail expands","promoted permalink clicks","promoted app opens","promoted app installs","promoted follows","promoted email tweet","promoted dial phone","promoted media views","promoted media engagements" "1181307096089972738","https://twitter.com/gwern/status/1181307096089972738","@SilverVVulpes fox are you ok (Elon might reply to you too someday. Ganbarre!)","2019-10-07 20:35 +0000","439.0","15.0","0.03416856492027335","0.0","1.0","5.0","1.0","0.0","0.0","8.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1181293776700096517","https://twitter.com/gwern/status/1181293776700096517","@davedixx @gdb You know that there's been like 6 other similar-scale (or larger) models trained since then, right? Conor's, skylion's, Grover, XLNet, CTRL, and Megatron (so far, off the top of my head).","2019-10-07 19:42 +0000","60.0","5.0","0.08333333333333333","0.0","1.0","1.0","3.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1181293398889816065","https://twitter.com/gwern/status/1181293398889816065","@gdb I think sometimes about when I reread old comments of mine I don't remember writing, and about reading the version of me from the GPT-2-117M we trained on our IRC channel logs. That would be a very spooky kind of gaslighting.","2019-10-07 19:41 +0000","629.0","11.0","0.017488076311605722","0.0","0.0","8.0","0.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1181204829424771073","https://twitter.com/gwern/status/1181204829424771073","@robkhenderson On the other hand, experimental results like https://t.co/3Gv9KgmAk4 suggest those people on the margin would've been better off going for it.","2019-10-07 13:49 +0000","724.0","30.0","0.04143646408839779","0.0","0.0","3.0","3.0","24.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1181015063366422528","https://twitter.com/gwern/status/1181015063366422528","@richardfuisz @ArtirKel @rivatez Yeah, beer, yogurt, kombucha, all these are easy. I use commercial Greek yogurt as a culture all the time. Surviving baking is a little more extreme. While SF microbiome surely contributed to the original culture, I doubt it's important ongoing. Cultures are highly resilient.","2019-10-07 01:15 +0000","101.0","5.0","0.04950495049504951","0.0","2.0","1.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1181014730963599360","https://twitter.com/gwern/status/1181014730963599360","@GeorgeTyler1998 @ArtirKel @rivatez Only one way to find out. I'm sure it's not worse than whatever starter you're using now, at least.","2019-10-07 01:13 +0000","48.0","4.0","0.08333333333333333","0.0","1.0","3.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1181004784284905472","https://twitter.com/gwern/status/1181004784284905472","@SteveStuWill https://t.co/D4ldaGsOk0 There's a lot of competition, though, of all kinds of beverages and drugs. Plus internal fratricide: may simply be stealing market share from each other, rather than going into unusual places advertising to non-alcohol drinkers.","2019-10-07 00:34 +0000","665.0","11.0","0.016541353383458645","0.0","0.0","2.0","0.0","8.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1181003658256891905","https://twitter.com/gwern/status/1181003658256891905","@ArtirKel @rivatez Since a sourdough starter is just a trade secret and you agree to no contracts when you buy the loaf, it's not like they can stop you in any way, and you can keep buying loaves until it works. (How do you know you have Tartine's? Well, if it's delicious, who cares if it's not?)","2019-10-07 00:29 +0000","266.0","10.0","0.03759398496240601","0.0","2.0","5.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1181003412999229443","https://twitter.com/gwern/status/1181003412999229443","@ArtirKel @rivatez I wonder if one could pirate the Tartine sourdough starter? A loaf would contain trillions of yeast; even after baking for 30+ minutes, there's a chance at least *some* are still alive? (Yeast can be tough.) If you blenderize it & begin a new culture...","2019-10-07 00:28 +0000","864.0","14.0","0.016203703703703703","0.0","1.0","6.0","0.0","0.0","0.0","7.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180964489388285953","https://twitter.com/gwern/status/1180964489388285953","@devonzuegel @zooko @rsnous Even with the 1% rule, there's still too many who want to put their thumbprint on something: https://t.co/zyPeaiNk4A","2019-10-06 21:54 +0000","823.0","52.0","0.06318347509113001","0.0","0.0","7.0","2.0","40.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180692978001678336","https://twitter.com/gwern/status/1180692978001678336","@collision @patrickc DATE: WAITER: ABSOLUTELY NO ONE: PATRICK: ""...Did you know median per capita income in this census tract is fully ????? percent of what it would be if American productivity progress had continued its pre-1971 growth trends¡ª""","2019-10-06 03:55 +0000","5580.0","234.0","0.041935483870967745","0.0","0.0","90.0","57.0","0.0","0.0","87.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180684932164804608","https://twitter.com/gwern/status/1180684932164804608","And that subreddit has been deleted as well, so here's a link to the main BigQuery export of DNM-subreddits' comments: https://t.co/FxCbj0y18G","2019-10-06 03:23 +0000","13138.0","252.0","0.019181001674531892","5.0","0.0","27.0","22.0","110.0","0.0","88.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180612629594869763","https://twitter.com/gwern/status/1180612629594869763","@ArtirKel @Miles_Brundage I dunno, I've played a fair bit of both chess and Go and I struggle to think of anything except the very most abstract concept of a 'two player game where you take turns' which really transfers.","2019-10-05 22:36 +0000","518.0","8.0","0.015444015444015444","0.0","1.0","0.0","0.0","0.0","0.0","7.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180600372009717761","https://twitter.com/gwern/status/1180600372009717761","@Miles_Brundage As always, ""AI is whatever we can't do yet.""","2019-10-05 21:47 +0000","406.0","5.0","0.012315270935960592","0.0","0.0","1.0","0.0","0.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180600267005292545","https://twitter.com/gwern/status/1180600267005292545","@ArtirKel @Miles_Brundage I don't see why you would expect finetuning to be all that useful there (beyond maybe initial CNN layers). Chess players do not become great Go players or vice-versa. E. Lasker, for example, never became more than a mediocre Go player, and had no speedup over a kid.","2019-10-05 21:46 +0000","457.0","7.0","0.015317286652078774","0.0","1.0","0.0","2.0","0.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180548048268808193","https://twitter.com/gwern/status/1180548048268808193","@tailcalled The typical mind fallacy is terrifying indeed: https://t.co/4CWjTkqpRK","2019-10-05 18:19 +0000","114.0","2.0","0.017543859649122806","0.0","0.0","0.0","0.0","2.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180547759860006912","https://twitter.com/gwern/status/1180547759860006912","@0K_ultra Yes, inner voice/monologue is another one of those I've compiled in my list: https://t.co/BEVFZwNwwW https://t.co/rkPRBTo5pU https://t.co/5lfdXnyyUn https://t.co/iHlvPCXsQ3","2019-10-05 18:18 +0000","340.0","20.0","0.058823529411764705","0.0","0.0","2.0","3.0","15.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180508521957335042","https://twitter.com/gwern/status/1180508521957335042","@KirkegaardEmil @samsaragon Doesn't have to be general. I wouldn't expect the phonological loop to help with reaction time, or crystallized knowledge like vocab. But it could help with some things involving manipulation of items in short-term or working memory (as RAPM stresses) by enabling 'chunking'.","2019-10-05 15:42 +0000","291.0","3.0","0.010309278350515464","0.0","0.0","0.0","0.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180309149424508928","https://twitter.com/gwern/status/1180309149424508928","@brianchoffman I think '90s kids will appreciate it the most. The Rotten Library was kind of like a snarkier Wikipedia pre-Wikipedia, wasn't it? Nowadays you can find anything on WP, but harder to find that sort of writing back then.","2019-10-05 02:30 +0000","264.0","3.0","0.011363636363636364","0.0","1.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180305645284798467","https://twitter.com/gwern/status/1180305645284798467","I've made a mirror of the https://t.co/kGlrLn4sWz Library: https://t.co/SLqXpDAVJq (Discussion: https://t.co/oGdHlnoJA2 )","2019-10-05 02:16 +0000","6478.0","297.0","0.04584748379129361","1.0","3.0","20.0","6.0","257.0","0.0","10.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180305361980596224","https://twitter.com/gwern/status/1180305361980596224","@HenryKrinkIe @ByYourLogic I've put up a mirror at https://t.co/SLqXpDAVJq","2019-10-05 02:15 +0000","79.0","3.0","0.0379746835443038","0.0","0.0","1.0","1.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180278774631415808","https://twitter.com/gwern/status/1180278774631415808","There's an SDAM subreddit now: https://t.co/9fasbbVPd8 because of course there is, why didn't it occur to me to search for it immediately.","2019-10-05 00:29 +0000","4917.0","157.0","0.031930038641448034","1.0","2.0","16.0","6.0","123.0","0.0","9.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180248930120065024","https://twitter.com/gwern/status/1180248930120065024","@BVLSingler @CSERCambridge But they're not entirely wrong, are they? The headline does demonstrate how much the image of the Terminator dominates public perception: want to know what CSER does? You can sum it up as 'Terminator studies' and expect the newspaper-reading public to understand instantly.","2019-10-04 22:30 +0000","575.0","4.0","0.006956521739130435","0.0","0.0","1.0","2.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180246757147582464","https://twitter.com/gwern/status/1180246757147582464","Refs on SDAM lack of long-term memory: https://t.co/lJuOqigVyW https://t.co/gJ7Rkuj5Jz https://t.co/zL6xcc2RvU https://t.co/tqcNEl4zIb https://t.co/O8OGa5dIn5 https://t.co/lghVF5LMXe","2019-10-04 22:22 +0000","6281.0","92.0","0.014647349148224804","2.0","1.0","12.0","5.0","71.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180246754849181698","https://twitter.com/gwern/status/1180246754849181698","Oh my god, there ??? people without any qualia of long-term memory: ""Severely Deficient Autobiographical Memory"" syndrome. And they say the exact same thing as aphantasics/anosmics/color-blind: https://t.co/epmJwhxGGP","2019-10-04 22:22 +0000","12723.0","313.0","0.024601116088972728","32.0","10.0","124.0","44.0","5.0","0.0","98.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180211165345001472","https://twitter.com/gwern/status/1180211165345001472","@zackwhittaker I wonder if that reflects the tech stack of the campaigns in general? Are all the Google users on Linux stacks?","2019-10-04 20:00 +0000","1617.0","1.0","6.184291898577613E-4","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180175384261939201","https://twitter.com/gwern/status/1180175384261939201","@SteveStuWill Looks like Twitter messes this one up by recompressing it or something. Original: https://t.co/Oj1XXew1gw","2019-10-04 17:38 +0000","548.0","23.0","0.041970802919708027","0.0","1.0","0.0","2.0","19.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1180161816854503425","https://twitter.com/gwern/status/1180161816854503425","September links: https://t.co/u4TklelgGy","2019-10-04 16:44 +0000","4878.0","129.0","0.026445264452644526","0.0","0.0","15.0","0.0","110.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179947544010473472","https://twitter.com/gwern/status/1179947544010473472","@konradzolna @DeepMindAI I found the TRAIL part of the paper very difficult to read. It describes the constrained optimization as focusing on 'meaningful behavior' but then just drops in a bunch of equations and moves on. (But what does all of that mean?)","2019-10-04 02:33 +0000","878.0","8.0","0.009111617312072893","0.0","1.0","1.0","6.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179877519027249152","https://twitter.com/gwern/status/1179877519027249152","@jmhessel @yoavgo You could look at the history of progress curve in BLEU for WMT, look at the GPU count in each, guesstimate at number of labs/researchers actively working on the problem, and back out a 'marginal researcher-GPU' trendline and see how many years +3.2 BLEU is.","2019-10-03 21:54 +0000","114.0","2.0","0.017543859649122806","0.0","0.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179860140197060608","https://twitter.com/gwern/status/1179860140197060608","@NeilLewisJr @StuartJRitchie @tuckerdrob Even more related - it doesn't: @StuartJRitchie et al (2015): https://t.co/lEhfUj76Ur","2019-10-03 20:45 +0000","1365.0","154.0","0.11282051282051282","0.0","0.0","7.0","7.0","135.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179837117440114690","https://twitter.com/gwern/status/1179837117440114690","@natfriedman Rereading https://t.co/RtbK2AHRUC & https://t.co/yNytUTn0zI reminds me how good they & _Close to the Machine_ were. (Also of how Salon didn't suck back then.)","2019-10-03 19:14 +0000","8893.0","68.0","0.0076464635106263355","1.0","0.0","11.0","0.0","51.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179834978424774656","https://twitter.com/gwern/status/1179834978424774656","@NoamJStein ""My likes serve...in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them¡ªas steps¡ªto climb beyond them to bluecheckdom. (He must, so to speak, throw away the ladder after he has climbed up it.)""","2019-10-03 19:05 +0000","1868.0","41.0","0.021948608137044967","1.0","0.0","13.0","23.0","0.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179795091684302855","https://twitter.com/gwern/status/1179795091684302855","@EthanZ Mirror: https://t.co/ThR74sn34R","2019-10-03 16:27 +0000","317.0","5.0","0.015772870662460567","0.0","0.0","0.0","0.0","5.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179602559151017984","https://twitter.com/gwern/status/1179602559151017984","@Meaningness I'm not sure about that. Contact between Tibet and Europe was very sparse. (They're still debating whether Hume was influenced or not.) The combination of 'red is maximal contrast' and 'red inks are very cheap & easy to make' explains parallel evolution of sacred rubrication.","2019-10-03 03:42 +0000","702.0","7.0","0.009971509971509971","0.0","1.0","1.0","0.0","0.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179600807123177472","https://twitter.com/gwern/status/1179600807123177472","@albrgr @InsidePhilanthr The all in Mastercard shares seems suspicious. What responsible charity puts all its assets into a single stock rather than an index or something? I assume the cynical explanation here is that it's a way to lock up shares indefinitely and prop up the price for insiders' benefit.","2019-10-03 03:35 +0000","744.0","4.0","0.005376344086021506","0.0","1.0","3.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179597490733105152","https://twitter.com/gwern/status/1179597490733105152","@tophtucker You joke about the shock collar but I know of someone who was trying that. (Semi-inspired by the Pavlok.) They thought it might be working...","2019-10-03 03:22 +0000","2175.0","12.0","0.005517241379310344","0.0","1.0","3.0","6.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179597167402590208","https://twitter.com/gwern/status/1179597167402590208","@anniefryman @fulligin @timhwang now you too know my sorrow","2019-10-03 03:20 +0000","244.0","12.0","0.04918032786885246","0.0","1.0","3.0","1.0","0.0","0.0","7.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179595133655883776","https://twitter.com/gwern/status/1179595133655883776","@anniefryman @fulligin @timhwang I saw this one as a kid and it *still* haunts me: https://t.co/FQwnf4brln","2019-10-03 03:12 +0000","194.0","35.0","0.18041237113402062","0.0","1.0","2.0","0.0","30.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179591031853060096","https://twitter.com/gwern/status/1179591031853060096","@LauraDeming People have done this in citation analysis by tracing copying of bibliographic errors (https://t.co/JzjZ69v56D), which indicates they didn't look it up or check the citation details: eg https://t.co/NbN61BPDRq estimates ""only about 20% of citers read the original.""","2019-10-03 02:56 +0000","939.0","15.0","0.01597444089456869","0.0","2.0","4.0","6.0","1.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179583245379411969","https://twitter.com/gwern/status/1179583245379411969","@antonhowes @ArtirKel @nayafia That makes sense for things which can be incremental. Start a novel, if it doesn't work out, fine. But I don't know how that'd work for many other subscribed things like the Dictionary or Arabian Nights.","2019-10-03 02:25 +0000","621.0","3.0","0.004830917874396135","0.0","0.0","0.0","0.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179582440488275968","https://twitter.com/gwern/status/1179582440488275968","@anniefryman It raises so many thought-provoking questions. Like: ""where is that cable going"".","2019-10-03 02:22 +0000","484.0","11.0","0.022727272727272728","0.0","1.0","1.0","7.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179578950357540865","https://twitter.com/gwern/status/1179578950357540865","@dpentecost So if red on white bg/black fg is fine, what is the equivalently contrasting color for white fg/black bg?","2019-10-03 02:08 +0000","152.0","2.0","0.013157894736842105","0.0","1.0","0.0","0.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179578633314344960","https://twitter.com/gwern/status/1179578633314344960","@sainieh It's not even consistent! The second one looks fine, it's the first one which looks awful. And they both are like 0.5MB anyway and really didn't need any compression (I run advpng/pngnq on all my PNGs in the first place).","2019-10-03 02:07 +0000","170.0","2.0","0.011764705882352941","0.0","0.0","2.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179565035800186882","https://twitter.com/gwern/status/1179565035800186882","Open question: as red appears to be the ??????????? ????????? ???????? color (https://t.co/DS7b9fxdoF), can red profitably be employed in dark mode to make up for issues? User reaction to red (https://t.co/EPnk54DC1n): ""looks like a vampire blog"".? https://t.co/7qF4dhpVrs","2019-10-03 01:13 +0000","6291.0","202.0","0.032109362581465586","1.0","9.0","15.0","0.0","62.0","0.0","5.0","0.0","0","0","0","0","0","110","110","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179563310070603778","https://twitter.com/gwern/status/1179563310070603778","- eg https://t.co/0iM0b7fGPM https://t.co/VFZ4VcYEQC - Berlin Kay: red first color named after B/W: https://t.co/FZsv0MPAKN - Tufte: Bryne's primary colors (black/red/yellow/blue) maximize contrast: https://t.co/4dGalDNxU0 Perhaps red just maximally contrasts with B/W pages?","2019-10-03 01:06 +0000","4755.0","47.0","0.009884332281808622","0.0","0.0","2.0","0.0","30.0","0.0","15.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179558464059854848","https://twitter.com/gwern/status/1179558464059854848","@ESYudkowsky Something something ultrafinite recursion: https://t.co/wdRaRJW9tx","2019-10-03 00:47 +0000","1506.0","102.0","0.06772908366533864","1.0","0.0","11.0","5.0","77.0","0.0","8.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179551734265454594","https://twitter.com/gwern/status/1179551734265454594","@goodfellow_ian @dwf If you believe barriers are a good thing, it's not hard to say 'I fully and without qualifications endorse these barriers, which are an actual solution, and strongly encourage the administration to implement additional solutions as well.' Instead of... what you have been saying.","2019-10-03 00:20 +0000","1325.0","24.0","0.018113207547169812","0.0","3.0","1.0","10.0","0.0","0.0","10.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179550725275373568","https://twitter.com/gwern/status/1179550725275373568","@goodfellow_ian @dwf Really? Because saying something is 'not an actual solution' sure does sound like 'you shouldn't do it', and neither of you chose to correct it in several tweets beforehand, although you had plenty of time to police and call out other Twitter users for disagreeing with you.","2019-10-03 00:16 +0000","1915.0","53.0","0.027676240208877285","0.0","3.0","8.0","14.0","0.0","0.0","28.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179549902440976386","https://twitter.com/gwern/status/1179549902440976386","@goodfellow_ian @dwf I don't see it as sad at all. I see it as sadder that people will take letting perfect be the enemy of better to the extent of encouraging suicide.","2019-10-03 00:13 +0000","1191.0","43.0","0.03610411418975651","0.0","2.0","5.0","5.0","0.0","0.0","31.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179549105930096641","https://twitter.com/gwern/status/1179549105930096641","@dwf You criticize it and advocate not doing something which will save kids' lives, kids like those dear friends. All those bridge jumpers were also dear friends of people, not 'body counts'.","2019-10-03 00:09 +0000","429.0","8.0","0.018648018648018648","0.0","1.0","3.0","2.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179548781626499074","https://twitter.com/gwern/status/1179548781626499074","@GabeNicholas @logic_magazine How exactly did they buy 'thousands' of ether in 2014 and are still at subsistence level? The 2014 ICO was $0.3, and it's now $180. That's a return of like $180k for each of those thousands. Maybe you should've pushed them harder on how 'profitable' their daytrading/mining is.","2019-10-03 00:08 +0000","480.0","2.0","0.004166666666666667","0.0","0.0","0.0","2.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179540495577604097","https://twitter.com/gwern/status/1179540495577604097","@marcoJcosta @s_r_constantin @deadalnix Indeed, doesn't have anything to do with 'intent to live'. It's about self-control and the long-term and maturing humanity. (Covered a little in https://t.co/efhO2HsKZu & https://t.co/cMsItzw8wX )","2019-10-02 23:35 +0000","112.0","2.0","0.017857142857142856","0.0","0.0","0.0","0.0","1.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179525161701249026","https://twitter.com/gwern/status/1179525161701249026","@TWLiterature Well, none of the vampires sleep so Rip presumably sleeps no more than anyone else, although I don't think the anime discusses what they all *do* do between WWII and attacking England - wait, was this a question about _Hellsing_ or behavioral genetics","2019-10-02 22:34 +0000","113.0","1.0","0.008849557522123894","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179515952112848897","https://twitter.com/gwern/status/1179515952112848897","@drderringer Sorry, the best I can offer on violence/aggression is that it's heritable, polygenic, and too much of the literature is candidate-genes or underpowered: https://t.co/23Izc1wHDI","2019-10-02 21:58 +0000","767.0","18.0","0.02346805736636245","1.0","2.0","4.0","1.0","9.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179431180954951680","https://twitter.com/gwern/status/1179431180954951680","@dwf So you criticize a proven solution for reducing 'body count' (what a remarkably cold dismissal of all those people alive now because of nets - just 'body counts'?) because it won't solve literally everything about life or education.","2019-10-02 16:21 +0000","898.0","27.0","0.030066815144766147","0.0","1.0","7.0","2.0","0.0","0.0","17.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179428536463085568","https://twitter.com/gwern/status/1179428536463085568","_Spatial Ability: Its Educational and Social Significance_, Smith 1964 https://t.co/o1999HTwNO https://t.co/SBCfYfgiVg https://t.co/B6YqL58L0T","2019-10-02 16:10 +0000","3650.0","20.0","0.005479452054794521","0.0","0.0","0.0","1.0","16.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179427292197658635","https://twitter.com/gwern/status/1179427292197658635","@dwf My understanding was that they actually are an 'actual solution', as many suicides are impulsive, and adding barriers to hotspots like bridges does not displace suicide but reduces them, and this was well known: https://t.co/WJRN3BCHkN https://t.co/MKBIYuEAxg","2019-10-02 16:05 +0000","2397.0","108.0","0.04505632040050062","0.0","2.0","16.0","10.0","13.0","0.0","67.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179425147494830081","https://twitter.com/gwern/status/1179425147494830081","@ttmariemia @zooko @ZcashFoundation @ElectricCoinCo No, you'd have the ZEC (which sit long-term in a static address/contract, increasing their vesting rights every day), and then tap into your ZEC *without moving them* by getting a loan from someone else to do any other stuff - possibly to buy unencumbered ZEC, even.","2019-10-02 15:57 +0000","173.0","1.0","0.005780346820809248","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179240615822925824","https://twitter.com/gwern/status/1179240615822925824","@climatebrad @GernotWagner Am I reading this paper correctly, that they assume zero effect of carbon taxes on economic growth and no opportunity cost?","2019-10-02 03:44 +0000","459.0","1.0","0.002178649237472767","0.0","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179187019706183680","https://twitter.com/gwern/status/1179187019706183680","@jschauma I don't think he said. My prejudices are that it probably was a *nix. (I can't imagine a Windows box running unattended for that long, not when dating from that era.)","2019-10-02 00:11 +0000","106.0","4.0","0.03773584905660377","0.0","1.0","0.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179186417404174337","https://twitter.com/gwern/status/1179186417404174337","@dbweissman So you don't actually know.","2019-10-02 00:08 +0000","101.0","6.0","0.0594059405940594","0.0","1.0","0.0","3.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179186178362413056","https://twitter.com/gwern/status/1179186178362413056","@mendel_random It is not a major problem, and it's incredibly obscure other than as a cliche example. I asked for real examples, not obscure diseases. And H pylori infections may be heritable, but >0.8 as I asked? I rather doubt that. These are the best you can do?","2019-10-02 00:07 +0000","414.0","10.0","0.024154589371980676","0.0","1.0","4.0","2.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179182785493393409","https://twitter.com/gwern/status/1179182785493393409","@andy_matuschak My notes on 'anti-spaced repetition': https://t.co/jD4SsY6VBW Also musings on how to optimize new vs rewatchs for movies: https://t.co/2sUIDUSaPK","2019-10-01 23:54 +0000","475.0","38.0","0.08","0.0","0.0","5.0","1.0","24.0","0.0","8.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179168256369840129","https://twitter.com/gwern/status/1179168256369840129","@devonzuegel I am very basic and always recommend the Noguchi Museum. I visit whenever I can when I'm in the area. (Over the years, I admire more and more Noguchi's efforts at doing all of sculpture and architecture and furniture and garden & interior design according to his vision.)","2019-10-01 22:56 +0000","2775.0","45.0","0.016216216216216217","1.0","0.0","25.0","7.0","0.0","0.0","12.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179136599839379458","https://twitter.com/gwern/status/1179136599839379458","@pjrvs HN suggestions: https://t.co/V1WD5RP2vn","2019-10-01 20:50 +0000","723.0","40.0","0.05532503457814661","0.0","0.0","1.0","0.0","39.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179136210293477376","https://twitter.com/gwern/status/1179136210293477376","@jamesjyu @nayafia @antonhowes https://t.co/toeoPEHxbs !","2019-10-01 20:49 +0000","338.0","9.0","0.026627218934911243","0.0","0.0","1.0","1.0","7.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179132885435060225","https://twitter.com/gwern/status/1179132885435060225","@ArtirKel @nayafia @antonhowes (You see this transition very starkly in _The Paris Review_ interviews when you go from the earliest 1910s-1920s authors to contemporary authors. More recent authors are flattered, confused, and grateful if anyone publishes their shorts at all. Certainly not cash cows.)","2019-10-01 20:36 +0000","444.0","3.0","0.006756756756756757","0.0","0.0","1.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179132533197418496","https://twitter.com/gwern/status/1179132533197418496","@ArtirKel @nayafia @antonhowes They did survive in SF for decades beyond that. SF authors would follow a path of writing shorts, getting a first novel, then publishing anthologies with backcatalogue. Now the SF short market is gone, and you get started with novels (ideally trilogies). ???","2019-10-01 20:34 +0000","281.0","1.0","0.0035587188612099642","0.0","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179132227969531904","https://twitter.com/gwern/status/1179132227969531904","@ArtirKel @nayafia @antonhowes Publishing economics are confusing. For example, short stories. They used to be *huge*. Not just Dickens but writers like Fitzgerald made their name (and literal fortunes) by selling short stories. Now they are notorious as the most guaranteed money-losers short of poetry.","2019-10-01 20:33 +0000","4185.0","32.0","0.007646356033452807","2.0","4.0","9.0","4.0","0.0","0.0","13.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179130713880383489","https://twitter.com/gwern/status/1179130713880383489","@nayafia @antonhowes Yes, it's interesting how that comes up all the time back then. Samuel Johnson's Dictionary, or special limited editions like Burton Watson's _1001 Arabian Nights_ (https://t.co/Xc4GyDijU3) (Note, incidentally, the rubrication/typographer's red in the title page samples.)","2019-10-01 20:27 +0000","847.0","3.0","0.0035419126328217238","0.0","1.0","0.0","0.0","2.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179114400235606026","https://twitter.com/gwern/status/1179114400235606026","@rictic @united I assume they are trying to implement DRM. There were 2 or 3 media entries which didn't require the app, and they were of course obscure TV things I'd never heard of. Everything you'd actually watch required the app.","2019-10-01 19:22 +0000","323.0","1.0","0.0030959752321981426","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179113991941046278","https://twitter.com/gwern/status/1179113991941046278","@mendel_random H. pylori is now heritability? You're trotting out PKU, really? And myopia doesn't even satisfy my question in the first place. All I asked was 3 examples of things like childhood obesity which are highly heritable and also solved by policy. Why is it so difficult to name 3?","2019-10-01 19:20 +0000","574.0","22.0","0.03832752613240418","1.0","1.0","7.0","1.0","0.0","0.0","12.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179113438875996160","https://twitter.com/gwern/status/1179113438875996160","@jschauma We asked him how it was on the Internet at all (he never said WWW, just Internet), and he didn't know. Strange, but it's a big world, and there are lots of examples with somewhat smaller ages, so...","2019-10-01 19:18 +0000","94.0","1.0","0.010638297872340425","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179112961593544704","https://twitter.com/gwern/status/1179112961593544704","@dbweissman So your first example is not an example (fatal infectious disease heritabilities are usually <0.8 for obvious reasons on both sides), and your other two examples are very rare monogenic diseases. Do you think that supports the claims being made for things like childhood obesity?","2019-10-01 19:16 +0000","136.0","1.0","0.007352941176470588","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179112430565302279","https://twitter.com/gwern/status/1179112430565302279","@JProtzko @rubenarslan @dbweissman @drewhalbailey The average long-run net effect of that seems to've been fairly small. (The randomized experiments you might be thinking of have outliers and were selected for heavily deficient regions.) And that was a century ago in the US now.","2019-10-01 19:14 +0000","181.0","4.0","0.022099447513812154","0.0","0.0","0.0","4.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179111740631584768","https://twitter.com/gwern/status/1179111740631584768","@dbweissman Longevity comes to mind. If we were fighting some sort of life-history tradeoff or programmed aging strategy, it'd be doomed. But it seems to be largely just noise and not particularly influenced by either shared environment (hinting at lifestyle) or genes. So, senolytics etc.","2019-10-01 19:12 +0000","76.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179104543814098946","https://twitter.com/gwern/status/1179104543814098946","Another good one: @united Airlines has removed TVs from chairs. Makes sense, saves weight, tablets/phone screens as good/better. You stream movies over a captive WiFi, of course. ...but you need to install a United 'app'. ...and you can't install it over their WiFi. Whups.","2019-10-01 18:43 +0000","6038.0","66.0","0.01093077177873468","1.0","1.0","18.0","3.0","0.0","0.0","43.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179103476791558144","https://twitter.com/gwern/status/1179103476791558144","@jschauma I was recently told by someone that scanning the Internet, they'd found a server in Pakistan that had been up for ~30 years. Apparently it turned out it had been drywalled-in and never turned off, so... Solid-state machines can be remarkably reliable.","2019-10-01 18:39 +0000","91.0","1.0","0.01098901098901099","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179088158098673665","https://twitter.com/gwern/status/1179088158098673665","@dbweissman @rubenarslan I wouldn't. What policies solved height and what height crisis was there that those were trying to solve, exactly? So it's not even 1 example.","2019-10-01 17:38 +0000","192.0","2.0","0.010416666666666666","0.0","2.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179087846113759232","https://twitter.com/gwern/status/1179087846113759232","@dbweissman @rubenarslan @mendel_random And they are equivocating on what 'environment' means. In the strict technical term, yes, perhaps 'environment' did increase childhood obesity. It does not then follow that *any* policy interventions are affecting the relevant 'environment'.","2019-10-01 17:37 +0000","83.0","7.0","0.08433734939759036","1.0","0.0","2.0","0.0","0.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179087246546407424","https://twitter.com/gwern/status/1179087246546407424","@rubenarslan @dbweissman Let's try it this way. If high heritability is irrelevant, and 'environment' so easily modified, then it should be easy for you to name 3 examples of traits with high heritabilities (say >=0.8, like height or intelligence), which have been solved by standard policies. Can you?","2019-10-01 17:34 +0000","418.0","12.0","0.028708133971291867","0.0","3.0","2.0","1.0","0.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179086702960353281","https://twitter.com/gwern/status/1179086702960353281","@dbweissman @rubenarslan @mendel_random There obviously isn't any working long-term solution, or they would have simply pointed it out yet, and we wouldn't have childhood obesity as a problem now. Which is why it's such a telling example against these speculations.","2019-10-01 17:32 +0000","101.0","2.0","0.019801980198019802","0.0","1.0","0.0","0.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179084702898118658","https://twitter.com/gwern/status/1179084702898118658","@dbweissman @rubenarslan It's the same endless equivocation on (nonshared) environment as being affected by hypothetical interventions when convenient, while all actual policy interventions which are feasible or implemented fail on average. eg developmental noise doesn't mean vending machine taxes work.","2019-10-01 17:24 +0000","146.0","5.0","0.03424657534246575","0.0","1.0","1.0","0.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179084287582314496","https://twitter.com/gwern/status/1179084287582314496","@dbweissman @rubenarslan Not in the least. Look at https://t.co/oJVAHUEOfw It's the same motte-and-bailey as ever. 'Policy interventions fail, especially with high heritability.' 'Ah, but heritability doesn't mean immutability - we could *imagine* interventions which do work! Check and mate.'","2019-10-01 17:22 +0000","505.0","15.0","0.0297029702970297","1.0","3.0","6.0","2.0","1.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179080282789552132","https://twitter.com/gwern/status/1179080282789552132","@rubenarslan @mendel_random ""Policy interventions, comrades, have never truly been tried.""","2019-10-01 17:07 +0000","157.0","5.0","0.03184713375796178","0.0","1.0","2.0","1.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179072063161733122","https://twitter.com/gwern/status/1179072063161733122","@fluk @5auth Remember that the SR1 forums were left up deliberately.","2019-10-01 16:34 +0000","66.0","1.0","0.015151515151515152","0.0","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1179072003573256194","https://twitter.com/gwern/status/1179072003573256194","@fluk @5auth Some people will be! You also can learn a lot from PMs and contacts, not to mention metadata like timing or shared PGP keys. You can use it for censorship, emphasizing FUD, or tracking people long-term. Lots of things you can do if you run a DNM forum.","2019-10-01 16:34 +0000","187.0","8.0","0.0427807486631016","0.0","2.0","2.0","4.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178828772147159041","https://twitter.com/gwern/status/1178828772147159041","@5auth All of that is consistent with him now being flipped, of course. And only the last one indicates it's him at all rather than someone pwning his computers (charitably assuming he didn't write it down).","2019-10-01 00:27 +0000","4568.0","33.0","0.007224168126094571","1.0","1.0","9.0","11.0","0.0","0.0","11.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178814959863685120","https://twitter.com/gwern/status/1178814959863685120","@celestelynpaul Oh sure, I've known about Waifu Labs for months. They use StyleGAN and Danbooru2018, even. They improve over my TWDNE by larger images and a slick web interface to editing, whereas I just punt to snowy halcy's Python scripts if anyone wants to do editing.","2019-09-30 23:32 +0000","129.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178814518845214723","https://twitter.com/gwern/status/1178814518845214723","@rubenarslan @mendel_random The original claim is about policy proposals, which typically propose policies which look like previously tested interventions, which have net effects as implemented in the real world of ~0 (cf the Metallic Laws), and do not look like 'let's re-enact Mao's Great Famine'.","2019-09-30 23:30 +0000","707.0","19.0","0.026874115983026876","1.0","2.0","8.0","1.0","0.0","0.0","7.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178752214321582086","https://twitter.com/gwern/status/1178752214321582086","@rubenarslan @mendel_random And that's part of the environment too. 'The environment is genetic', remember? If people don't want to go to the genocide-park, how are you going to stop them, as Yogi Berra might've said.","2019-09-30 19:23 +0000","745.0","7.0","0.009395973154362415","1.0","1.0","3.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178751293676019714","https://twitter.com/gwern/status/1178751293676019714","@slimdipping Yes. Predictit is already down to 30% non-impeachment, and skimming, I suspect it's still underestimating. Oh well. I figured he'd beat the Mueller report & learn a lesson, but I guess he couldn't help himself.","2019-09-30 19:19 +0000","138.0","2.0","0.014492753623188406","0.0","1.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178750492782006273","https://twitter.com/gwern/status/1178750492782006273","@antonhowes Yeah, it would've been fun if it was a real-life DAC, but if they only cover losses, then it doesn't have the right game-theory properties to solve public goods provision. Might be a good question what that sort of guarantee fund *does* solve under what circumstances?","2019-09-30 19:16 +0000","268.0","1.0","0.0037313432835820895","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178747310928465920","https://twitter.com/gwern/status/1178747310928465920","@KirkegaardEmil @NickWolfinger I don't recall any offhand, but I would be surprised if no one had done any by now.","2019-09-30 19:03 +0000","300.0","3.0","0.01","0.0","0.0","0.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178746647985168388","https://twitter.com/gwern/status/1178746647985168388","@rubenarslan @mendel_random In what sense is that an 'effective intervention'? We still have it, do we not? If you claim something is easily modified, people want to modify it, and it's not modified, then you are the one redefining words.","2019-09-30 19:01 +0000","1046.0","13.0","0.0124282982791587","2.0","1.0","7.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178746389175689217","https://twitter.com/gwern/status/1178746389175689217","@samsaragon No. Expensive and difficult to blind.","2019-09-30 19:00 +0000","41.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178746121067339776","https://twitter.com/gwern/status/1178746121067339776","@hentaikid I see it as the other end of the spectrum, similar to Charles Bonnet syndrome or how many people report low-level visual hallucinations as regular things. (I'm not familiar with it, but it sounds like a story one doesn't need to know.)","2019-09-30 18:59 +0000","143.0","1.0","0.006993006993006993","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178745591427485696","https://twitter.com/gwern/status/1178745591427485696","@tambourine Have you considered using Google Colab? The free GPUs work quite well for finetuning and people already have lots of notebooks written to do that sort of thing.","2019-09-30 18:57 +0000","72.0","4.0","0.05555555555555555","0.0","1.0","0.0","0.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178745157157556231","https://twitter.com/gwern/status/1178745157157556231","@nathanbraun @michael_nielsen I use clean native-digital PDFs or EPUBs when they're available online or on Libgen, of course, no need to reinvent the wheel; but for countless books, a scan+OCR is all that will ever be possible. I describe how I do it here: https://t.co/4ggIV2HClK","2019-09-30 18:55 +0000","196.0","5.0","0.025510204081632654","0.0","0.0","1.0","0.0","2.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178744754995154949","https://twitter.com/gwern/status/1178744754995154949","@JoshPritsker Yes... Between old people and the lizardman constant, I don't know if one would find those all *that* surprising.","2019-09-30 18:53 +0000","93.0","1.0","0.010752688172043012","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178744397707583490","https://twitter.com/gwern/status/1178744397707583490","@ImperiumRequiem @KirkegaardEmil @woke8yearold 'TBC'? As far as genome synthesis goes, I believe https://t.co/O2BfqMW3Ob represents roughly SOTA.","2019-09-30 18:52 +0000","130.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178739558197972999","https://twitter.com/gwern/status/1178739558197972999","@antonhowes I see. It sounded like it was an award for the method from the wording, but still an interesting example of the RSA popping up. Incidentally, the Kickstarter-like thing I mentioned: https://t.co/qKlBgNn11N","2019-09-30 18:33 +0000","205.0","3.0","0.014634146341463415","0.0","1.0","1.0","0.0","1.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178738966851473409","https://twitter.com/gwern/status/1178738966851473409","- how often do mouse studies replicate or then transfer to humans? https://t.co/nLS1g8ngIF - unusual Internet subcultures related to aphantasia: creating 'tulpas' https://t.co/djFcvEC5tW https://t.co/dt2TJr8fLY https://t.co/P5shvVYuzs","2019-09-30 18:30 +0000","4669.0","95.0","0.020346969372456628","0.0","1.0","9.0","0.0","85.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178738965345689601","https://twitter.com/gwern/status/1178738965345689601","- Interesting offbeat datasets: the OKCupid scrape, and the Pony Voice Preservation Project https://t.co/nAuEUPL9To - Essence of hacking/security: 'seeing through': https://t.co/ELq1EKVn7V - OSS funding: quadratic voting https://t.co/ODRpRjbhAf eg Gitcoin https://t.co/HKw0HzS2W3","2019-09-30 18:30 +0000","5280.0","49.0","0.009280303030303031","0.0","1.0","3.0","4.0","35.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178737049320468481","https://twitter.com/gwern/status/1178737049320468481","Travel observation: reviews are a tragedy of the anti-commons. I have 7 emails nagging me for reviews from the airlines/hotels etc (not counting feedback requests from services en route like airport WiFi) - which repeat! I can answer 1, but 7 makes one want to give up entirely.","2019-09-30 18:23 +0000","11450.0","69.0","0.0060262008733624455","0.0","2.0","30.0","22.0","0.0","0.0","15.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178546995717296128","https://twitter.com/gwern/status/1178546995717296128","- optimal review of movies or notes: https://t.co/2sUIDV9LHi https://t.co/jD4SsYowtu - Notes towards dissolving Pascal's mugging with dynamic programming & high-mem AWS instances: https://t.co/YA7Ne0uJ8M","2019-09-30 05:47 +0000","4936.0","56.0","0.011345218800648298","1.0","1.0","4.0","4.0","44.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178546994685472768","https://twitter.com/gwern/status/1178546994685472768","- Do writers prefer mornings? https://t.co/cCOh7IlYZn - Why aren't empirical fields like CS/AI using inadequate methods as irreplicable as social psych or candidate-genes? Perhaps their evolutionary grounding: https://t.co/LvIL2eU1n3","2019-09-30 05:47 +0000","4951.0","87.0","0.017572207634821247","1.0","1.0","7.0","8.0","67.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178546993586556928","https://twitter.com/gwern/status/1178546993586556928","- Aphantasia et al: https://t.co/4CWjTkI0Ji https://t.co/RRuzs9A2Xg https://t.co/56xgvEU1Ez https://t.co/9TxXTHRU5J - lrn2google: https://t.co/hnBd6b5wja - CO2? https://t.co/R7HFCbCrXW https://t.co/ufeoS7E7sh - Human steganography: https://t.co/Po4BFLl87S https://t.co/GKTDtUfMul","2019-09-30 05:47 +0000","2897.0","150.0","0.05177770107007249","0.0","1.0","1.0","2.0","143.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178546992600891399","https://twitter.com/gwern/status/1178546992600891399","Links for my conversees this weekend: - Ads' 10% user penalty: https://t.co/bIFtVR6Jfr - Anime NNs: https://t.co/3nM685Ft8n https://t.co/GRWAA2tqof https://t.co/Ua5Uq4XP9b https://t.co/sgxDdGMQEj - RL-based poetry generation: https://t.co/FzOQkE4g1G https://t.co/gk5E5vTLJY","2019-09-30 05:47 +0000","10202.0","164.0","0.016075279356988825","2.0","1.0","21.0","13.0","104.0","0.0","23.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178505094943297537","https://twitter.com/gwern/status/1178505094943297537","@ArtirKel 'How much can the microbiome matter' is interestingly analysed by https://t.co/8KK6Xhk5fA (one of the examples of the benefits of a variance component analysis approach I mention in https://t.co/1mIXMoFVpr )","2019-09-30 03:01 +0000","633.0","23.0","0.036334913112164295","0.0","0.0","6.0","3.0","11.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1178504687357550592","https://twitter.com/gwern/status/1178504687357550592","@antonhowes Modern bread manufacturing, using a faster more hygienic chemical process, was advertised as 'alcohol-free' after being invented & winning a RSA silver medal: https://t.co/YfF9VJzNm6 (Found following a German improvement mention in https://t.co/7NPkLUTA2x )","2019-09-30 02:59 +0000","751.0","15.0","0.019973368841544607","0.0","1.0","3.0","3.0","6.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1177322401098620928","https://twitter.com/gwern/status/1177322401098620928","@ArtirKel Oh yeah. Every methodological problem you can think of with human research in psych/economics/medicine? It's worse in animals. The high-powered invasive stuff you can do with animals (like killing them) only partially compensates.","2019-09-26 20:41 +0000","2817.0","36.0","0.012779552715654952","1.0","0.0","9.0","8.0","0.0","0.0","18.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1177318366454919169","https://twitter.com/gwern/status/1177318366454919169","@michael_nielsen Yes, that's the downside... I scan most of my books after reading, so the only books I have fall into the categories of (1) haven't read, and (2) too precious to give away. (If anyone thinks I'm hauling along one of my _Absolute Sandman_ volumes to give away, they're crazy!)","2019-09-26 20:25 +0000","1488.0","15.0","0.010080645161290322","0.0","1.0","2.0","6.0","0.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1177316513407295490","https://twitter.com/gwern/status/1177316513407295490","@ArtirKel https://t.co/nLS1g8ngIF","2019-09-26 20:18 +0000","2354.0","167.0","0.07094307561597281","1.0","2.0","11.0","12.0","121.0","0.0","20.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1177030019580944384","https://twitter.com/gwern/status/1177030019580944384","@tambourine Yes, admittedly a finetuned model by design loses the original multimodality of GPT-2 (unless your finetuning also incorporated multimodal text to maintain those capabilities, like how Grover or CTRL includes metadata for source). Could switch between multiple models as needed.","2019-09-26 01:19 +0000","160.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1177018574898192384","https://twitter.com/gwern/status/1177018574898192384","@tambourine @alexhern Have you tried using my GPT-2 models finetuned on old poetry? https://t.co/gk5E5wbn8y Or is the idea here to use the baseline TalkToTransformer model to get a more prose-y look?","2019-09-26 00:34 +0000","176.0","10.0","0.056818181818181816","0.0","1.0","1.0","2.0","1.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1177016925777473537","https://twitter.com/gwern/status/1177016925777473537","@AndreTI The new 'learning preferences' version of GPT-2 might be a good tool for that. You could coax it towards rhyming more, or if you have a rhyming tool which can automatically measure number of rhymes in a text, auto-generate comparisons to force rhymes.","2019-09-26 00:27 +0000","715.0","4.0","0.005594405594405594","0.0","1.0","2.0","1.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1177009659389124608","https://twitter.com/gwern/status/1177009659389124608","@arjendre @epkaufm @TOOEdit Preprint, though, so harder to read. Mirror of the published version: https://t.co/f1YmWfuHNq","2019-09-25 23:59 +0000","648.0","18.0","0.027777777777777776","1.0","0.0","3.0","2.0","12.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176959183977144320","https://twitter.com/gwern/status/1176959183977144320","Vaguely related: XKCD asks what are the maximally uncontroversial & most popular polled questions? https://t.co/tqi42tGfYF","2019-09-25 20:38 +0000","6494.0","267.0","0.04111487526947952","0.0","0.0","13.0","2.0","241.0","0.0","11.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176618854136635392","https://twitter.com/gwern/status/1176618854136635392","@alexeyguzey It always shellshocks me to remember that one of my favorite games back then, like _Streets of Rage_ for Sega Genesis, would have cost somewhere ~$100 in 2019 dollars. And then used games were only like half the price! Compare this to the Steam sales...","2019-09-24 22:06 +0000","1285.0","24.0","0.01867704280155642","1.0","0.0","12.0","3.0","0.0","0.0","8.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176616410430279680","https://twitter.com/gwern/status/1176616410430279680","@madebymaze That'd be nice. I assume you just link to a URL which does the equivalent of 'git log https://t.co/ekrpHRyKLw'? (Unfortunately, I found Github to be too restrictive a host to keep my repo there, and don't want to set up my own Git repo hosting.)","2019-09-24 21:56 +0000","168.0","5.0","0.02976190476190476","0.0","1.0","0.0","0.0","2.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176615293130612747","https://twitter.com/gwern/status/1176615293130612747","@danielwithmusic Oh. No. (I mean, unless people think it looks better? Then it's not a bug it's a feature.)","2019-09-24 21:52 +0000","200.0","4.0","0.02","0.0","1.0","0.0","0.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176615041904369677","https://twitter.com/gwern/status/1176615041904369677","@dbabbitt I would have to see it on the page, but I think it's usually darker?","2019-09-24 21:51 +0000","135.0","2.0","0.014814814814814815","0.0","0.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176614807711223815","https://twitter.com/gwern/status/1176614807711223815","@rubenarslan @mendel_random If you say 'akshully high heritability is consistent with modifiability!' and your example is not, in fact, modifiable in practice, you should consider whether ""one man's modus ponens is another man's modus tollens"" (https://t.co/nWDxH1G9L1)...","2019-09-24 21:50 +0000","1300.0","41.0","0.031538461538461536","2.0","1.0","11.0","2.0","17.0","0.0","8.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176614390700892160","https://twitter.com/gwern/status/1176614390700892160","@rubenarslan @mendel_random If we can't point to interventions which do modify things of high heritability, then it seems a fortiori true that not only is the specific case false because public schools do not fix anything they try to fix (such as, say, childhood obesity), we should expect that in general.","2019-09-24 21:48 +0000","1097.0","22.0","0.020054694621695533","2.0","2.0","7.0","8.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176614105102372866","https://twitter.com/gwern/status/1176614105102372866","@mghwaz I believe they all get inverted, since it's just a CSS filter.","2019-09-24 21:47 +0000","107.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176510979015958529","https://twitter.com/gwern/status/1176510979015958529","@evansd66 @Richie_Research Using '¨¦lan vital' also has the benefit of connotations of sarcasm and pointing out that '¨¦lan vital' has been debunked and reminding the reader of the ever-present cognitive biases towards anthropomorphism and against reductionism & materialism.","2019-09-24 14:57 +0000","131.0","5.0","0.03816793893129771","0.0","0.0","2.0","0.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176505193103577089","https://twitter.com/gwern/status/1176505193103577089","@mendel_random How are those childhood obesity interventions going? Have we solved childhood obesity yet? If not, maybe you need a different example here...","2019-09-24 14:34 +0000","2003.0","80.0","0.0399400898652022","3.0","3.0","21.0","18.0","0.0","0.0","35.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176503275505561600","https://twitter.com/gwern/status/1176503275505561600","@AndrewCutler13 The problem with her is that we don't know how much work or custom tools or hand-tweaking goes into her, at least in any of the articles I've read. There have been photorealistic CGI or paintings for many decades. Is she '200ms GAN forward pass' or 'SFX team working for a week'?","2019-09-24 14:26 +0000","1420.0","18.0","0.01267605633802817","0.0","2.0","11.0","3.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176319224886833152","https://twitter.com/gwern/status/1176319224886833152","@gitcoin @BlackGirlsCode Are the green squares' total area supposed to reflect the current allocation via the quadratic, or indicate the marginal change from a $1 contrib? It's most easily explained geometrically, I had the impression, so a UI might want to use that.","2019-09-24 02:15 +0000","2261.0","3.0","0.0013268465280849183","0.0","0.0","2.0","1.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176318030739136512","https://twitter.com/gwern/status/1176318030739136512","So my phone upgraded to Android 10 and I can finally see native browser dark mode for https://t.co/LC5JQL86wv. I didn't realize just how much of a difference a higher-DPI display made, especially for the fonts, but it looks razor-sharp and simply beautiful. ?","2019-09-24 02:10 +0000","7057.0","76.0","0.010769448774266686","1.0","3.0","8.0","9.0","44.0","0.0","11.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176316462790782977","https://twitter.com/gwern/status/1176316462790782977","@digitalviscera 'performative'?","2019-09-24 02:04 +0000","107.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176290697810460673","https://twitter.com/gwern/status/1176290697810460673","@theshawwn I dunno. Hadn't even thought about code samples. For those, I usually break them up or rewrite them. (That's less of an option for tables or images.)","2019-09-24 00:22 +0000","211.0","3.0","0.014218009478672985","0.0","0.0","1.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176283711085731842","https://twitter.com/gwern/status/1176283711085731842","Borrowed another feature from Tufte-CSS (https://t.co/v7iicCC3cc): full-width images, like full-width tables. Particularly with time-series, they need as much horizontal space as possible. https://t.co/0h0oaXtUm4","2019-09-23 23:54 +0000","5680.0","133.0","0.02341549295774648","0.0","1.0","14.0","4.0","18.0","0.0","10.0","0.0","0","0","0","0","0","86","86","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176282689806229505","https://twitter.com/gwern/status/1176282689806229505","Nifty Japanese example: https://t.co/jAwpmZU2Hp From 1920s collection: https://t.co/lTrMqBVrJW Like Butterick on Stempel catalogues: https://t.co/7I88jcN31j 'Just this page. You could set up a whole de?sign prac?tice to?day rip?ping off this one page. You¡¯d be fine for 20 years.'","2019-09-23 23:50 +0000","9815.0","107.0","0.010901681100356598","3.0","3.0","16.0","4.0","68.0","0.0","13.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176262902409584645","https://twitter.com/gwern/status/1176262902409584645","@TheClarksTale @Scholars_Stage I'm convinced the movies made the difference. Who would read _The Wizard of Oz_ today if not for the movie? HP has a sequence of 8 triple-A Hollywood blockbuster movies which can grow up with kids, as part of an ongoing franchise. (Have you see the HP World crowds?)","2019-09-23 22:31 +0000","195.0","2.0","0.010256410256410256","0.0","0.0","1.0","1.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176196055202971649","https://twitter.com/gwern/status/1176196055202971649","@robinhanson re #1: https://t.co/d2lqImToJD has you covered. (And is available in the usual places.)","2019-09-23 18:06 +0000","2018.0","115.0","0.05698711595639247","0.0","1.0","17.0","2.0","91.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176172809481007106","https://twitter.com/gwern/status/1176172809481007106","@anderssandberg Incidentally, I found that particular archive using the IA site-search operator (`$domain.com/*`), which is covered in my Internet search guide: https://t.co/hnBd6aNVrC (After checking Nordhaus's sites & Google & the journal & the IA of the URL failed, fellback to domain search.)","2019-09-23 16:33 +0000","300.0","3.0","0.01","0.0","0.0","1.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176162477635186690","https://twitter.com/gwern/status/1176162477635186690","@anderssandberg At this point, depositing it *anywhere* but personal websites is an improvement. The IA misses a lot of stuff, but not for fundamental reasons (no reason it couldn't've picked up the original URL, it just didn't). Souping up the IA would go a long way.","2019-09-23 15:52 +0000","302.0","4.0","0.013245033112582781","0.0","0.0","2.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176160894184828930","https://twitter.com/gwern/status/1176160894184828930","@dileeplearning Yes... except none of that did anything. I read about that DARPA project not long ago ( https://t.co/XV7Rwl1sWX ) and was impressed how much more NNs do now than then. eg the CMU self-driving car couldn't drive across empty asphalt parking lots w/o being confused by sunlight.","2019-09-23 15:46 +0000","164.0","5.0","0.03048780487804878","0.0","0.0","1.0","2.0","2.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176153521198157825","https://twitter.com/gwern/status/1176153521198157825","@anderssandberg (It is dismaying looking at #nordhaus's website in IA and seeing just how many spreadsheets, papers, and other URLs he let die. Thank goodness for the IA.)","2019-09-23 15:17 +0000","278.0","4.0","0.014388489208633094","0.0","1.0","1.0","2.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176152031003926529","https://twitter.com/gwern/status/1176152031003926529","@anderssandberg Made a mirror at https://t.co/XLDe3PrIyD","2019-09-23 15:11 +0000","434.0","7.0","0.016129032258064516","0.0","2.0","1.0","0.0","4.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1176145721596219393","https://twitter.com/gwern/status/1176145721596219393","@anderssandberg https://t.co/cRuYtlw8Zf is the closest I can find.","2019-09-23 14:46 +0000","1011.0","7.0","0.006923837784371909","0.0","2.0","1.0","0.0","4.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175886843268816903","https://twitter.com/gwern/status/1175886843268816903","@glenweyl @nielsrosenquist Georgism is definitely very interesting, but I'm not sure he'd really be in the mainstream of economics either then or now (eg he was an autodidact, who never even went to college), or that Chicago could be blamed for a lack of Georges. Are there any other good examples?","2019-09-22 21:37 +0000","678.0","9.0","0.01327433628318584","0.0","1.0","0.0","1.0","0.0","0.0","7.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175873684839550977","https://twitter.com/gwern/status/1175873684839550977","@nielsrosenquist @glenweyl Blaming it on any particular school implies things were better pre-school. What exciting or revolutionary uses of IT-like tech or thinking in economics do you associate with pre-Chicago-school economists, that the Chicago school killed?","2019-09-22 20:45 +0000","396.0","5.0","0.012626262626262626","0.0","1.0","0.0","2.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175811145728516096","https://twitter.com/gwern/status/1175811145728516096","@carlzimmer If we wildly exaggerate and compare some gardening research prototypes not working to tens of thousands of people receiving falsified medical test results, then yeah, sure, I can see how it sounds a bit like Theranos.","2019-09-22 16:36 +0000","1203.0","12.0","0.00997506234413965","0.0","0.0","5.0","5.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175785726904078337","https://twitter.com/gwern/status/1175785726904078337","@robinhanson https://t.co/zsvs4LxxpT Vizio seems to think it's a decent chunk of money over the lifetime of a TV and its multiple watchers, especially in comparison to existing margins (6%, so $60 of your example).","2019-09-22 14:55 +0000","1633.0","34.0","0.020820575627679118","0.0","1.0","9.0","2.0","10.0","0.0","12.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175532737953980420","https://twitter.com/gwern/status/1175532737953980420","@qntm Sounds repetitive.","2019-09-21 22:10 +0000","311.0","5.0","0.01607717041800643","0.0","0.0","1.0","1.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175495812941516805","https://twitter.com/gwern/status/1175495812941516805","@SilverVVulpes Don't be silly. We know the IAT must be false because no physical laws supervene on the operation of the mind! ?","2019-09-21 19:43 +0000","714.0","9.0","0.012605042016806723","0.0","0.0","7.0","1.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175438595378831360","https://twitter.com/gwern/status/1175438595378831360","@snoreydude @jonathanstray @RokoMijicUK That graph shows little net change over 'the last two decades', doesn't take into account hedonics or transfers or any kind of progress ('income'), and as I already pointed out, 'households' are not even a constant unit in the first place.","2019-09-21 15:56 +0000","196.0","8.0","0.04081632653061224","0.0","2.0","0.0","0.0","0.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175419583467073536","https://twitter.com/gwern/status/1175419583467073536","@EllieAsksWhy While you're there, maybe see if the ALE bug recently found works on an arcade machine: https://t.co/iPs0MrTX1Z","2019-09-21 14:40 +0000","672.0","11.0","0.01636904761904762","0.0","1.0","1.0","1.0","8.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175418722615472133","https://twitter.com/gwern/status/1175418722615472133","@Ada_Palmer Plus to permanently store specific points in text files, 'bookmarks': https://t.co/LFJPRxNohR Another thing I do is have Emacs automatically syntax highlight in bright red some keywords like 'TODO', to make them hard to miss while reviewing.","2019-09-21 14:37 +0000","362.0","2.0","0.0055248618784530384","0.0","0.0","0.0","2.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175418394486738944","https://twitter.com/gwern/status/1175418394486738944","@Ada_Palmer In Emacs, you can 'set the point' and then pop it to return. So you are writing, realize something, hit `C-space C-space`, go off to wherever, and when you're done, hit `C-u C-space` to pop back, recursively: https://t.co/4EtSQs7ETe & storable: https://t.co/8o5Q7SffBi","2019-09-21 14:35 +0000","961.0","3.0","0.003121748178980229","0.0","1.0","2.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175417415041245192","https://twitter.com/gwern/status/1175417415041245192","@SQLGene That's very close to the linguistics idea of a 'graded reader': which attempts to sort texts by used words to increase vocab as slowly as possible. https://t.co/hQkIqEwI2K https://t.co/SiXjnmj7IG https://t.co/Ty3qgMyYcb","2019-09-21 14:32 +0000","160.0","6.0","0.0375","0.0","1.0","1.0","0.0","4.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175243152166477824","https://twitter.com/gwern/status/1175243152166477824","@hbarovertwo So, tags? Lots of work on multi-label classification with NNs.","2019-09-21 02:59 +0000","136.0","1.0","0.007352941176470588","0.0","0.0","0.0","0.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175160907128872961","https://twitter.com/gwern/status/1175160907128872961","@Gary_Lewis1 @psforscher That, however, didn't stop people from continuing to run loads of candidate-gene studies long after and looking under the lamp post, so to speak. The question is, what institutional or economic incentives or other factors made it *actually* happen, in the real world?","2019-09-20 21:32 +0000","184.0","10.0","0.05434782608695652","0.0","0.0","1.0","1.0","0.0","0.0","8.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175160366545952768","https://twitter.com/gwern/status/1175160366545952768","@seanpmackinnon Ah. Well, glad to help.","2019-09-20 21:30 +0000","130.0","3.0","0.023076923076923078","0.0","0.0","0.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175160098328629248","https://twitter.com/gwern/status/1175160098328629248","@pulkomandy @shauna_gm @sterlingcrispin Indeed. And so... As I said. MIT is 'compatible' with proprietary fully-closed licenses. So 'compatibility' is a very poor criterion.","2019-09-20 21:29 +0000","168.0","2.0","0.011904761904761904","0.0","0.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175159842073468929","https://twitter.com/gwern/status/1175159842073468929","@hbarovertwo (Danbooru2018 is a 365k-label tagging problem.)","2019-09-20 21:28 +0000","134.0","1.0","0.007462686567164179","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175100525198610432","https://twitter.com/gwern/status/1175100525198610432","@pulkomandy @shauna_gm @sterlingcrispin That's not how the MIT license works at all. It's not even per-file, to begin with, and including a copy of the license itself doesn't mean all derivative works are also MIT licensed. That's the entire point of the *GPL* - enforcing that because MIT and BSD don't.","2019-09-20 17:32 +0000","221.0","3.0","0.013574660633484163","0.0","1.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175087878474469378","https://twitter.com/gwern/status/1175087878474469378","@pulkomandy @shauna_gm @sterlingcrispin No, it only requires that a license *notice* be included. The code can be modified and all the new derived versions have additional copyrights, which can be proprietary. That's what it means to not be viral. Good luck 'extracting' some proprietary code long ago based on MIT code.","2019-09-20 16:42 +0000","238.0","3.0","0.012605042016806723","0.0","1.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175087320170733573","https://twitter.com/gwern/status/1175087320170733573","@yoavgo @sebastien_wood @catherineols Yes, it is different: an assembly line is way more expensive to shut down, and harder to shift workers to other tasks, than ML annotation. I'm not sure that's the point you wanted to make, though.","2019-09-20 16:40 +0000","727.0","6.0","0.008253094910591471","0.0","2.0","1.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175076807286165506","https://twitter.com/gwern/status/1175076807286165506","@pulkomandy @shauna_gm @sterlingcrispin I'm not sure that argument shows what you think it does. MIT licensed code is 'compatible' even with fully proprietary patent-encumbered code (because you can relicense & close MIT stuff). Does that mean fully proprietary patent-encumbered code is ""better"" than open source code?","2019-09-20 15:58 +0000","467.0","4.0","0.008565310492505354","0.0","1.0","0.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175068514157305856","https://twitter.com/gwern/status/1175068514157305856","@rjhaier I've always wanted to read an insider account of how things like SSGAC and UKBB happen. From the outside, it's mysterious how things go right for once.","2019-09-20 15:25 +0000","704.0","7.0","0.009943181818181818","0.0","0.0","5.0","1.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175066709880975361","https://twitter.com/gwern/status/1175066709880975361","@yoavgo @sebastien_wood @catherineols That would depend on how they are being paid. And did Toyota start paying its workers *less* when it became *more* successful due to things like the andon...?","2019-09-20 15:18 +0000","576.0","6.0","0.010416666666666666","0.0","1.0","0.0","0.0","0.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1175039032834822145","https://twitter.com/gwern/status/1175039032834822145","@yoavgo @sebastien_wood @catherineols They don't have to read increasingly disturbing machine-generated smut.","2019-09-20 13:28 +0000","657.0","3.0","0.0045662100456621","0.0","1.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174886154564460545","https://twitter.com/gwern/status/1174886154564460545","@seanpmackinnon Er, why do you have to change your monitor resolution for any reason? What does that do that the usual R method of saving images with specific resolution/size like 'png(file=""foo.png"", width = 580, height = 600); foo; https://t.co/zG7YGjBKNr()' doesn't do?","2019-09-20 03:21 +0000","817.0","3.0","0.0036719706242350062","0.0","1.0","1.0","1.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174878205330755589","https://twitter.com/gwern/status/1174878205330755589","@sebastien_wood @catherineols In this context, like an assembly line, it'd just stop. That would avoid wasting ratings until the researchers have a chance to sanity-check everything. In other contexts, perhaps there would be a fallback to a baseline model or reverting to an older model.","2019-09-20 02:49 +0000","799.0","6.0","0.007509386733416771","0.0","1.0","3.0","2.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174874525084868608","https://twitter.com/gwern/status/1174874525084868608","@sebastien_wood @catherineols Literally perversely correct behavior. But no, an 'Andon cord' lets any factory worker stop the whole assembly line. So here it would be a big red button any of the https://t.co/thrJqXR55t workers could push to stop training/rating if it starts looking really weird/bad.","2019-09-20 02:34 +0000","670.0","9.0","0.013432835820895522","0.0","1.0","1.0","2.0","3.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174847976335925248","https://twitter.com/gwern/status/1174847976335925248","@the8472 His summary is that there doesn't seem to be any way to lay it out correctly but to insert stuff and check, and batching wouldn't be doable in any obvious way he knows of.","2019-09-20 00:49 +0000","161.0","1.0","0.006211180124223602","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174845112016429057","https://twitter.com/gwern/status/1174845112016429057","@the8472 I'm afraid I'm almost totally ignorant of JS, especially layout/DOM stuff, but I'll forward your comments to Obormot and maybe he'll have some ideas on optimizing it.","2019-09-20 00:37 +0000","172.0","1.0","0.005813953488372093","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174844465527361537","https://twitter.com/gwern/status/1174844465527361537","@hbarovertwo I don't read a ton about unsupervised learning because it seems like something of a minefield and at least for Danbooru2018, it'd be semi-supervised learning (so something like MixMatch ought to work brilliantly).","2019-09-20 00:35 +0000","135.0","3.0","0.022222222222222223","0.0","1.0","1.0","0.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174792830402908160","https://twitter.com/gwern/status/1174792830402908160","@AndreTI @OpenAI I dunno about that. How do you train that adversarial Transformer? By distinguishing real/fake? Even assuming you have the relevant corpus, then that's just a GAN, and those don't work well for sequences like text.","2019-09-19 21:10 +0000","3334.0","5.0","0.0014997000599880025","0.0","1.0","1.0","3.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174792429427445764","https://twitter.com/gwern/status/1174792429427445764","@mja So? The global average across all traits for c^2 using ACE in Polderman et al 2015 is still only 0.17: https://t.co/nlr3JeQW9H","2019-09-19 21:08 +0000","531.0","2.0","0.003766478342749529","0.0","0.0","1.0","0.0","1.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174730405125263362","https://twitter.com/gwern/status/1174730405125263362","@LeoBman @DrLukeOR Yes, a bit of reinventing the wheel here. Of course people have wondered if there's serious overfitting... But there's not - not even on MNIST! eg https://t.co/HUPkmu6sJx https://t.co/uOIPINRb6D https://t.co/5Kd69tPHPn https://t.co/YxXuqBm8Lb https://t.co/pCkS03fC3q","2019-09-19 17:02 +0000","3192.0","234.0","0.07330827067669173","2.0","0.0","27.0","8.0","188.0","0.0","9.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174718269636431878","https://twitter.com/gwern/status/1174718269636431878","@ludwigschubert @font https://t.co/ngElxWNTdZ","2019-09-19 16:13 +0000","188.0","3.0","0.015957446808510637","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","3","3","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174717286940327937","https://twitter.com/gwern/status/1174717286940327937","@OpenAI ????????. Excited to see if the code is usable for improving poetry generation.","2019-09-19 16:10 +0000","5743.0","53.0","0.009228626153578268","0.0","1.0","17.0","22.0","0.0","0.0","13.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174702994065186816","https://twitter.com/gwern/status/1174702994065186816","@i_contemplate_ I don't know about that. Vaping was the great hope for legacy tobacco to switch into a healthy but still profitable and growing niche. Tobacco isn't exactly a growth industry (at least in the first world).","2019-09-19 15:13 +0000","160.0","2.0","0.0125","0.0","0.0","1.0","1.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174693312739106817","https://twitter.com/gwern/status/1174693312739106817","@i_contemplate_ The Baptists I see, the usual anti-smoking advocates, but who are the bootleggers? Juul and the legacy tobacco companies are being cutoff at the knees by *nicotine* bans, and the marijuana people are all fly-by-nights or illegal or way too tiny to be doing this lobbying.","2019-09-19 14:34 +0000","220.0","3.0","0.013636363636363636","0.0","1.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174690145527902208","https://twitter.com/gwern/status/1174690145527902208","@i_contemplate_ The panic is very puzzling. It's obviously the marijuana stuff, and all the coverage admits it doesn't even have anything to do with regular legal nicotine vaping fluid (as should be obvious since if it'd hospitalize or kill you within weeks, that'd've happened decades ago!).","2019-09-19 14:22 +0000","439.0","8.0","0.018223234624145785","2.0","1.0","0.0","2.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174499476855304192","https://twitter.com/gwern/status/1174499476855304192","@XiXiDu https://t.co/N59xI7jBD7","2019-09-19 01:44 +0000","397.0","12.0","0.030226700251889168","1.0","0.0","1.0","1.0","8.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174496445677867010","https://twitter.com/gwern/status/1174496445677867010","@Miles_Brundage Wait, before you take off - https://t.co/unLSvxh5Y0 !","2019-09-19 01:32 +0000","750.0","58.0","0.07733333333333334","0.0","0.0","1.0","1.0","56.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174475629477212160","https://twitter.com/gwern/status/1174475629477212160","@ludwigschubert Oh, come to think of it, I actually did that too. I added to the compilation script a sed call to prefix in 'font-display: swap;' to the '@font-families' in the MathJax-generated CSS. (I regret nothing.)","2019-09-19 00:09 +0000","275.0","1.0","0.0036363636363636364","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174450286649384960","https://twitter.com/gwern/status/1174450286649384960","@ludwigschubert Anyway! Between that and upgrading to Cloudflare Pro, I should've roughly halved initial load time for mobile users, especially ones outside the USA.","2019-09-18 22:29 +0000","267.0","3.0","0.011235955056179775","0.0","1.0","1.0","1.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174450064552542210","https://twitter.com/gwern/status/1174450064552542210","@ludwigschubert Part of the problem is that the usual web optimization guides (*cough* Google Pagespeed's blocking-CSS recommendations *cough*) don't even mention this is possible. I assumed no one would resort to these crazy JS hacks if there was any remotely plausible option. Apparently not.","2019-09-18 22:28 +0000","212.0","3.0","0.014150943396226415","0.0","1.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174394916333641728","https://twitter.com/gwern/status/1174394916333641728","@janerikstrasser @gdb @lt0gt @ilyasut I was definitely told it was by Nick when I was there, but maybe it got garbled along the way: he did the cover but not the original drawing, or something like that. (Many leprechauns are like that in dropping complexity & simplifying.)","2019-09-18 18:49 +0000","105.0","3.0","0.02857142857142857","0.0","0.0","1.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174386292886515713","https://twitter.com/gwern/status/1174386292886515713","@janerikstrasser @gdb @lt0gt @ilyasut Most biological owls aren't made out of rivets, nor inhabit forests made out of computer code. Nor would it make any thematic sense for a symbol of *wisdom* in a book about machine intelligence to then be biological. (Possibly the original artwork makes this clearer in person.)","2019-09-18 18:14 +0000","156.0","2.0","0.01282051282051282","0.0","0.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174376675573927936","https://twitter.com/gwern/status/1174376675573927936","@gdb @lt0gt @ilyasut Many people have hidden talents. Did you know Nick Bostrom drew that mechanical owl on the cover of _Superintelligence_? I thought it was a great cover, and had no idea Nick did it until I saw it hanging in the FHI offices.","2019-09-18 17:36 +0000","1578.0","32.0","0.020278833967046894","0.0","2.0","18.0","7.0","0.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174294570307346436","https://twitter.com/gwern/status/1174294570307346436","@mattparlmer I think the general attitude there is ""it may not be legal for me to be hosting it, and anyway, the journals & libraries have copies of anything *real* or *important*, also I am way too busy trying to not perish by publishing to think about the long run of said publishing"".","2019-09-18 12:10 +0000","253.0","5.0","0.019762845849802372","0.0","0.0","3.0","2.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174140778278465536","https://twitter.com/gwern/status/1174140778278465536","@Duderichy One of these days. I've got too much to read and write to check out rando's websites.","2019-09-18 01:59 +0000","138.0","4.0","0.028985507246376812","0.0","0.0","1.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174136288619966465","https://twitter.com/gwern/status/1174136288619966465","Optimization: I always assumed that to 'lazy load' CSS & avoid blocking, you had to use insane JS hacks (like Google wants). You couldn't just stick CSS at the bottom of a page like JS. TIL: this is explicitly allowed by the spec? https://t.co/l2JkqbP611 https://t.co/w2A6YSN3fY","2019-09-18 01:41 +0000","7379.0","72.0","0.009757419704567014","0.0","2.0","8.0","0.0","47.0","0.0","15.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174130726222139394","https://twitter.com/gwern/status/1174130726222139394","@TeaGeeGeePea @gcochran99 @RokoMijicUK @KirkegaardEmil Schizophrenia is interesting because it's also now a confirmed fitness hit, but we do see SCZ pop up all the time in selective sweeps, with consistent genetic correlations across cohorts. So why SCZ? Maybe 'background selection' drags up SCZ mutation load! https://t.co/D79rbA205k","2019-09-18 01:19 +0000","390.0","12.0","0.03076923076923077","0.0","0.0","0.0","1.0","6.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174118740289576960","https://twitter.com/gwern/status/1174118740289576960","@gcochran99 @RokoMijicUK @KirkegaardEmil Sure, but I was pointing out what selection gets you on that remaining part of the variance which does seem to be genetic (and is captured in the current PGS). No reason to think the PGS is a complete fluke and not tagging any causal variants.","2019-09-18 00:31 +0000","600.0","9.0","0.015","0.0","0.0","3.0","1.0","0.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174113174397865985","https://twitter.com/gwern/status/1174113174397865985","@M_PF Is it the moustache, or the chinfur?","2019-09-18 00:09 +0000","520.0","3.0","0.0057692307692307696","0.0","0.0","1.0","2.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174102224001257472","https://twitter.com/gwern/status/1174102224001257472","@mattparlmer 'Decay'? That's how it's always been on the Internet, for everything.","2019-09-17 23:25 +0000","988.0","13.0","0.013157894736842105","0.0","1.0","7.0","5.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174095361004244993","https://twitter.com/gwern/status/1174095361004244993","@mark_riedl It is known. (But I will never get tired of watching GIFs of hilarious AI exploits which I can add to my list.)","2019-09-17 22:58 +0000","750.0","11.0","0.014666666666666666","0.0","0.0","6.0","3.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174090218804318208","https://twitter.com/gwern/status/1174090218804318208","@EvaMonkey Fascinating! So we finally have a primary source confirming the damaged/missing leg. This also seems to be different from the AERA 08/31/1998 Asahi interview which was supposedly the source of that. Looks like this is much more detailed, too.","2019-09-17 22:38 +0000","571.0","2.0","0.0035026269702276708","0.0","0.0","1.0","0.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174082271671988224","https://twitter.com/gwern/status/1174082271671988224","@rubenarslan @annemscheel @WiringTheBrain @MarcusMunafo @StuartJRitchie @ent3c @AllardThuriot ""Forget it, Jake. It's Chinatown."" See the talk page.","2019-09-17 22:06 +0000","284.0","6.0","0.02112676056338028","0.0","1.0","1.0","1.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174064606358179840","https://twitter.com/gwern/status/1174064606358179840","@vectorgen @XiXiDu Well, if you didn't know about it, maybe it refers to someone else then! Are there ????? critiques we should know about?","2019-09-17 20:56 +0000","108.0","2.0","0.018518518518518517","0.0","1.0","0.0","1.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174056332196028416","https://twitter.com/gwern/status/1174056332196028416","@XiXiDu I noticed the editorial warning at the bottom of the paper, which is unusual. I assume it refers to https://t.co/KdXaiTcMdU","2019-09-17 20:23 +0000","859.0","11.0","0.012805587892898719","1.0","2.0","4.0","1.0","1.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174047256099676160","https://twitter.com/gwern/status/1174047256099676160","@michael_nielsen @internetarchive I've been getting sporadic issues myself, but typically waiting and reloading will get me in.","2019-09-17 19:47 +0000","1414.0","7.0","0.0049504950495049506","0.0","0.0","1.0","2.0","0.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1174035797294682119","https://twitter.com/gwern/status/1174035797294682119","@rubenarslan @WiringTheBrain @StuartJRitchie @ent3c How are machine-scorings of maze performances not 'blinded'? Even Rosenthal in his expectancy book advocates using automated scoring to reduce bias.","2019-09-17 19:02 +0000","427.0","5.0","0.0117096018735363","0.0","0.0","1.0","0.0","0.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173760111107616768","https://twitter.com/gwern/status/1173760111107616768","@C_Harwick @biorxivpreprint @CrossrefOrg @theshawwn /sigh. The mobile & Chrome & Firefox bugs have proven to be intractable. The transforms part of the CSS spec is completely bizarre and apparently both Chrome *and* Firefox implement it in subtly broken ways. We've reverted to the last Aug version, sacrificing sections.","2019-09-17 00:46 +0000","249.0","7.0","0.028112449799196786","0.0","0.0","1.0","1.0","0.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173748512317759490","https://twitter.com/gwern/status/1173748512317759490","@gojomo Just found the cover for the rerelease of _A Mind Forever Voyaging_: https://t.co/uJNaev9rLc","2019-09-17 00:00 +0000","409.0","8.0","0.019559902200488997","0.0","0.0","1.0","0.0","2.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173744606401503233","https://twitter.com/gwern/status/1173744606401503233","@jackinlondon @Of_Of_Of_Of_Of @ESYudkowsky Exact solutions are often mystifying and hard to understand. This was perhaps first noticed in chess endgame databases ( https://t.co/rsp4yWWNAL ): they are optimal by construction and can't be beaten, but weeeird. I don't intuit the timing of many bets in mugging games.","2019-09-16 23:44 +0000","174.0","4.0","0.022988505747126436","0.0","0.0","0.0","0.0","2.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173736634661113858","https://twitter.com/gwern/status/1173736634661113858","@ESYudkowsky So we have a more optimized version which can evaluate as deep as 74 rounds: https://t.co/yGs1xSHqbW Trying to run it on some AWS high-mem spot instances to go deeper, but it has none, it says. Infuriating. It is my right as an American to rent 4TB of RAM for $8 whenever I want!","2019-09-16 23:13 +0000","2026.0","40.0","0.019743336623889437","0.0","0.0","14.0","3.0","17.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173731350140137472","https://twitter.com/gwern/status/1173731350140137472","@mattblaze @matthew_d_green Given the way the NSA handmedowns stuff to the DEA sometimes, one day that may be true - ??? ???? ???'?? ?? ?????!","2019-09-16 22:52 +0000","2968.0","6.0","0.0020215633423180594","0.0","0.0","2.0","2.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173607486961725440","https://twitter.com/gwern/status/1173607486961725440","@1hplovecraft @torinmccabe @themightypuck @KirkegaardEmil (Meehl discusses this nicely for psychology, but Derek Lowe has noted this also happens in chemistry sometimes; some areas of research just 'mysteriously' stop being done. Nobody ever refutes it formally, but people get the word.)","2019-09-16 14:40 +0000","289.0","3.0","0.010380622837370242","0.0","0.0","1.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173606029826887680","https://twitter.com/gwern/status/1173606029826887680","@1hplovecraft @torinmccabe @themightypuck @KirkegaardEmil You should also remember that in the soft sciences, nothing ever gets refuted; like old soldiers, they just fade away. The lack of direct replications or extensions shows (pace the Many Labs PMs) that the field knew perfectly well it's extremely dubious.","2019-09-16 14:34 +0000","330.0","7.0","0.021212121212121213","0.0","1.0","0.0","0.0","0.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173605685965275136","https://twitter.com/gwern/status/1173605685965275136","@1hplovecraft @torinmccabe @themightypuck @KirkegaardEmil You could, I've seen some cool ML-based camera-tracking systems on Biorxiv, and you can genotype each mouse to trace exact ancestry etc. But you couldn't get funds for it because it'd be 'unethical', as one person who tried told me the IRB killed their attempt.","2019-09-16 14:32 +0000","558.0","1.0","0.0017921146953405018","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173430027188396032","https://twitter.com/gwern/status/1173430027188396032","@KirkegaardEmil Still a cute finding. Might not be huge here, but 'indifference of the indicator'. A trace of the big C.","2019-09-16 02:54 +0000","653.0","7.0","0.010719754977029096","0.0","0.0","1.0","1.0","0.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173423728946028544","https://twitter.com/gwern/status/1173423728946028544","@torinmccabe @themightypuck @KirkegaardEmil It's those too.","2019-09-16 02:29 +0000","499.0","4.0","0.008016032064128256","0.0","1.0","1.0","1.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173405368724152322","https://twitter.com/gwern/status/1173405368724152322","@brogram_ @Kartikitrak @paulg @peterthiel @zaoyang @vkhosla @rabois @jordanbpeterson @patrickc @jtoomim @demishassabis @balajis @sentientist @foundmyfitness @RayDalio @eladgil Uh oh. As an outlier, by regression to the mean, that suggests I don't deserve to be in the list and am included due to sampling error!","2019-09-16 01:16 +0000","499.0","9.0","0.018036072144288578","0.0","1.0","2.0","3.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173405143561379841","https://twitter.com/gwern/status/1173405143561379841","@ctphoenix Oh no, GPT-2 repeats a ton. And image models like PixelRNN or PixelCNN are autoregressive image prediction/generation, so it seems like DeepDream-like tesselating ought to be possible. So I wonder why the image models avoid repeating 'words' while LMs always are vulnerable.","2019-09-16 01:16 +0000","115.0","1.0","0.008695652173913044","0.0","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173332317001195528","https://twitter.com/gwern/status/1173332317001195528","@the8472 That's a big enough image, with enough annotations, that it would be quite a challenge for any existing CNN architectures. Definitely not going to get good performance with a single pass, I suspect.","2019-09-15 20:26 +0000","87.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173290205853421571","https://twitter.com/gwern/status/1173290205853421571","@ctphoenix Another idle question about likelihood losses: all likelihood-trained text sequence NNs fall into repetition very easily. So why don't we ever seem to see image likelihood models like Glow or PixelCNN etc have failure modes like tessellating dog heads in an image?","2019-09-15 17:39 +0000","130.0","2.0","0.015384615384615385","0.0","1.0","0.0","0.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1173273816761602051","https://twitter.com/gwern/status/1173273816761602051","@RichardWiseman @KPUNews Caroline Watt has provided photos which I've cleaned up into PDFs: - ""Models of control and control of bias"" https://t.co/uSo7EYmplM - EJP editorial https://t.co/WNZoExfaGk - ""On Publication Policy Regarding Non-Significant Results"" https://t.co/FcCqzVA6XT","2019-09-15 16:34 +0000","417.0","5.0","0.011990407673860911","0.0","0.0","0.0","0.0","2.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172978315717689344","https://twitter.com/gwern/status/1172978315717689344","@Scientific_Bird @SilverVVulpes @hsu_steve Hm. https://t.co/Rb377ewM56 ? Alternative: https://t.co/vzyK1FsbDu","2019-09-14 20:59 +0000","460.0","18.0","0.0391304347826087","0.0","1.0","1.0","1.0","15.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172942396281315335","https://twitter.com/gwern/status/1172942396281315335","@Meaningness Yes, and lots of possibilities for small-scale researchers too. For example, why not take all the text adventure games and do imitation learning on automatically-generated or recorded game trajectories and learn the dynamics?","2019-09-14 18:37 +0000","847.0","5.0","0.0059031877213695395","0.0","0.0","3.0","1.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172941201131147264","https://twitter.com/gwern/status/1172941201131147264","@Meaningness I'm looking forward to more progress on LMs and integrating with memories and/or knowledge databases. It'll open up entirely new kinds of text adventure games, and in particular, with enough world modeling, create a DM-in-a-pocket.","2019-09-14 18:32 +0000","866.0","9.0","0.010392609699769052","0.0","1.0","4.0","0.0","0.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172908553440583682","https://twitter.com/gwern/status/1172908553440583682","@FelixThoemmes But muh heterogeneity and WEIRD!","2019-09-14 16:22 +0000","251.0","5.0","0.0199203187250996","0.0","0.0","1.0","3.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172879202519080960","https://twitter.com/gwern/status/1172879202519080960","@EileenOrmsby So my theory is that much of the humor of groin shots is a 'there but for the grace of God' effect.","2019-09-14 14:26 +0000","267.0","7.0","0.026217228464419477","0.0","0.0","1.0","0.0","0.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172878983517679616","https://twitter.com/gwern/status/1172878983517679616","@EileenOrmsby The thing about getting hit in the nuts is that 95% of the time, it hardly hurts; then the 5% is the worst agony one's ever experienced. From being kicked in TKD, I'd say it's worse than having my 4 wisdom teeth out. (I quickly developed Zen master levels of groin protection.)","2019-09-14 14:25 +0000","904.0","8.0","0.008849557522123894","0.0","1.0","2.0","5.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172718732092891137","https://twitter.com/gwern/status/1172718732092891137","It makes even more sense to do this for Bitcoin, so I'm happy I finally got that implemented. It'll be really useful on my Silk Road page and a few others. https://t.co/yqENkuR43X","2019-09-14 03:48 +0000","9066.0","311.0","0.034303992940657405","0.0","0.0","12.0","4.0","20.0","0.0","49.0","0.0","0","0","0","0","0","226","226","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172681927964471297","https://twitter.com/gwern/status/1172681927964471297","@dhruvbansal @clarkmoody @danheld @bquittem @TuurDemeester @BuckPerley @MartyBent @TaylorPearsonMe @RyanTheGentry Maybe I am missing something, but why does there need to be any conflict between Bitcoin and currencies localized elsewhere? Atomic swaps already exist and neuter any possible conflict. Going to Earth? Swap some Muskcoins for Bitcoins. Attacks get factored into timelock margins.","2019-09-14 01:22 +0000","684.0","12.0","0.017543859649122806","0.0","1.0","6.0","0.0","0.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172673712665088000","https://twitter.com/gwern/status/1172673712665088000","@tonysheng Mirror of the 2nd: https://t.co/YgYALHPjU0","2019-09-14 00:49 +0000","213.0","6.0","0.028169014084507043","0.0","0.0","0.0","0.0","3.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172652818077011968","https://twitter.com/gwern/status/1172652818077011968","@Aceinmysleeve1 @robinhanson As required by McKean's law...","2019-09-13 23:26 +0000","574.0","3.0","0.005226480836236934","0.0","0.0","1.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172625872874082305","https://twitter.com/gwern/status/1172625872874082305","@ntborgen @AmirSariaslan Not sure I follow the argument here. If any collider bias here is attenuating the relationship (rather than reversing it), then surely that makes finding a relationship even stronger evidence for there being a (larger) causal effect, and thus justification for a ban?","2019-09-13 21:39 +0000","506.0","4.0","0.007905138339920948","0.0","1.0","1.0","2.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172623286217449472","https://twitter.com/gwern/status/1172623286217449472","@hsu_steve Cow breeding estimate of limits to selection I don't think you've mentioned before: even optimizing just one chromosome's haplotypes yields +1-2SD: https://t.co/yp4JWsla39 Total lower bound, 36SD; expected, 97SD; upper bound: 436SD. https://t.co/WcP9T1tbV7","2019-09-13 21:29 +0000","1392.0","101.0","0.07255747126436782","2.0","1.0","12.0","12.0","17.0","0.0","18.0","0.0","0","0","0","0","0","39","39","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172577877491146753","https://twitter.com/gwern/status/1172577877491146753","@eiaine @alt_kia You didn't even get past the customer service reps; good simjackers do. A hacker who doxed me worked his way up to internal chatrooms with the ISP's sysadmins to get info on my IP, & noted to me that it would have been trivial to simjack me. (I removed all SMS 2FA the next day.)","2019-09-13 18:28 +0000","792.0","7.0","0.008838383838383838","0.0","0.0","1.0","1.0","0.0","0.0","5.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172538550073663488","https://twitter.com/gwern/status/1172538550073663488","@SilverVVulpes We're not exactly talking about Standard Model predictions here, after all...","2019-09-13 15:52 +0000","244.0","4.0","0.01639344262295082","0.0","0.0","1.0","1.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172538383446544384","https://twitter.com/gwern/status/1172538383446544384","@SilverVVulpes Sure but 'some guys once got some fudge factors wrong in multiplying a bunch of numbers to estimate something about small weird asteroids they couldn't observe anything about' is not that much of a stretch. Sure, maybe they were a couple of OOMs off. Why not?","2019-09-13 15:51 +0000","1290.0","11.0","0.008527131782945736","0.0","2.0","5.0","3.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172536147458842629","https://twitter.com/gwern/status/1172536147458842629","@SilverVVulpes One man's modus ponens... Combined with the Great Silence, a second observation so quickly proves it's not aliens.","2019-09-13 15:42 +0000","1199.0","22.0","0.01834862385321101","0.0","1.0","18.0","3.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172335123175313408","https://twitter.com/gwern/status/1172335123175313408","@parisreview Seems to be working now, thanks.","2019-09-13 02:24 +0000","562.0","1.0","0.0017793594306049821","0.0","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172323753759395842","https://twitter.com/gwern/status/1172323753759395842","@Of_Of_Of_Of_Of @ESYudkowsky Well, if your sensibility is that refined and delicate, maybe just read https://t.co/IFHubMrGOa and then do the obvious extension to Pascal's mugging.","2019-09-13 01:38 +0000","408.0","30.0","0.07352941176470588","0.0","1.0","1.0","1.0","24.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172320531414044674","https://twitter.com/gwern/status/1172320531414044674","@ESYudkowsky Some work on solving Pascal's mugging: https://t.co/KeKDHwObJj","2019-09-13 01:26 +0000","3616.0","128.0","0.035398230088495575","1.0","2.0","13.0","7.0","99.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172311153365569536","https://twitter.com/gwern/status/1172311153365569536","@RichardWiseman @KPUNews Thanks.","2019-09-13 00:48 +0000","660.0","1.0","0.0015151515151515152","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172231251413143552","https://twitter.com/gwern/status/1172231251413143552","@ABM_paints @JJ_Emerson Of course you can.","2019-09-12 19:31 +0000","370.0","2.0","0.005405405405405406","0.0","0.0","0.0","2.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172204764525989888","https://twitter.com/gwern/status/1172204764525989888","@ABM_paints @JJ_Emerson A snowbank of snow and ice, as described, is not a good place to get liquid water. Where's all that extra heat going to come from?","2019-09-12 17:46 +0000","405.0","1.0","0.0024691358024691358","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172190144272707584","https://twitter.com/gwern/status/1172190144272707584","@ABM_paints @JJ_Emerson Where was Freuchen going to get a bucket of pure liquid water when he was freezing to death?","2019-09-12 16:48 +0000","392.0","3.0","0.007653061224489796","0.0","1.0","0.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172189907495862272","https://twitter.com/gwern/status/1172189907495862272","@ArtirKel @rubenarslan @GordPennycook @dingding_peng Yes, this is done sometimes by meta-analysts (figuring out the optimal set of trials for a given n or budget), although more typically the quantity being minimized is something like posterior predictive intervals or total variance than direct decision analysis.","2019-09-12 16:47 +0000","736.0","5.0","0.006793478260869565","0.0","2.0","0.0","2.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172187190635446272","https://twitter.com/gwern/status/1172187190635446272","@ArtirKel @rubenarslan @GordPennycook @dingding_peng On the other hand, heterogeneity & moderators, plus other useful things like quantifying bias in the field.","2019-09-12 16:36 +0000","479.0","5.0","0.010438413361169102","0.0","1.0","3.0","0.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1172169634222628865","https://twitter.com/gwern/status/1172169634222628865","@randomuserbr @paulg ""regression to the mean"" is also a phrase sorely needed in this NYT article.","2019-09-12 15:26 +0000","433.0","10.0","0.023094688221709007","0.0","0.0","2.0","6.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171966908956119042","https://twitter.com/gwern/status/1171966908956119042","@hardmaru @nature @schmidhubered ""...collect more citations than the most influential founders of their field"" ?","2019-09-12 02:00 +0000","1850.0","11.0","0.005945945945945946","0.0","2.0","2.0","5.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171942518180724736","https://twitter.com/gwern/status/1171942518180724736","@polyaletheia @JJ_Emerson Unfairly, I thought. In a world in which https://t.co/GSS4CXNY1K exists, can we really dismiss poo-chisels or poo-knives just because one group of researchers couldn't replicate on their first try? (Would that be convincing for, say, stone knapping?)","2019-09-12 00:24 +0000","5873.0","60.0","0.01021624382768602","0.0","2.0","22.0","6.0","24.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171903211067838464","https://twitter.com/gwern/status/1171903211067838464","@JJ_Emerson In the Inuits' defense, their use of frozen feces was also replicated by Peter Freuchen, who chipped his way out of an ice trap with a poo-chisel: https://t.co/s3O8z2rTvL","2019-09-11 21:47 +0000","7979.0","121.0","0.015164807620002507","1.0","4.0","24.0","5.0","74.0","0.0","13.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171890960051187712","https://twitter.com/gwern/status/1171890960051187712","@RokoMijicUK @XiXiDu Mentally, we live in a Minecraft world.","2019-09-11 20:59 +0000","136.0","2.0","0.014705882352941176","0.0","1.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171890716039163904","https://twitter.com/gwern/status/1171890716039163904","@minimaxir @RichardSocher Sure, I was expecting to have to use CPU mode. But I was trying to figure out if it was yet another error in getting it running or an issue with CPU mode itself. Hard to tell since it might've been OOMing on GPU before it hit whatever issue I was having on CPU. Gave up after that","2019-09-11 20:58 +0000","317.0","1.0","0.0031545741324921135","0.0","0.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171888497726283776","https://twitter.com/gwern/status/1171888497726283776","@minimaxir @RichardSocher I don't think so. With the patch, I get 'tf does not exist' errors. Without the patch on GPU (1080ti/11GB), I get OOMs; and without the patch on CPU, I eventually get entirely other errors.","2019-09-11 20:49 +0000","468.0","3.0","0.00641025641025641","0.0","1.0","1.0","1.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171869124613591040","https://twitter.com/gwern/status/1171869124613591040","@RichardSocher For anyone getting 'invalid_grant: Bad Request' errors downloading the models, remember that gsutil credentials do expire so you may need to 'gcloud auth login' again...","2019-09-11 19:32 +0000","4215.0","23.0","0.005456702253855279","0.0","0.0","12.0","10.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171863290873548800","https://twitter.com/gwern/status/1171863290873548800","@parisreview Website bug: ""Moxomenon"" https://t.co/VkC2ATsl9e seems to be missing pg166-170 from the print edition. The web edition just stops at ""My whole life but with hands like this my repertoire"". (Other things like ""Why Visit America"" in the same issue seem fine online.)","2019-09-11 19:09 +0000","794.0","6.0","0.007556675062972292","0.0","1.0","0.0","1.0","3.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171781373771886592","https://twitter.com/gwern/status/1171781373771886592","@Artbreeder It seems to be working now. Great! Now all I want for Christmas are the BigGAN & StyleGAN anime models... ?","2019-09-11 13:43 +0000","772.0","7.0","0.009067357512953367","0.0","0.0","5.0","1.0","0.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171780870824439808","https://twitter.com/gwern/status/1171780870824439808","@KirkegaardEmil Evolution is hard: https://t.co/uL6YIdvhvz I wonder if Meehl would've found abiogenesis etc easier to accept if he had available more computational examples of creativity/evolution like https://t.co/5vyV9Yj3yE https://t.co/PoEeidcPSY","2019-09-11 13:41 +0000","706.0","18.0","0.025495750708215296","0.0","0.0","6.0","0.0","8.0","0.0","4.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171605996931866624","https://twitter.com/gwern/status/1171605996931866624","@briantemerowski @jjding99 Yes, and I was like the 7th person, at most, to generate poetry snippets with GPT-2 too... The only 'first' I could claim was finetuning/retraining. AFAIK, everyone else was using prompts with the raw model, they didn't retrain (because nshepperd's codebase didn't exist).","2019-09-11 02:06 +0000","744.0","9.0","0.012096774193548387","0.0","0.0","5.0","1.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171559036740919301","https://twitter.com/gwern/status/1171559036740919301","@RichardWiseman @KPUNews Amazing find. Amused that the rate of significance drops just like in RRs elsewhere - 'parapsychology, the control group for science' indeed... Do you have copies of the 2 Johnson articles proposing RRs? EJP digitized archives don't go nearly that far back, it seems.","2019-09-10 23:00 +0000","242.0","9.0","0.0371900826446281","0.0","1.0","0.0","1.0","1.0","0.0","6.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171548891147317249","https://twitter.com/gwern/status/1171548891147317249","@RealtimeAI @robinhanson Skeptical, & envious. Particularly in the _Paris Review_ interviews, a lot of them reply to the work habits question by expressing the wish to be able to work long regular hours, but say they just can't do more than the usual 3-4 hours a day or work at more than irregular times.","2019-09-10 22:19 +0000","228.0","4.0","0.017543859649122806","0.0","0.0","4.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171548271841558528","https://twitter.com/gwern/status/1171548271841558528","@RealtimeAI @robinhanson My point there is that even people who are stereotyped as intelligence supremacists don't exhibit the caricature Hanson claims is modal. The emphasis on energy is one shared by layman and researcher alike, and has been since Galton started discussing the psychology of eminence.","2019-09-10 22:17 +0000","290.0","2.0","0.006896551724137931","0.0","1.0","1.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171547853946335234","https://twitter.com/gwern/status/1171547853946335234","@RealtimeAI @robinhanson None of them say anything absurd like 'Oh yeah, you can totally work 20 hours a day like Danielle Steele if you just try really hard.' (I think Steele must be a short-sleeper, incidentally, she fits the profile perfectly.)","2019-09-10 22:15 +0000","127.0","2.0","0.015748031496062992","0.0","0.0","2.0","0.0","0.0","0.0","0.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171547661620666368","https://twitter.com/gwern/status/1171547661620666368","@RealtimeAI @robinhanson The attitude most take is that all you can do is make the most of however many hours a day you get, and arrange your habits and do your best to seize the writing hours. Many are skeptical of anyone claiming to work many hours.","2019-09-10 22:15 +0000","257.0","6.0","0.023346303501945526","0.0","2.0","2.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171547216667992064","https://twitter.com/gwern/status/1171547216667992064","@RealtimeAI @robinhanson For example, I've been reading through ~400 interviews with successful writers so far for https://t.co/cCOh7I4o7P . Many emphasize the importance of hard work. I can think of very few who say anything like 'it's easy to work harder! Just use the ""try harder"", Luke!'","2019-09-10 22:13 +0000","210.0","13.0","0.06190476190476191","0.0","1.0","7.0","0.0","2.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171544682066796545","https://twitter.com/gwern/status/1171544682066796545","@RealtimeAI @robinhanson I don't agree with that either. You can find lots of people, including such minor obscure figures as Francis Galton or Arthur Jensen, who don't say that in discussing the joint necessity of energy and intelligence in great achievement.","2019-09-10 22:03 +0000","430.0","8.0","0.018604651162790697","0.0","3.0","3.0","0.0","0.0","0.0","2.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171505659214364678","https://twitter.com/gwern/status/1171505659214364678","@DegenRolf Mirror: https://t.co/8DJqXJ9uTP","2019-09-10 19:28 +0000","1226.0","39.0","0.03181076672104405","0.0","2.0","5.0","0.0","22.0","0.0","10.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171493320750583819","https://twitter.com/gwern/status/1171493320750583819","@robinhanson Disagree. We hear tons about putting in long hours, being super hardworking, skipping sleep, 'sweat equity', getting up at 4AM to do email etc, in coverage of both science and business, to the point of backlashes where people boast about only working 40 hour weeks.","2019-09-10 18:39 +0000","2350.0","62.0","0.026382978723404255","0.0","1.0","27.0","7.0","0.0","0.0","27.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171492717907439616","https://twitter.com/gwern/status/1171492717907439616","@genomic_pred Sounds a lot like https://t.co/Cz2M58Cjnf . An unsupervised autoencoder predicting as much of visual variance as possible is going to serve as a high-dimensional fingerprint/ID/similarity, enabling heritability much like GCTA etc.","2019-09-10 18:36 +0000","218.0","5.0","0.022935779816513763","0.0","0.0","2.0","0.0","2.0","0.0","1.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171454664048381953","https://twitter.com/gwern/status/1171454664048381953","@Miles_Brundage This could actually be really useful for me. Instead of popping up a screenshot for links without an annotation/summary, the page could be dumped into this!","2019-09-10 16:05 +0000","12149.0","33.0","0.002716272944275249","0.0","0.0","20.0","2.0","0.0","0.0","11.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" "1171223264833810432","https://twitter.com/gwern/status/1171223264833810432","@whyvert @WonkaWasRight I'm not sure about that. Inbreeding isn't useful for additive variance (I mention this in my Dune discussion), you're very limited in your selection if you choose among relatives so not meritocratic, and you get some inbreeding depression.","2019-09-10 00:46 +0000","765.0","10.0","0.013071895424836602","0.0","0.0","6.0","1.0","0.0","0.0","3.0","0.0","0","0","0","0","0","0","0","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-","-" He interrupted her. Close at hand is a stable where two beautiful ponies are kept. They are snowy white, and are consecrated to the goddess Ku-wanon, the deity of mercy, who is the presiding genius of the temple. They are in the care of a young girl, and it is considered a pious duty to feed them. Pease and beans are for sale outside, and many devotees contribute a few cash for the benefit of the sacred animals. If the poor beasts should eat a quarter of what is offered to them, or, rather, of what is paid for, they would soon die of overfeeding. It is shrewdly suspected that the grain is sold many times over, in consequence of a collusion between the dealers and the keeper of the horses. At all events, the health of the animals is regarded, and it would never do to give them all that is presented. On their return from the garden they stopped at a place where eggs are hatched by artificial heat. They are placed over brick ovens or furnaces, where a gentle heat is kept up, and a man is constantly on watch to see that the fire neither burns too rapidly nor too slowly. A great heat would kill the vitality of the egg by baking it, while if the temperature falls below a certain point, the hatching process does not go on. When the little chicks appear, they are placed under the care of an artificial mother, which consists of a bed of soft down and feathers, with a cover three or four inches above it. This cover has strips of down hanging from it, and touching the bed below, and the chickens nestle there quite safe from outside cold. The Chinese have practised this artificial hatching and rearing for thousands of years, and relieved the hens of a great deal of the monotony of life. He would not have it in the scabbard, and when I laid it naked in his hand he kissed the hilt. Charlotte sent Gholson for Ned Ferry. Glancing from the window, I noticed that for some better convenience our scouts had left the grove, and the prisoners had been marched in and huddled close to the veranda-steps, under their heavy marching-guard of Louisianians. One of the blue-coats called up to me softly: "Dying--really?" He turned to his fellows--"Boys, Captain's dying." Assuming an air of having forgotten all about Dick¡¯s rhyme, he went to his place in the seat behind Jeff and the instant his safety belt was snapped Jeff signaled to a farmer who had come over to investigate and satisfy himself that the airplane had legitimate business there; the farmer kicked the stones used as chocks from under the landing tires and Jeff opened up the throttle. ¡°Yes,¡± Dick supplemented Larry¡¯s new point. ¡°Another thing, Sandy, that doesn¡¯t explain why he¡¯d take three boys and fly a ship he could never use on water¡ªwith an amphibian right here.¡± Should you leave me too, O my faithless ladie? And years of remorse and despair been your fate, That night was a purging. From thenceforward Reuben was to press on straight to his goal, with no more slackenings or diversions. "Is that you, Robin?" said a soft voice; and a female face was seen peeping half way down the stairs. HoMElãñÔóÂÜÀ­³ó ENTER NUMBET 0016iuptnb.com.cn
fbpghr.com.cn
www.jtchain.com.cn
tianyu0.com.cn
tgergz.com.cn
skicms.com.cn
www.szsubway.com.cn
pttjui.com.cn
nthnxszp.org.cn
pupemi.com.cn
插你肥穴 制服丝袜旗袍涶乱的小电影 日本兽王无码全集 朋友们开撸吧欧美图片 � 人兽性交影片 维族人黄色片 日本女孩小嫩穴摄影 三八人体图片基地 美国 女人体图片欣赏 女人洞人体艺术 强奸做爱短片小说 父媳乱来 成人乱伦小说狠狠 91自拍电影网 我的美艳岳母小说撸二哥 谁有仓井空的电影 堂成人免费视屏 最大 拾贰月重磅巨献高清私拍安徽合肥随妻5p av女优超市 tube兽人 人体少女成人艺术 丝袜淫娃女空姐 欧美熟妇刚交图 新号百度云怎么看片 李忠瑞影音先锋下载 日本女人超黄色人体 男女明星日逼图片 sss性交图片 超级玛丽兄弟电影三级片 gaoqingtoupaiyiyuannvce 擢撸少妇色图 美女逼图欣赏 他州色图 麻岛美几 jav365熟女俱乐部 夜夜摸东京热 7788dcom 亚洲成人电台 美女自己玩射了18p 亚洲美女掰鲍图 色妹妹影视下载 丝袜美女露阴门图片 小说 吃咪咪 黄色电影全草免费 爱爱医学网 干骚逼 射进肉洞 迅雷长片狠狠撸 大胆97人体 WWW_PS31_INFO 胖妹子和瘦妹子做爱有什么不同 WWW_ZARA_COM 明星18avday 乳交 乱伦 鸡巴 女儿的嫩洞逼 人妖变装图片 忧忧人体艺术导航 日韩考屄视频 粗鸡巴日骚屄 快播52淫 超诱惑骚妇奶子图片 黑巨屌插烂插扯阴道 四季美农贸城 招商银行客户端下载 情债电视剧 时空猎人电脑版下载 黑男白女种子 谁有亚洲美图网站2014 黑丝电影有哪些 苍井空的逼放大图 鼘掠牒趯? 就要搞妹妹 撸撸网王老撸 大鸡巴肏死小骚屄了 狼友迷奸自拍视频 韩国a片观看 徐莹性爱视频ed2k 女人和畜生做爱的小说 日本爽片种子 极品淫荡骚货 操逼有害不可大力日比 大香蕉很狠撸 喜欢日肥女人 小说享受泰国浴 亚洲成人10p 草小少妇屁股 李宗瑞100级 少女的阴蒂儿阴道 六九av影院 林心如两腿分开 汤芳最大胆全裸体清晰图 哪个网可看黄色图片 狂插妹妹15p 影音先锋黑人幼幼 WWW_211VV_COM av怎么看不了 给德国处女开包 夜色贵族成人小说区时间 美女爱骚逼 爱爱小说乱伦 欧美女人全裸照片 幼幼被3p 大吊黑逼 粉色骚逼 我插过的女人 日本熟女友田真希电影影音先锋播放 草裙操比 我要肏贾玲 黄色电影影音 小姐的阴道夹的我好爽 ww66avav 国语影视先锋 手机皇色视频网站大全 WWW59QRCOM 五月亚州婷婷炸天撸 张筱雨被男人插人体写真 天天色偷拍自拍 欧美图库亚洲图库 欧美人妻性爱写真 成人电影网址youjizz 美国骚逼想唉操 爸爸插错洞吉吉 操少女小说网 视频偷拍逼 成人网黄色片下载 跟老婆操奶操逼 大鸡巴狠插黑黑的骚逼 骚妇粉色木耳 四房开心播播图片 雨宫琴音qv电影 久久视频在线播放观看( 去干网1月25日 亚洲性交色视频 爱爱帝国钟和区图片 粉嫩写穴 快播伦理影院全国若妻 WWW578133COM 日韩孕妇做爱 成人在线电台 大学生妹妹穴穴 快播户外乱伦电影 亚洲女色网伦理 婶婶与侄子伦理小说 苍井空秘密搜禅观 德川重男种子 佐佐木春香 qvodk8经典强奸乱伦 女鲍鱼16p 幼嫩系列在线 内射欧美 我爱插逼淫淫网 啊淫水图 奸恋图片 我和小姨在车里 骚妇淫 欧美大鸡巴照片 漂亮女优伦理电影 性爱小说97 国产呦呦100部精选集 艳女淫妇 尻逼电影 乱伦乱摸乱交配 哪个明星的逼最好看 春色校 妻子让朋友干出尿 无需播放器在线观看西山希 色喜大胆人体 国产自拍偷拍在线播放若怒 成人孕妇性爱电影网 美女掰骚鲍www51gannet 骚逼蜜穴 家庭乱伦动漫电影 盗撮协和影视 射死你天天日百度 儿子射进来 怡春院人兽 操岳母色色网 少女嫩逼激情小说 自拍偷拍欧美美色图 家庭乱伦之巨乳妈妈 美女游戏封面相片 黑丝avi码合集 帅的男同性恋做爱的视频 本色国际 西西人体高清18岁 人体艺术舒淇艳门照 影视先先锋 亚非成人电影 究极人渣素人 weimeiqingchunccom 老头和妇女公园不雅视频 来点黄色的videombaiducom 色漫画之丝袜少妇 第一放映室伦理欧美 狗肏人网 农村tongjian 能直接在线观看图片的网站 操你妹手机电影 妇色小说 国产抽插开档黑丝女友 色狠窝 朋友的处女女友m06xscom 伦全集狠狠射 tt老婆 yingqi 制服超短裙番号迅雷下载 变态图区精品 谷露英语 日本一级毛片magnet 徐东东丝袜图片 AV区亚洲AV欧美AVwww9966acomwwwc5508comwwwshcateringcom 女性全裸近照无遮挡物 短篇剧情网站 亚洲无码女教师撸 午夜A片播 大尺度强暴性爱 色网站色网站 久草在线被监禁了的身体玩具 欧洲学妹性交视频 当着外人的面插入下载 日本美女性交24p图片 renyudongwuxingjiao capron马蓉出轨 明星合成欧美色色欧美性爱区欧美骚妇欧美成人色图欧美大奶欧美大胆人体 真人做色性动态图片 八戒街拍裙底 春药潮吹痉挛视频 思思热资源站 久草在线视频7mwwwgsobc1024cnmyiwancom 比基尼淫 花井美纱RKI252磁力链接 少妇熟女搞逼 裸体女奴小说 两性故事淫荡家庭 日韩少妇人妻系列 人妻的禁忌 riyelu国语 耽美骚货出水 17岁少女跪嫁68岁老汉 性交口交天天撸 制服丝袜1页日日摸 b3jjj 哥哥综合影院雅蠛蝶五次郎 偷拍自偷人妻自慰 西方人与狗性交片 艳照门视频华为网盘 偷拍性生活祝频 WWW_621YY_COM 欧美女生阴道艺术 sek115苞米地 影音先锋a9电影 亚洲成人动漫av 97成人在线乐子 2015Tv丶C0M 星河大帝同人小说 夜插 日本熟女性爱偷拍xjo530tumblrcom 美女b比图片 小泽玛利亚与人妖图片 成人大胆诱惑人体艺术图片 大色se片 偷拍自拍内衣12p 少女小山雀69Xx wwwfafa98om 美国发布站2K3HH wwww336bbbcom 成人做爱有哪些影院 亚洲777www69rhcom 尼姑云影视 制服丝袜图片网站 vr美腿丝袜 天天操日日艹 破处番号 偷拍影音先锋电影网 意淫强奸人妻女友 AV电影百度云 ggggbb 北原多香子全集种子 97资源www688bxcom 正在播放87午夜孕妇 无码一级黄视频 强奸乱伦 哥哥干 色狗狗综合视频网站 先锋影音成人动 日本av百人群Pav 成人3d卡通动漫在线 wwwjjzzocom 成人性全免费视频 网55eekcom 色和尚被偷拍的女主播 微信可以处理违章吗 a图片大全 欧美成人裸体网 欧美家庭伦理片资源下载mp4 妓女av网 欧美图片激情小说 1333ccc ee214magnet 最新日本av在线视频 姐姐爱爱集 狠狠爱五色天大香蕉 哈尔滨狼友 乱伦日韩欧美制服手机在线 wwwavsj 64sm 偷偷拍自拍15p 国产偷拍厕所在线视频 强奸军人母亲 日本有个是主播做爱的片子吗 亚洲第一在线黄色网站 东方东方伊甸园苍井空 蜜桃丁香 www2Csex872com 231avav 爱吧综合 黄色网站空姐被强奸下载 啊扣你的小穴 五月天激情网址 大香蕉精品520 色中色AV在线视频哥哥干 奇米成人影视色和尚、色尼姑 新西兰成人片 500av大香蕉 黄色1级片 看欧美AV剧情片的网站 性插网页 av大b视频 撸波波映院 人妖当着妈妈和爸爸 v2视频网站 千百撸综合 教室调教老师 另类小说专区在线 av在线总站 www725bb 摩擦影院5566 女人的穴是万能的 乱伦性交激情小说 亚洲1AV 高清av1080在线 大香蕉波波在线视频 淫汁Y 清纯唯美哥哥撸 sex888欧美片 依依社区影院撸lu射 995hao m3344tt 草民盒子 西西人体偷拍另类 狠狠操骚b 亚洲偷拍自拍最新网站 制服丝袜av天天 3d黄片 kk999pw下载 WWWavtb789com 男同志免费影片 琪琪电影网李宗瑞 色在线奶奶 579路com wwwyoujizzcom美女 有没有幼女的视频网站 欧美私处艳照 亚洲图片欧美图片日本av成人有声 国产小黄片网站 搜索www123456789 撸一撸美女人体 操小屄还玩老屄 淫淫色色色色 深爱萝莉 花样困绑虏待女人方法 很哈图热图 wwwcom84 wwwsyy31con 草莓社区2016新地址 国产av91自拍 鸡鸡和光屁股 夜一夜撸一撸 91luCOM 最新乱伦比较黄的肉文 嫩妹妹色妹妹干妹妹 av英语老师 美女合集早时候的图片 韩国浴池在线播放 奶大性爱基地 L三级片 美女的逼怡院 影音先锋偷看媳妇中文 91qq1com wweuucom 美国伦理在线 狼窝激情影院影音先锋 韩国邪恶动做片mp4 古代激情17P 男操男小说 干哥哥插妹妹逼 成人激情12P 好色视频超碰碰碰 yaolu导航最新地址 同性恋舔我小穴 在线视频主播AV 手机看片www555com 整部的黄色小说 男人强奸女人操逼的视屏 色黑木耳穴 经典乱伦性爱故事 色四图吧 广濑奈央美 无码 国产成人在线视频网站 女知青风流小说 高兴为什么呢哪俩唱的 女人wumaoyingbutu 儿子与母亲的乱伦性爱故事情节 爽图在哪里看 屄 亚洲 hongkongav 乱伦熟女做爱视频 欧美破处视频区 爆插长腿黑丝妹50p 后宫社区最新最新地址 最新肥胖人体艺术摄影 淋浴做爱av 李宗瑞性侵录像网址 热香屄 女主播艳门照 车上逼逼图 为什么我喜欢让情人操我 性爱av哪里能看 影视先锋伦理电影 国内老女人av在线视频 欧美黄色网站视频第一页 美女乱论小说 日本人操b图片 神木轮奸犯王伟 与狗爱爱网聊视频 嫩乳妹妹人体艺术 1238090com 郭富城早期经典视频 女忧快播影音先锋 色色偶性爱自拍 肏嫩屄舔雪白屁股插进去 叶欣桐图片库 国产视频操胖老奶奶10306 激情爱爱1qq号是多少 酒色视频网最新网址 成人电影微信群 苍井空的电影名字 想几个男人操逼逼 在酒吧把泥酔巨乳女带回家狂插先锋影音 色mm激情万种 美女人体馒头 明星华为网盘 老妈与我乱伦 亚洲色图 女大学生性爱 毛色黄片网站 非洲美女陆逼网战 偷拍亞洲路上幼小情侶做愛相片圖? 女人白虎屄 bb性爱视频 大鸡吧干衅 日本成人强奸乱伦电影 在哪能找艳照门 欧美嫂子qvod 我喜欢大奶女人 亚洲爱色套图 遥子里被操的图片 色爱在线视频 36周了感cnbb头己下了 不举总裁的盗种妻 美国最年轻的总统 wendao 甚至导致严重的肌腱炎。建议不要长时间发短信, 梦幻西游69群p 武器战士输出手法 韩国人体艺术新闻 关于怡红院的网名 少女18艺术图片 色沟影院 女人杂交视频 大鸡巴姐姐肏骚妹妹 日本人体艺1 上原结衣的片去哪个网站找 makeloveveido 91快播电影国外处女 人体图全女 女子健美漏屄图片 美足自慰 畸形逼 吉泽明步orrent 人妻性爱色图影音先锋 女艺术图 qovd幼交电影 韩国女明星被潜规则爱爱电影 快播免费成人禁片 日本强奸资源 熟女尿射口 鑹叉儏缃戠珯链夎皝鐭ラ亾 午夜性交图片 走光露屄毛的女人 操妹妹逼 撸撸涩涩图片 惩罚骚货熟女 小岳母的淫水 狠狠干激情 人体艺术女同性恋视频 暴操成都艺校妹妹 我爱和妈妈和妹妹乱伦小说 成人极品小说合集 人体射精艺术图30p 三级片3级片四级片视频 美女的下部肉肉图 和女仆干真爽电影 在线视频录像 WWWAAACOMYAZHO 奇艺人体艺术 人狗性交私处照 插妹妹自拍偷拍 RE789COM 鲁鲁射射 深爱激情网免费视频播放 清纯唯美中文 橹一橹小说阅读 骚女嫂子 幼女小说长篇 屄都什么样 咪咪网王老撸 手机浏览器看不了在线视频 婶婶在果园里把我了 少妇熟女乱伦强奸 粉红亚洲妹偷拍自拍 WWWWWNNVCOM 国产av大片 一圾片2013推存 日本少妇伦理小说 黄昏去操逼俄罗斯 玉女情挑吉吉音影 色四月婷婷网五月天肉文 亚州成人大奶综合电影 老江吃嫩草20 亚洲怡红院首页 日本城人电 色哥百度 WWWGAN521COM 激情网站wwwhuangsedianyingnet 强暴乱伦爱爱 httwww8888tpcom 台湾无码A片欲火焚身 婷婷五月亚洲Avwww317avcom wwwAVKUTV 本站只适合18岁以上观看 tom色播网 免费成人网绐 www111eecom四虎影裤成人精品 小苍奈奈网站 妻子地铁失身 印度毛片ed2k 儿子插完妈妈插妹妹 xingjiaoluanai WWWWEIBOCOM 淫姐姐骗了弟弟操她逼 小鲜肉操熟女 老色哥奇米影视26uuu 插进妈妈的肥逼 网友自拍偷偷师爷 成人动漫亚洲色图意淫强奸亚洲色图意淫强奸 好骚色色re 不小心小孩插大人影音先锋 韩国少女掰屄图 wwwqylsp1 YY6080PAPAPA 中国空姐干逼视频 观看美女自慰直播的网址 2017搜片神器是那个 强奸色女网 日韩人与兽爱爱 母淫网姐姐 亚洲娃娃 猛操骚货痴女护士文 亚洲大奶老女人 韩国电影熟女的欲望 偷窥厕所漫画 儿媳妇与公公三级片电影 好看站撸啊撸撸 首页偷拍自拍朋友女友的嫩穴毛还是初长成那种 李宗瑞生殖器大吗 肥乳性爱 有声小说ok网 笑傲华夏有声小说 春色夜生活 知春色 樱井莉亚下马 樱井莉亚体重 樱井莉亚123 小泽玛利亚迅雷 大家给个h网 东京热n0570 东京热n0393rmvb 可以用快播看的黄片 哪里可以看黄片 男人为什么喜欢看黄片 川村まや 33链导航 性生活知识 一点色导航 tv69在线 插插洞插插B 多情成仁高清 群交乱伦电影 色人居成人网 性乐汇色电影 淫色网 yeye撸 撸撸侠网站 苏紫紫干露露全套 干海参怎么发 Av天堂色电影 4099y影院 3d爆乳女教师西瓜 精品ady9 52AVAV好色色屌丝 国产精品情侣av 西野翔被偷窥 怡红院福利二区 在线阿v在线澳门 人人操美丽人妻 人人摸人人揉视频在线 在线无码私人影院 午夜福利1000部757 咪咪av电影 海外啪啪啪视频 南方来信免费观看 100国产自拍 rbxjdh 七色沙耶香番号 芹泽直美诱惑女教师视频 青青草原影视免费观看 桥本美步百度云 青草碰碰 青青草人碰人视频在线观看 在线看片 欧美自拍 青青 日福利av 日本情爱电影 孕妇在家孕交视频乳交 樱濑奈bt 日本少妇cosplay母狗 日本桐冈市 日本男女抽插插福利免费动态资势 在线男人福利免费视频 爱青岛在线观看 在线小萝莉 美女主播自慰出水视频 VEC-360部长奥-八木梓 麻仓优112部全集 迷夜惨案在线观看 痴女诱惑 稀缺资源 九州av 夜趣阁AV导航 无码av种子在线观看 luoliai 黄色片大胸 日本在线强奸乱伦 淫妇骚屄欠肏视频 色午夜v 小操黄色免费管看视频网址 日本亚洲欧美无码免费视频 草莓网 日本冒死厕拍视频 大香蕉一本道富二代 天天影院色小 凹凸視頻免賛在線 女主播叫床视频在线观看 行交视屏 黄色三级轮奸高清 宅男伦理 在线观看成人影院在线 xinxin68 org 国产自拍在线更新av 蓝沢润手机免费视频 你懂百度资源 和充气娃娃爱爱视频 龙头蛇尾 童颜巨乳 色和尚在线视频久和综合色色 江川美奈子在线视频 3344df 怡夏影院成人快播 1204手机在线播放下载 4438x一全国大色 露谷试看 色女孩影视最新网址 欧美群交网 波多野结衣来广州 日本she射精视频 自拍内裤爱液 伦理片自拍短视频 爱草色综合 彩美旬果xfplay 亚洲 自拍 porn 亚洲种子av 苍井空AⅤ下载迅雷 磁力链 下载 rki435 488成年网站 52av2av av天5 极品红人cutie queen 浪货成人网 91p app 上原亚衣黄色 中文字幕av12免m看 www路AV 亚洲36D 福利视频偷拍 大香蕉在緌人妻 福利avi 福利roufan动漫在线观看 大香蕉s视频 迪厅里的黑人男女裸体交配视颖 大神夜店大山极品美女带回高级寓所阳台一直 xxx-av 20147-20148 青鱼视频2自拍 国产人妖 鬼父有无码的么 韩国三级电台在线直播 劲暴欧美 一本道 东京热 自拍 国产 无码 quanqiuzuidadechenrenwangzhan 午夜小影院。 看看福利视频92 快播AV海量资源666 亚洲激情无码视频 俄罗斯群交之影院 义母吐息在线1mm 贵妃网伦理看天长网 自拍偷拍 微信小视频 526成人网 东方影库66 韩国三级在线 成人频道 永泽真央美医院BD视频中文字幕版 miae-183星奈爱 黑人 中国情侣自拍视频对白刺激 91色老板福利电影 自拍口交网 人与兽迅雷下载 大学情侣自拍视频在线 日本一本道高精无码 护士献身捐款视频百度云 欧美incest在线视频 老牛吃嫩草大尺度爱床上视频 wwwf994 哥哥搞在线播放 爱狠狼 58国产自拍在线 120214-749在线 aapp8988视频全集 成人国产视频导航 亚洲自拍91 嫂子内射 肥佬影视 福利制服诱惑在线 啪啪免费福利网站 1さつ册のみち道电影 青欲乐视频 视频操逼i 花间小视频 magnet 日本操逼小说 九州风云4电影下载 magnet 艺校妹子全裸淫语挑逗 台湾黄色毛片 破处mp4 下载 干到一半 偷偷换人的av 欧美极品AV束缚 协和影视 julia www11peng 色女香蕉 CAO96 gvg436在线播放 露脸怒草发廊高颜值小姐 www,637Df 干莉莉 我们立足于美国法律,受美国法律 小野遥 邪恶甄姬同人 小视频啪啪来电话女的说他免费 性感漂亮的女神级嫩模黑丝情趣装大尺度私拍视频曝光这身材样貌看的让人受不 新版spanking视频 性感网袜美女床上诱惑视频 先老弟影院 涩图片97 光棍影院,色戒 美丝馆福利视频 12级高清黄片 伊人谷姐干岳电影 a4yy午夜福利网神马 爸爸你的鸡巴太大轻点我好痛 主题宾馆偷情偷拍在线 缓存淫乱视频 2O19午夜福利合集 国产黄福利自拍在线视频 欧美x片视频在线观看 性爱伦理黄色一级电影 好美呀别停舒服死了 h极品卡通 胖老外酒店嫖妓白领兼职妹子鸡巴插着逼还手指抠屁眼听呻吟小姐吃不消 靠逼福利店 成人轮奸强奸网址 国产自拍大学城美女 啪啪啪青草视频免费观看 露脸超嫩的UU7后萌萌脸美女 在沙发上被大鸡巴不停抽插 天天橾夜夜橾bAV 淫妻迅雷 80s污动漫 火色院线 538日韩精品 京香JULIA ftp 超漂亮美女AV mp4 不知火舞无翼恋母 国产小视频2015 操我求你了好痒用震动棒干我好爽飞了高潮了哦哦啊啊奥视频 操教师在线 无码小说伦理电影图片 亚洲第一福利在线视频 椎名由奈无码 中文字幕 日本av播放器 户外女主播迅雷磁力链接 强插漫画 日本无码动漫AV迅雷下载下载 人体艺术6080强插漫画 悠悠资源共享 夜色邦另类图片 ipz-746中文字幕 泷泽萝拉福利在线 caoprom超碰vr专区 色尼姑免费AV 香港古装av无码种子 贵妇的沉沦女王反转版 段艺璇 大尺度 wwwmimikkk 床戏 体验区 avav177 一本道Th 多罗罗天下之战在线观看网址 vr天海翼视频在线观看 超乳爆乳巨乳影片福利 成人刺激抽插视频 S视频在线观看 车模聂子雨 板扎福利 UD-636R mp4 VIP私人玩物粉穴 草根艳舞团 县城全裸淫荡火爆 小刀电影伦理片 久久成人国产视频精品在线 在线视频播放你懂 御姐里番acg 色爱区在线视频网站 国模依诺娃与摄影师销魂视频 黑人大香蕉在线视频 操干网 高清大尺度丝袜老师 女主播资源公众号 无内丝袜福利在线 嫩模啪啪在线观看 最新国产自拍做爱 巨乳日本人妻還欠債被 538在线精品牛牛 黑皮书第一放映室 光棍影院最新福利在线看B 婷婷午夜成人视频 全裸性交 mp4波多野结衣下载短视频 国产视频58pao 我是检察官种子 磁力链 青青在线葡京视频 2018国产欧美在线va 日本啪啪视频无马赛克视频网站 欧美同性视频vibes 韩宝贝在线播放 韩国美女性解放与性交视频 国产主播特大阳具插入视频 韩国美女大胆生殖器23p 不堪欺辱的小林初花磁力链 国内在线自拍 国外另类大尺度福利网站 国产视频偷拍a在线观看 操逼视步 国产自拍丝袜 被强行占有第一次视频磁力 爆乳寡妇替欠债亡夫用身体肉偿 1曰本巨乳清木视频 有部小说男主在一个部门拿手机录下他上司和女员工做爱又拿录象和女员工做爱是 luoli写真 国产福利最新 蓝导航全福利导航 偷拍视频视频在线观看 刘可颖视频在线播放 美女翘臀肛交视频 亚洲香蕉视频在线播天堂网房 夜射猫精品av自拍在线 日本视频欧美一本道 国产大胸美女链接 mp4 2淫色人妻乱伦电影 上原亚衣 喷水无码在线 ibw518z ftp wwwjingziwou8 吉永尤利娅磁力 美女裸体直播迅雷下载 亚州无码中文字幕 www锛巃aakkk88锛巆n t午夜电影 东京热菜菜绪爱在线观看 rctd 3838pao,com 图激情五月婷婷丁香 欧美性爱大片黄 人妻众大香蕉视频在线播放了 女神一级毛片 www在线观看视频av草樱 luo女aa 射野之王在线 美国guolunlipianhuangshan 乱伦在线播放 裸聊直播在线免费观看 色老爹在线影院 鲁管电影 小视频 美国少妇与水电工 激情亚洲自拍偷拍 激情影院的祼体照片 精品小旅馆偷拍貌似很饥渴大白天偷情男女进来就迫不及待开草玩的姿势还不少 吉林市sm群 奸尸视频 magnet 拳交高清无码视频 街拍厕所 mp4 情侣激情爱爱自拍视频 极度高潮yyxf视频 AAAA快播 91金龙鱼最新-气质漂亮的良家少妇苏小小偷偷兼职做楼凤 一个更新快的色综合 av高清在线播放 偷窥自拍 清纯唯美 邱县特级黄片 男女做爱经典自拍免费 强奸乱伦影音先锋影片 免费观看爆乳家政妇 人猿泰山h版无删减mp4 小黄瓜资源站 91蟹哥 下集在线观看 黑丝袜裸体开档裤视频 欧美性交电影 黄图男在上女在一下 里番福利导航 97影院在第一页 男人透女人伦片 偷惰男女手机在线观看 手人人在线视频 阿v大香蕉性爱视频 朴妮唛在线视频全集 av在线,亚洲av 刺激影院亚洲hot 欧洲se 2艹高中女孩 男女做爱真实叫床视频 免费v片在线观看2320 哪里有a片资源 韩哥福利视频 男女抽插日p高清视频网站 バック体位で恥じらいのお尻まる出し下品な連続潮吹きjk 小島みなみ 日笔视频100部 jaⅴ颜射 友崎亚希 ftp 276ff这个网址换成什么了 大乳房美女奶头视频 一本道电影日本 自拍区偷拍亚洲最新777ap 椎名由奈 456 5388x最新更新 迷情影院一楼一凤首页 一级棒的美国j毛片 一本道日本电影天堂 91lu在线观看 普通话户外直播野战迅雷种子 东京干www12oa,me 三级黄久久免费视频 两性做爱高清写真 激情五月激情成人激情妹妹 90后人体艺术网 日本乱伦肛交灌肠颜射放尿 美女秘书加班被xo图片 xxx虐待阴茎视频 家庭乱伦都市激情校园春色自拍偷拍 丰满骚娘们视频 美女动态私处照 五月天97ss 日韩美女主播网站 欧美性爱歪歪色影网 911scc 亚洲性交30p 赶尸艳谭女角色办演着 全国成人色 美女的超短裙快播 很很累苍 xxxx333nm 美乳22p 美妞屄屄 丁香成人十次啦导航 绣由衣乱伦小说 狠狠撸操逼图片