prev <-> next
[2018-11-20 12:55:18] - a: That's why I don't have bonds. Not sure what advantage they provide over CDs and bonds have added risk. -Paul
[2018-11-20 12:54:53] - Daniel: I keep a fair amount of money (probably too much, honestly) in our checking account which gets 0%, so that's the trade-off, I guess. -Paul
[2018-11-20 12:54:08] - a: Yeah, I think I agree with Daniel. It depends on your risk tolerance. The market could drop 50% in a hurry, so it would suck if you lost your job AND your emergency fund got cut in half at the same time. If you're willing to take that risk, then I guess there is no problem. -Paul
[2018-11-20 12:39:41] - 2% jesus. that is so crazy high. like wtf, last time i checked savings accounts and CDs were both very sub-1%. like sub-0.5%. bonds don't seem nearly as awesome now if savings accounts and CDs are spitting out 2%. ~a
[2018-11-20 12:32:50] - Paul: Those are better rates than I get in my savings account but not a ton. Currently get 2% in my Amex online savings account. So probably yeah just depends on the liquidity vs return opinion. Seems fair. -Daniel
[2018-11-20 12:30:16] - a: Depends on your feelings and the if the volatility in the taxable account is a concern. Andrea is pretty risk averse so we keep a chunk of change just in an online savings account and pay for things like having a tree removed from our yard out of that instead of having to sell anything. I don't think its a huge thing though for it to be your emergency fund. -Daniel
[2018-11-20 12:26:42] - . . . i guess worded differently: can my vanguard account *be* my emergency fund? ~a
[2018-11-20 12:25:49] - paul/daniel: so is an emergency fund even necessary if most things can go on a credit card, and i can get money out of my (taxable) vanguard account in like ~1 or 2 days? ~a
[2018-11-20 12:17:24] - Daniel: Here are the APRs that I have on my CDs right now: 2.00% 2.35% 2.45% 2.65% 2.75%. I've found that those rates generally beat what I can find for online savings accounts, and I'm willing to sacrifice the slightly lower liquidity for the slightly higher returns. -Paul
[2018-11-20 12:14:33] - Paul: Whats your return on the CD ladder? You find it worth it to do that vs a online savings account? -Daniel
[2018-11-20 11:52:47] - a: And I guess I have like half a bitcoin for my pure speculation.
-Paul[2018-11-20 11:52:30] - a: I don't have any bonds, I transferred what few I did over to equities a few years ago. In terms of "investments" (and I don't quite know how I am using that term), I have my retirement funds 100% in equities, I have a fair chunk of our emergency fund in a CD ladder, I have a small bit of the emergency fund invested in what I consider safer equities outside of a tax deferred account... -Paul
[2018-11-20 11:39:24] - paul: so i have been (very minorly and slowly over the months) timing the market by moving bonds to equities during big drops. and slowly moving them back during spikes. probably not a great idea but it's what i've been doing. i don't have a lot of bonds, really (12%) but it does end up being a lot of money in dollars. ~a
[2018-11-20 11:09:25] - a: Yeah, it's definitely preferable to the alternative. I'm just jealous of the people on twitter who are excitedly talking about putting their cash balance to work.
-Paul[2018-11-20 11:06:10] - paul: hmmm. i don't think that's a bad thing though. i think it's better to have zero cash to buy on dips, than to have a buffer of cash to buy on dips. selling one thing to buy another thing should always be the norm? ~a
[2018-11-20 11:01:33] - a: Yeah, that's the worst part of being fully invested at all times: Having to sell something to buy on dips like this. -Paul
[2018-11-20 10:57:51] - paul: well, honestly, you didn't say not to buy nvda. you just said it wasn't on your list of things to buy. also, i think i had already decided not to buy nvda anyways: since i'd probably have to sell something else to buy nvda, which i didn't want to do. ~a
[2018-11-20 10:54:07] - a: Heh, and NVDA is up 4% right now, so goes to show you shouldn't listen to me.
-Paul[2018-11-20 09:53:21] - a: Yeah, I don't think we'll be dropping 50% like back then, but it feels like we could easily drop more from here. I'm kinda restricting myself to making one move a week right now because I don't want to overreact to what has so far been a pretty minor drop. -Paul
[2018-11-20 09:44:42] - paul: agreed. if you go into logarithm mode, and look back to the crashes in 2001-2002 and 2008, this is nothing. each of those drops were *over* 50%. we still haven't reached 10%! ~a
[2018-11-20 09:30:55] - a: It's interesting because as bad as this drop has felt, we're still up for the year and looking out over 5 years, this still seems like a tiny blip. I think I'm "feeling" it worse because I'm acutely aware of how bad the first Freedom Portfolio check-in is going to look.
-Paul[2018-11-20 07:24:14] - Really wish bitcoin wasn't moving in concert with the market, though. Too many things that look appealing to buy! -Paul
[2018-11-20 07:23:19] - a: Geez, looks like another brutal day is shaping up today. I've already made a few moves to sell positions that haven't dropped as much to load up on some stocks (SQ) that have. I probably am going to see how the rest of the week goes before doing anything else. -Paul
[2018-11-19 16:39:06] - a: Oof, man, now is the time to buy anything.
NVDA isn't tops on my list of things to buy (too many other things I like as well), but it's hard to deny that a 50% sale looks appealing. -Paul[2018-11-19 16:10:10] - is now the time to buy nvda? ~a
[2018-11-19 11:38:11] - a: I think it's all NVDA. They were on fire until this most recent quarter. -Paul
[2018-11-19 11:30:26] - which we can actually look at . . . it's not hard for us to use march 1st. tmfc would still be beating us in that case: you're at +3% and i'm at -4% (wow, i guess i had a really great first quarter). ~a
[2018-11-19 11:27:41] - paul: i don't know if that matters, much: i think i agree with you in general that this is apples and apples. but digging deeper, if your beta isn't exactly the same as the s&p500 beta, i imagine it's not fair, unless you use the same starting date. ~a
[2018-11-19 11:16:03] - a: Maybe. By March 1st, wasn't the S&P pretty much back to where it was around Jan 1st? -Paul
[2018-11-19 11:13:57] - paul: well . . . tmfc didn't exist on 2018-01. they missed out on the shitty january/february we all had. ~a
[2018-11-19 10:25:21] - Sneakily, though, TMFC (the Fool 100 index) is beating us all so far, it looks like. -Paul
[2018-11-19 10:24:31] - a: I've taken the (unlikely) lead in 2018! Mixed feelings, since I have NVDA in the Freedom Portfolio so its fall has hurt me too. The important thing is that we're both crushing Daniel, though. -Paul
[2018-11-16 17:47:47] - https://www.nbcsports.com/washington/capitals/capitals-fan-travels-antarctica-remind-actual-penguins-who-stanley-cup-champions-are
[2018-11-16 16:58:55] - wp page already ~a
[2018-11-16 16:46:51] - a: the due process angle is something I hadn't considered, and the precedent seems pretty tight nit. - mig
[2018-11-16 16:26:30] - acosta has press pass returned . . . ordered a judge that was appointed by trump. ~a
[2018-11-16 15:00:50] - Yeah I was sad by all the prices back then too. I'm sure its worse now if they still haven't gone down. -Daniel
[2018-11-16 14:54:19] - I'm seeing a GTX 1070 for $370 and that's an amazon special (MSRP $500!). Have no clue how these cards haven't depreciated by now. - mig
[2018-11-16 14:53:17] - daniel: i'm set to replace mine, haven't been bitten much by the .5gb bug but feels like with what's coming out eventually I'm going to start running into those issues. 4 years seems like a good run though for it. - mig
[2018-11-16 14:38:43] - mig: My 970 died and I moved to a 1050 because that was what I was able to get for not an arm and a leg back in April. -Daniel
[2018-11-16 14:26:41] - https://github.com/minimaxir/big-list-of-naughty-strings/blob/master/blns.json#L499 here's a list of naughty strings which do things like XML injection, javascript injection... and line #499 is, i guess psychosis injection? i really want to know the backstory behind that one - aaron
[2018-11-16 14:21:21] - xpovos:
yeah it was hilarious seeing how far people took it. when i first heard the rules i just thought, "oh so people will i don't know... just get really paranoid when someone's in range of their tank? or they'll try to move out of range of everyone?" it really threw me for a loop - aaron[2018-11-16 14:12:23] - https://www.nebula.org/?utm_source=twitter&utm_medium=social&utm_campaign=launch-announcement This seems interesting, although I already got my DNA sequenced with 23andMe... -Paul
[2018-11-16 13:31:03] - daniel: are you still on the gtx 970? - mig
[2018-11-16 13:24:29] - aaron: I'm enjoying the video so far. -- Xpovos
[2018-11-16 13:20:13] - Speaking of nvidia gpus, seems like there's still pricing wackiness. The 10 series has been out for around 2.5 years and the cards still haven't seen any real price drops even with the next gen 20 series dropping 2 months ago. - mig
[2018-11-16 11:31:59] - a: Yeah, I can't tell you much beyond that. I just felt like that demand was bound to fall and so NVDA might see some flat performance going forward. It helps that it had already more than doubled for me, so I felt good taking some profits. Didn't think it was going to drop as much as it did. -Paul
[2018-11-16 11:23:09] - paul: hmmm thanks that's interesting. they don't give any specifics to which cryptocurrency, so i have to just be surprised that there are that many people are stupidly using graphics cards when there are two generations of systems better suited for cryptocurrency mining: fpga and asic. ~a
[2018-11-16 11:13:53] - a: https://www.techradar.com/news/nvidia-says-gamers-come-first-as-it-moves-against-miners-bulk-ordering-graphics-cards But here is an article from earlier in the year about it. -Paul
[2018-11-16 11:13:05] - a: Ah, sorry, guess my time period was off. Forgot it was December of last year. Let's say 18 month or so ago instead. -Paul
[2018-11-16 11:11:58] - paul: i'm not entirely sure what you're talking about
the big bitcoin run-up wasn't earlier in the year. do you have a link i can see? i'll admit that there were probably people buying graphics cards way after they were profitable, but 2018 seems a little weird? ~a[2018-11-16 11:09:55] - a: I'm not entirely sure what you're talking about, but I know a lot of the chatter around NVDA and AMD during (and even after) the big bitcoin run-up was about how gamers couldn't get their cards because miners were buying them all up. -Paul
[2018-11-16 10:58:22] - https://www.youtube.com/watch?v=t9WMNuyjm4w a game designer developed a tabletop battle royale game. you could only make one move per day, unless you saved up moves from previous days, or people gave you their moves. it got to where alliances and strategy meetings consumed 75% of people's workday and then they stopped the game - aaron
[2018-11-16 10:02:53] - aaron: god even if they were, i think the algorithm would cross-check against actual words. ~a
[2018-11-16 10:02:25] - aaron: none of the inner parts of words or names themselves are algorithmicly generated? ~a
[2018-11-16 10:00:46] - paul: "crytpocurrency mining craze" and "earlier in the year" sound weird to me. even scrypt mining stopped being profitable with graphics cards three or four years ago. ~a
[2018-11-16 09:57:07] - mig: that's crazy that something like "farmer of racism" could make it onto a card... i'd understand a combination like "karnath: devourer of orc meat" where each part is independently innocuous but someone would have had to explicitly OK the word "racism", why would they do that? - aaron
[2018-11-16 09:52:38] - a: Yeah, I had cut my holdings in half earlier in the year. I was worried they were getting too caught up in the crytpocurrency mining craze and it would come back to hurt them later. Didn't think it would hit this hard, though. -Paul
[2018-11-16 09:49:52] - paul: in other good news, i've been selling a lot of my nvda (sold twice above the current price). i only have small exposure. ~a
[2018-11-16 09:48:10] - paul: yeah i had a strong feeling that the 2018 challenge wasn't wrapped up yet. didn't forsee this exactly, but i knew that we were only a few percentage points away from eachother a few days ago. ~a
[2018-11-16 09:44:14] - a: And by strong, I kinda mean the opposite. Sorry if I inadvertently got your hopes up. It's in the freedom portfolio too so I feel your pain. -Paul
[2018-11-16 09:43:31] - a: Wow! NVDA coming in strong today to put the 2018 back within reach for me! -Paul
[2018-11-15 15:24:21] - aaron: https://techraptor.net/content/unfortunately-named-keyforge-decks-are-being-recalled this sounds like a familiar problem. - mig
[2018-11-14 21:01:06] - https://www.realclearpolitics.com/video/2018/11/14/trump_announces_criminal_justice_reform_redemption_is_at_the_heart_of_the_american_idea.html Seems relevant to an earlier discussion about Trump and criminal justice reform. -Paul
[2018-11-14 21:00:42] - a: I didn't edit anything, so I should be safe. Believe it or not, I randomly stumbled onto the page because I was curious about his age and what films he had been in. -Paul
[2018-11-14 17:29:09] - paul: slate has a different take on crystal city: that it sucks. therefore: "Crystal City is a perfect place for a corporate behemoth like Amazon to remake in its own image"
~a[2018-11-14 16:56:00] - paul: i guess since it's relevant, i should mention the 3RR. you can be banned (1 day ban) for reverting one article too much in 24 hours. they're pretty strict about it. ~a
[2018-11-14 16:54:06] - paul: according to miller himself, he is fine with he/him/his. ~a
[2018-11-14 16:52:46] - there have been nine edits today. mostly regarding pronouns. so . . . i'm curious, how you found that article? probably from the same place as a few other people maybe? ~a
[2018-11-14 16:51:05] - also i should probably mention that there has been some discussion on pronouns on the talk page. "yeah, good luck with that", heh. even an edit war apparently. ~a
[2018-11-14 16:50:15] - a: Oh, that's also the problem. I have zero confidence that I am right and would be changing it to something better. In fact, I'm more on the side that I am wrong somehow, but I just don't know how. -Paul
[2018-11-14 16:49:12] - well you're free to change it. especially if you think you're changing it to something that is more readable and is more correct from a "modern language" (but not bleeding edge) perspective. ~a
[2018-11-14 16:47:16] - a: Uh... I was more confused by the fact that it changes throughout the article. Not sure if that's "right", either in terms of wikipedia style or what Miller prefers. -Paul
[2018-11-14 16:45:04] - paul: i'm not sure if i answered your original question. i think most people would say to use the pronouns that are modern. which is to say, do whatever you think miler would be most comfortable with? what does miller prefer? it could be that miller would prefer you use the gender that is associated with the time the article was referring? i'm guessing though. ~a
[2018-11-14 16:37:51] - paul: more generally, you're referring to a manual of style. above the manual of style is even more general stuff: policies and guidelines. or, if you want to learn more there's the general introduction or five pillars might help. ~a
[2018-11-14 16:18:34] - paul: Wikipedia:Gender-neutral language. ~a
[2018-11-14 16:11:05] - a: https://en.wikipedia.org/wiki/Ezra_Miller Question about Wikipedia... etiquette? That's not the right word. Maybe its not even a Wikipedia thing. Anyway, it looks like WP goes out of their way to use "they" as Miller's pronoun in the latter half of the article, but uses "he" earlier on. Is... that right? -Paul
[2018-11-14 16:01:31] - aaron: So I don't know but probably several? -Daniel
[2018-11-14 16:01:18] - aaron: I just get my burrito's with extra red sauce. -Daniel
[2018-11-14 16:00:40] - for those of you who use hot sauce on taco bell tacos/burritos, what is the correct hot sauce packet to burrito ratio? - aaron
[2018-11-14 15:15:51] - aaron: I did. I always made the wrong decision.
-Paul[2018-11-14 14:43:00] - did anybody ever play Papers Please? the morality puzzles of this web site sort of remind me of decisions in certain pivotal decisions in Mass Effect and Papers Please - aaron
[2018-11-14 14:35:55] - reprogram the cars!
~a[2018-11-14 14:34:12] - xpovos: yeah, i had the same thing happen -- with only 13 questions, it turned out i coincidentally spared every single fat person, and ran over every jogger. so they deduced like, "wait, wait, this guy HATES fitness! reprogram the cars!" - aaron
[2018-11-14 14:15:52] - a: When I took it, it turned out that all of my choices were to also save the person not violating the law, but that was coincidental, a side of effect of other more prominent choice branches. -- Xpovos
[2018-11-14 14:14:28] - daniel: it shouldn't have been part of your calculus!
~a[2018-11-14 14:05:01] - Daniel: It's probably not something I would consider in a split second decision, but it's definitely something I think would sway me if I had time to think about it. -Paul
[2018-11-14 14:03:49] - I'm running the car into the barrier on part 2. -Daniel
[2018-11-14 14:03:21] - Whats partly funny is that I didn't even notice the signs to indicate jay walking or not. It wasn't even part of my calculus. -Daniel
[2018-11-14 13:55:50] - The dog is jaywalking! And you can certainly not fault a cat for faulty brake lines. I say the dog gets it - aaron
[2018-11-14 13:53:38] - 2 animal passengers injured vs 1 animal pedestrian dead? i choose 2 animal passengers injured. ~a
[2018-11-14 13:49:29] - http://moralmachine.mit.edu/hl/en/browse/-635155056 part 2 - aaron
[2018-11-14 13:49:22] - Yeah, I really love the trolley problem (or, in this case I guess, the MIT site) because it drives me crazy. I can't come up with a good, consistent moral compass to use. There's no right answer! -Paul
[2018-11-14 13:47:01] - aaron: hitting the dogs as well.
~a[2018-11-14 13:45:48] - aaron: Hitting the dogs! Sorry dogs. -Daniel
[2018-11-14 13:45:18] - Its interesting that we all came up with differing rules for the mit site. I mean thats the point I suppose but nonetheless interesting to read them. -Daniel
[2018-11-14 13:40:58] - http://moralmachine.mit.edu/browse/-1598914774 my scenario: would you swerve to run over two dogs if they were obeying traffic signals? - aaron
[2018-11-14 13:30:18] - aaron: "i would not swerve to run over a citizen who is doing nothing wrong" here we differ. really the person making the legal real mistake was the car with the broken brake lines. so, i decided to ignore lawfulness when deciding who died. ~a
[2018-11-14 13:25:57] - most of my decisions came down to whose fault it was, and punishing the people at fault, which is probably heartless. but it seems horrible to think of like, "well why did so-and-so die?" "oh, he standing on the sidewalk when 30 assholes all jaywalked nearby, so he was run over by an autonomous vehicle which swerved onto the sidewalk" - aaron
[2018-11-14 13:23:58] - also that moral machine site was interesting. when it comes down to "run over the 1 lawful jogger or the jaywalking family of 5" i definitely value following traffic laws above all else. if that family of 5 is jaywalking it is cruel but i would not swerve to run over a citizen who is doing nothing wrong - aaron
[2018-11-14 13:22:23] - a: yeah that's my logic, a DUI will not necessarily end two lives but it will like, end 1.25 lives or so. (average DUI driver probably has about 40 years remaining, and average prison sentence would probably take away about 25% of that.) and yeah injuries would need to be included too, i agree that's a factor - aaron
[2018-11-14 13:17:32] - aaron: ignoring my link though, i still don't think i get what you're saying. i think you're saying that a DUI will cause the end of two lives: both the driver and the victim are fucked. whereas with the autonomous car: only the victim is fucked. is that your logic? if that is your logic, then i think we would first need to include *injuries* before we include "ruined lives". ~a
[2018-11-14 13:14:35] - i might have those numbers slightly flipped for what i'm trying to say, but i guess my point is you're weighing "killing a random person" against "killing a random person and punishing a negligent person" so it's weird. maybe if autonomous vehicles are less safe, it's still an OK bargain, but i doubt people would ever accept that - aaron
[2018-11-14 13:13:06] - aaron: yeah in the mit link (that just scrolled off the page), they have you "value" human life. i decided to not go that route: each human is exactly equal. hell, the MIT link even lets you value males and females separately
~a[2018-11-14 13:11:25] - i think humans being humans, they'll probably take the 2% over the 1%, rationalizing something like, "oh who cares about drunk drivers, those assholes belong in jail, autonomous cars aren't safe enough yet." and there's probably merit to that opinion, but it's messy - aaron
[2018-11-14 13:10:10] - so what's worse; a 1% chance of killing a guy, or a 1.2% chance of killing a guy and also ruining someone's life. or what if it's 2%, but the people whose lives are ruined are sort of negligent pieces of shit who deserved it probably, but they're still humans - aaron
[2018-11-14 13:09:04] - it's also a weird math problem because if you think of a lot of accidents, they involve deaths (the people who are hit or the passengers) and often jail time or life-altering consequences (someone can no longer drive for the rest of their life, or has to spend 10 years serving a prison sentence for DUI) - aaron
[2018-11-14 13:05:13] - a: Damn my assumptions! Unrelated, where would we be without Wikipedia? -- Xpovos
[2018-11-14 13:01:15] - xpovos: no, actually all three of the tesla autopilot deaths have been passengers. wp. ~a
[2018-11-14 12:47:43] - a: Good data! I'm surprised by these numbers, somewhat. More by how low the human data is, in fact. I think it goes to show how well-engineered the cars are. Most human-driven deaths result from vehicle-vehicle collisions. Those have been make amazingly safe considering the forces involved. The autopilot deaths seem to be vehicle-pedestrian collisions. Pedestrians have no such engineering. -- Xpovos
[2018-11-14 11:19:39] - palu: ah, oops, i guess tesla did publish the number of miles driven by their autopilot: no, tesla autopilot (as of 2016) killed more humans per mile (1 human per 105m km) than humans did (1 human per 160m km). ~a
[2018-11-14 11:17:35] - paul: humans kill about 1 person every 160m km. there have been four deaths due to self-driving cars. but since we don't know how many miles tesla autopilot have driven, and we don't know how many miles uber testers have driven, we can't know the accuracy of self driving cars. ~a
[2018-11-14 10:11:03] - I haven't seen any good studies on how safe they are yet. I know Elon Musk had a tweet some months ago saying it's already safer, but there was some push-back on that and he's known for hyperbole. Still, humans are notoriously bad drivers, so it wouldn't surprise me if we're close to that point already. -Paul
[2018-11-14 10:10:14] - While I am interested in knowing why the car failed in this specific instance, I'm more interested in seeing how the safety of self-driving cars compares overall to the safety of human drivers. There will always be one off accidents, but if self driving cars are twice as safe, I think the trade-off is worth it. -Paul
[2018-11-14 09:47:54] - a:
- aaron[2018-11-13 20:47:06] - Yah I would too. We're left guessing on a lot of the details. ~a
[2018-11-13 18:01:24] - a: Right. I'd love to know more, but I don't think my source gave much information on why the algorithms chose not to do those things. Presumably because of other erroneous decisions. E.g. it thought she was a bike/bicyclist for a while. Therefore it made assumptions about her rate of speed, which when it turns out to be not true, it failed to take additional actions. WHY? That is critical, and it's in the software. -- Xpovos
[2018-11-13 14:58:32] - from reddit: "this [marijuana] is too strong, I smoked too much and hallucinated that Ryan Reynolds was gonna play Pikachu in a live action Pokemon movie"
~a[2018-11-13 14:54:34] - xpovos: i follow now, thanks. i still think we're ignoring a few other problems though: emergency-braking wasn't the only issue. it saw the human for 6 full seconds. it didn't need to emergency-brake. it didn't even need to swerve. it needed to slow down and change lanes and it did neither of those things. it didn't slow down (the 9% is close to a negligible-change given the huge amount of time it had). that's so fucked up. ~a
[2018-11-13 14:47:24] - a: Emergency braking by an autonomous vehicle might result in a higher rate of accidents from human driven vehicles nearby that aren't expecting emergency braking and are themselves driving unsafely because they usually get away with it. That kind of a problem is less pronounced if a majority (plurality?) of vehicles are all algorithmically driven. -- Xpovos
[2018-11-13 14:42:09] - xpovos: can you further explain the "say 65%" thing? i don't follow that part. ~a
[2018-11-13 14:39:41] - better before it can be implemented widely. We don't seem to have the data. Which is why I had that (in testing) question, below. If it's just for testing and generally good, that's a human intervention problem. If it's because otherwise it causes problems, and might continue to cause problem until there's a critical mass of autonomous vehicles (say 65%), then it's not actually solved yet. -- Xpovos
[2018-11-13 14:38:32] - Daniel: The reason why the emergency brake was turned off was because of the potential for misreads. So the ultimate question is: are emergency brake incidents as a result of misreads a serious and regular problem? If they aren't, then it seems we have a pretty good algorithm and it's ready to move forward. If there are too many emergency braking incidents, for trashbags, as the example cited, then the detection algorithm needs to be [..]
[2018-11-13 14:38:14] - daniel: IMO, yes that counts as algorithms. though, to be fair to you and paul, i think it's an algorithm failure that can be easily remedied. there will always be "configurations" and "testing" and "turned off systems" that can cause deaths. the lasting question i think will be: how long (how many car-hours or car-km) until another human is killed? ~a
[2018-11-13 14:34:28] - a: I think Pauls point (if I'm understanding) is that the algorithms were good in that they identified her and wanted to brake but some engineer who was doing testing turned that off, so I guess the question is does that count as part of the "algorithms". -Daniel
[2018-11-13 14:33:28] - paul: yes, i think we all agree on the majority of these things. if self-driving cars are better than average-humans, than we need to move forward for the better of everybody. the part where we might disagree is if we've reached that point today. even if you completely ignore arizona and herzberg, i might still not feel like we've reached that point. ~a
[2018-11-13 14:32:06] - xpovos: "Is the set of algorithms driving the vehicle better than the average human driver?" if you're talking about uber using the configuration the had in arizona, than i say emphatically no. ~a
[2018-11-13 14:29:38] - We know that the least safe part of a car is the driver. I'm all for removing that from the equation as soon as possible, as long as it's not making things actively worse. People will still die in car accidents, but fewer, and that's good, right? But the issue of liability is a major concern. -- Xpovos
[2018-11-13 14:29:12] - a: I think we agree, but we're quibbling over semantics. The self driving car only failed because the emergency braking was overridden by an engineer. It identified the problem and wanted to break. In my mind it was human error that overrode the car. -Paul
[2018-11-13 14:28:44] - a: I think the questions I have are still basically the same. Is the set of algorithms driving the vehicle better than the average human driver? If not, what needs to be improved, still? If so, what can we do to further improve it so that it's better than any human driver?--while still implementing it as widely as possible. -- Xpovos
[2018-11-13 14:20:38] - maybe some day we'll get to good self-driving cars, but if i've learned anything from that wired article, 2018 is not that time. ~a
[2018-11-13 14:16:58] - xpovos: thanks for that link. so much to unpack there. though, i don't think i agree with paul: that we should let the self driving car off the hook. in the end, it didn't brake. it didn't swerve. it didn't change lanes. it didn't even notify the safety-driver. (basically because it was instructed *not* to do these things) ~a
[2018-11-13 14:10:33] - Xpovos: Heh, so the self driving car actually WAS superior to the human driver even in this instance? -Paul
[2018-11-13 14:09:38] - a: Did we ever find out what happened? Did it think it was a bird or something or some imaging ghost? -Paul
[2018-11-13 14:09:33] - It goes to show that in the end, any system is only as robust as the humans who program it. -- Xpovos
[2018-11-13 14:08:44] - a: https://www.wired.com/story/uber-self-driving-crash-arizona-ntsb-report/ The software determined it needed to emergency brake, but was prohibited from doing so by an overriding command because they want to avoid emergency braking for misreads (in testing?) and were relying on the human operator at that point. -- Xpovos
[2018-11-13 13:57:01] - paul: the thing about that crash, that blows my mind, is that it had thousands (millions?) of data points referring to an object right in front of it, and it still continued straight ahead. talk about a total breakdown of engineering. ~a
[2018-11-13 13:48:59] - a: nm, you just addressed that.
-Paul[2018-11-13 13:48:24] - a: I assume you're referring to the Uber car crash? Isn't it generally agreed that that was an accident and not an intentional act? There was no reason for the car to have hit her had it recognized she was there. -Paul
[2018-11-13 13:44:03] - anyways, the herzberg situation is totally different, sorry to confuse things. i doubt it decided hitting her was the best solution: it just didn't get to the point of actually making that decision. ~a
[2018-11-13 13:41:48] - it'll be rare. but because of the number of miles/day we're talking about across the world, it'll come up enough that it makes sense to work out the kinks. ~a
[2018-11-13 13:40:38] - paul: well, let's ask herzberg. ~a
[2018-11-13 13:33:56] - a: I would hope that situations where a self driving car "has" to consider hitting pedestrians is pretty rare. It's hard for me to imagine a non-rare scenario where that's preferable to rear-ending another car or driving off into the grass or whatever. -Paul
[2018-11-13 13:32:11] - paul: one of my coworkers was in a car with brake failure. he down-shifted (even though it was an automatic car, you can switch to "2" and "1" even when the car is moving), and turned uphill. nobody was hurt. i hope the self-driving car will know that hopping up onto a curb or driving off the road is better than hitting a pedestrian even if it means damaging itself. but, after the herzberg situation, i'm not hopeful. ~a
[2018-11-13 13:31:36] - paul: which is why MIT is making this game
~a[2018-11-13 13:30:25] - a: Probably, but it's such a grey area that it feels impossible to legislate. This simplifies things because of 100% fatalities (it sounds like), but that's a huge simplification. What if it's 60% chance of passenger death and 40% chance of pedestrian death? I don't envy whoever has to make those decisions. I have no good answers. -Paul
[2018-11-13 13:26:36] - or more likely the legislators. ~a
[2018-11-13 13:26:17] - paul: which is why it'll probably be left up to the courts. ~a
[2018-11-13 13:24:56] - There was a good article before that I thought I posted here but maybe I didn't. It talked about if people would want to ride in self driving cars that "prioritize" pedestrian safety over passenger safety and if car companies might advertise if their cars didn't. -Paul
[2018-11-13 13:22:20] - daniel: yes that's interesting, but for the purposes of the game we've been guaranteed 100% fatality. if fatality numbers are <100%, then we need to be given that information. i still think pedestrians > passengers in a tie though, because passengers chose to get in the high speed death-machine, and pedestrians only chose to cross the street. ~a
[2018-11-13 13:20:18] - my own internal rules: gender, age, fitness, and "kind" of humans do not matter. rule of law does not matter. 1 animal = 1/1000th of a human (animals do not matter unless there is a tie in the human count). pedestrians = passengers (unless there is a tie, then give preference to pedestrians). non-intervention > intervention only if there is a tie. ~a
[2018-11-13 13:19:45] - Cars hit barriers not pedestrians. Cars have seat belts and air bags. People on sidewalks don't. That was what I put in the end. I pretty much choose car into barrier on all choices I could over pedestrians. -Daniel
[2018-11-13 13:19:24] - yeah, exactly. except they turned it into a "game" that is collecting data about what people believe, and that could potentially affect how the trolly's are actually governed
~a[2018-11-13 13:08:40] - a: The trolley problem! I love that. -Paul
[2018-11-13 13:02:41] - http://moralmachine.mit.edu/ . . . watch the video first. it has you imagine brake failure on a self-driving car, and you must help the self-driving car decide who lives and who dies. brings up a lot of really interesting scenarios, where a human wouldn't have enough time to weigh the results of a split-second decision, but a self-driving car AI would have enough time to weigh the results. interesting! ~a
[2018-11-13 12:38:49] - Going for NYC puzzles me honestly. - mig
[2018-11-13 12:33:19] - regarding the discussion on infrastructure, here is info on the potomac yard metro station which looks pretty active. i also forgot that crystal city effectively already has four metro stops if you count the airport and the bus station at pentagon. five metro stops would be adequate for amazon, imo
~a[2018-11-13 11:07:30] - on the other hand, maybe there will be unexpected effects: recruiting would actually get easier (and our salaries will not go up) if a bunch of tech people move to our area. ~a
[2018-11-13 11:05:50] - paul: yeah maybe, depending on how many people get hired. we'll also probably also see a bump in salaries and hire-ability. the only big thing that will suck will be the roads. and recruiting will get harder. and (imo) the tax breaks will suck. ~a
[2018-11-13 10:56:01] - aaron: I heard about it and assumed it was a joke. I'll have to watch later. -Paul
[2018-11-13 10:55:46] - a: Yeah, apparently you are right. Maybe I'll still get a bump in housing value.
-Paul[2018-11-13 10:55:29] - aaron: I hadn't watched it till just now. I'm actually a little intrigued which I did not expect at all. -Daniel
[2018-11-13 10:51:05] - also has anybody seen the new "detective pikachu" trailer? i haven't seen a pokemon movie since 2000 but i might have to make an exception, it looks really well done. i thought it was some kind of a joke when they said ryan reynolds was voicing Pikachu but then it turns out they were just going for a very, very weird kind of movie - aaron
[2018-11-13 10:50:56] - paul: my guess, based on their decision, is that "housing prices" is not *actually* a priority. expanding more in seattle doesn't get them the extra talent/connections they'll find in dc+nyc. ~a
[2018-11-13 10:47:28] - mig: I suppose. I guess I just thought that inexpensiveness in terms of housing prices was a higher priority. Not sure the point of having a second (and third) HQ in NYC and DC when you could've just expanded in Seattle. -Paul
[2018-11-13 10:24:09] - paul: I think it makes a whole lot of sense. The area still has a lot of room for growth. It also has metro access now and they can pull in potential workers from DC and MD at that location. Even when the silver line does finally reach dulles, it's not a convenient trip for anyone outside of VA. - mig
[2018-11-13 10:18:16] - a: Burn him!... er... You! -Paul
[2018-11-13 10:14:27] - so it's all my fault hq2 is in crystal city. ~a
[2018-11-13 10:13:22] - the reverse commute on 66 has always been a problem too. i remember talking to somebody on reddit about the reverse commute on 66. that'd be funny if they were actually an amazon exec pretending to be a regular joe. ~a
[2018-11-13 10:13:18] - a: I thought one of the big motivations was that Seattle was getting too congested and expensive. Plopping down in the middle of Arlington and New York doesn't seem to fix that.
-Paul[2018-11-13 10:12:41] - a: Sure, and admittedly I don't know much about Crystal City. I just don't see any particular benefit to being SO close to DC. And as not-crowded traffic is and cheap housing might be, I have to imagine both would be even better out west. They have a superior airport in Dulles and the Silver Line will eventually make it out to where they can still access DC. -Paul
[2018-11-13 10:12:15] - it's a balancing act i'm sure. pick an area too far out from the city, and nobody wants to live there. pick an area too close to the city, housing prices and commutes start to suck. ~a
[2018-11-13 10:10:48] - i lived in crystal city for six months (recently), and i think crystal city is a great area for hq2. the kids will love being so close to dc without having to pay dc prices. ~a
[2018-11-13 10:09:21] - paul: knowing a little bit about crystal city might help. there's a shit-ton of recent infrastructure improvements in southern crystal city close to potomac yard (and in potomac yard). there has always been talk about adding a metro in potomac yard, i'm not sure if that's still on the table or not. new housing projects have exploded in that area. also crystal city has never seemed (to me) to be an overly-busy area during rush-hour. ~a
[2018-11-13 10:03:57] - I've always thought Northern Virginia was a good match, but I don't get the Crystal City location. Further out west around the Dulles/Sterling/Ashburn/Leesburg area always seemed like a much better fit. -Paul
[2018-11-13 09:59:18] - https://www.cnn.com/2018/11/12/tech/amazon-hq2-new-york-virginia/index.html amazon's officially coming to northern virginia - aaron
[2018-11-13 09:39:05] - well i guess we'll see if november and december end up being recovery months or not. ~a
[2018-11-13 09:15:58] - a: I get what you're saying, but Mazor did even worse. They had hit $76 a share in the past year and ended up getting bought at $58, which isn't very close. -Paul
[2018-11-13 08:12:36] - the 50% yesterday is a reversion to previous (pre-recession) prices. the 2018-07 price was 36; the 2018-10 price was 37. i'm surprised they could get such a good deal in this recession, though. ~a
[2018-11-12 14:48:50] - a: I think the APTI buyout solidifies another stock market challenge win for you. The 60% gain there dwarfs my 13% gain for MZOR. -Paul
[2018-11-09 23:56:07] - a: No need to apologize. I wasn't offended that you found out my secret plan to convince you that all bicyclists must die.
-Paul[2018-11-09 13:31:50] - I think the difference lies between perfect justice, which humans can't achieve, and objective justice, which we try, and still often fail. For ojective justice we do have to punish outcomes rather than intents. -- Xpovos
[2018-11-09 11:43:30] - sorry paul
~a[2018-11-09 11:20:08] - "It's really hard for me to blame somebody too much for what is an accident and not intentional" not for me. people are reckless every day, and one day, their time comes up. the intent was in the recklessness. they intended to be reckless. this is a perfect example of recklessness too: he was in the middle of running a red light when it happened. ~a
[2018-11-09 10:58:01] - paul: you've voiced that you think pedestrians and bikes should not be on the road, because you're not expecting to see them. and that's a safety issue. now there's a bicyclist that's been killed by a car, and you're wondering why the driver is getting punished more harshly than a person with the same intent, but didn't kill a cyclist. sorry if i'm seeing dots that aren't there. ~a
[2018-11-09 10:56:05] - paul: no, i doubt we are. ~a
[2018-11-09 10:55:45] - a: *Sigh* Nevermind. I don't think you and I are on the same page here. -Paul
[2018-11-09 10:54:43] - a: And it feels similar to this, in that to me it involves weighing outcome versus intent a bit. I kinda think Mike Vick was worse because of the intent. It's really hard for me to blame somebody too much for what is an accident and not intentional. -Paul
[2018-11-09 10:54:38] - paul: you think peoples should get a less harsh sentence because it isn't his fault that hollowell died? i mean, hollowell was an old dude and was probably going to die anyways? i feel like the "90% agree with you" is top-cover for the fact that you think hitting a dodgy bicyclist who just happens to die, isn't really totally the red-light runners fault. or at the very least he shouldn't be punished. what exactly are we punishing? ~a
[2018-11-09 10:54:17] - apples and oranges? ~a
[2018-11-09 10:53:36] - a: And Dave and I were discussing the differences in reactions and punishments. Is it "worse" to have completely and accidentally killed a human? Or to have very purposefully and intentionally killed dogs? -Paul
[2018-11-09 10:52:29] - a: The latter was not a hit and run, the player took full responsibility and was super remorseful from the get-go. I believe it was also a situation where the pedestrian ran out from between two cars in the middle of the night... -Paul
[2018-11-09 10:51:31] - a: I remember talking to Dave awhile ago and we were comparing Mike Vick (purposefully and intentionally ran a cruel dog-fighting operation) and.... some other football player who struck somebody with their car and killed them. -Paul
[2018-11-09 10:50:29] - a: Honestly, I'm not even really talking about practical cases, I'm more talking ideologically speaking: What exactly are we punishing? -Paul
[2018-11-09 10:49:29] - a: Or it works the other way, what if somebody gets hit so hard that it would normally kill like 99% of people, but this person happens to go flying into.... I dunno a pillow factory and happens to live. Should they luck out? I guess since we're talking outliers, it's not a bit deal... -Paul
[2018-11-09 10:47:37] - a: Well, situations like that where (from information available) it seems like all else is equal, I'm more okay with a difference in punishments. It's what are probably admittedly outlier cases. Maybe somebody suffers a bad, but not life threatening injury, but the ambulance takes forever to show up or there is medical malpractice and the person dies... -Paul
[2018-11-09 10:45:45] - paul: of course, there might be some other better way. but, save us not finding that way, i'm not uncomfortable with legal ruling based on lucky outcomes, because our legal system is chocked full of them. two men beat their wives to unconsciousness: one wife wakes up, and the other dies. are you ok with the widower getting more time? ~a
[2018-11-09 10:30:10] - ok. ~a
[2018-11-09 10:11:34] - a: Right, listen, I 90% agree with you. I have no counter proposal. Just wondering if there might be some better way. -Paul
[2018-11-09 10:06:30] - we punish outcomes every day. you can drive while texting every day and not be punished. you hit a kid in a crosswalk, and now you're punished. should you be treated differently than the person who texted *just as much* and didn't hit a kid in a crosswalk? ~a
[2018-11-09 10:03:32] - paul: ok, well then i say, i'm fine with the law the way it is. and until you propose something better, we'll keep it that way. ~a
[2018-11-09 10:00:44] - a: I don't have a thesis, this is purely me wondering aloud. -Paul
[2018-11-09 10:00:27] - a: Should those be treated differently? The same? I honestly don't know. Obviously the outcome of the 90 year old getting hit was worse, but that would seem to be pure luck, and I'm a little uncomfortable with legal rulings based on lucky outcomes. I don't know what the alternative is, to your point, though. -Paul
[2018-11-09 09:59:32] - paul: part of me wonders your thesis . . . that we should punish people based on what they did and not the outcome . . . meaning? what should change? ~a
[2018-11-09 09:59:15] - a: Sure (and keep in mind I mostly agree with you, but also see the other side), but let's take a different hypothetical. Two different drivers hit two different cyclists in the exact same way. One is a healthy 20 year old who gets maybe a broken arm and some cuts and bruises. The other is a 90 year old who hits their head and dies. -Paul
[2018-11-09 09:53:53] - there's some luck there, of course. but the real problem was hitting hollowell in the first place (and driving off in the second place). the luck, that hollowell hit people's windshield, and whether he died from it or not, is minor by comparison. ~a
[2018-11-09 09:51:57] - we see if hollowell died, because there's no alternative. ~a
[2018-11-09 09:51:05] - paul: we're trying to punish based on what's done. but to determine what was done, we have to *take into account* the outcome. because otherwise we would have no way of determining what was done! we can't ask the defendant (and get an unbiased answer) how hard did you hit him? if you had hit 100 people that hard, would 50 of them have died? so, instead we see if hollowell died and that's the law. done. ~a
[2018-11-09 09:47:41] - a: Are we trying to punish based on what was done? Or the outcome? Put that way, I lean towards the former. -Paul
[2018-11-09 09:47:19] - a: I could easily argue both sides. The thing that bothers me is that oftentimes whether or not somebody dies could be pretty independent of the seriousness of what you did. Like, maybe the cyclist has cat-like reflexes and rolled the right way to minimize injuries. Why do you get a lesser sentence because you luckily hit somebody like that? -Paul
[2018-11-09 09:45:43] - a: On the surface, I don't like the idea of felony murder. Seems to muddy the water too much to somebody who doesn't know much about law. I like murder to require intent. -Paul
[2018-11-09 09:34:15] - paul: to address your original statement though, yes it should matter (imo). if you're worried about making an arbitrary law, with rules that are subjective, then "if somebody died or not" is perfect. ~a
[2018-11-09 09:32:24] - paul: "felony murder" is the perfect example of what you're talking about. you were doing some serious shit and somebody died. now you're in trouble. i'm kinda on the fence about felony murder: especially since we're pretty fucking loose what we call a "felony". ~a
[2018-11-09 09:20:20] - One thing I always have trouble balancing is the seriousness of what happened versus how intentional it was. Like, let's say the exact same thing happened to somebody else, but the person survived with just minor injuries. Should the punishment be different? In some ways, I think so, but in other ways, it doesn't seem like it should. -Paul
[2018-11-08 22:32:08] - daniel: nope. paul: yep. ~a
[2018-11-08 16:20:54] - a: Wait, I can't remember, did the other guy die? I guess involuntary manslaughter implies he did. -Paul
[2018-11-08 16:17:24] - a: I mean that still is between 1 and 10 years in jail. That seems like a "big deal" so is there more that you need for "justice"? -Daniel
[2018-11-08 16:14:00] - paul: that's right. i'm not a dc lawyer, but he pleaded guilty to "involuntary manslaughter" and did not plea to the "hit and run" (which in some areas is a second charge). ~a
[2018-11-08 16:02:50] - a: Wait, he plead guilty to something less than hit and run even after they basically had text evidence of him confessing? -Paul
[2018-11-08 15:47:12] - "Peoples at first denied involvement in the hit-and-run, saying a friend had borrowed his car that day while he had taken an Uber to work. Detectives later figured out from text messages Peoples sent to an unidentified person that he had hit someone with his car, according to court documents. He wrote, 'I jus hit somebody bad . . . It’s glass all over me.'" ~a
[2018-11-08 15:45:03] - xpovos: the hit and run we discussed in september (and october when he was caught) made it to trial. he pleaded guilty to a lesser crime. ~a
[2018-11-08 13:10:55] - mig: "you mean legally?" i didn't. i meant any at all. "I'm not sure any laws were broken in doing so" i guess i'll differ here. i can't tell you what law exactly was broken, but i'm sure at least one or two have been broken here. she used falsified documents in her official capacity. fraud is illegal. so is breaking ethics laws. libel is illegal. ~a
[2018-11-08 13:08:53] - For the record: the article https://www.newyorker.com/news/letter-from-trumps-washington/trump-the-2018-midterm-elections-jeff-sessions-firing -- Xpovos
[2018-11-08 13:03:10] - otherwise, i could see some scenarios where the WH is forced to walk this back, especially given the way they've tried to justify it. - mig
[2018-11-08 12:59:15] - if you mean legally? Almost certainly none. As repugnant as what the WH did pimping that video, I'm not sure any laws were broken in doing so. - mig
[2018-11-08 12:57:35] - a: I went to get the link and was re-reading it. I inadvertently mischaracterized it. The article says nothing about the removal of Acosta's privileges, pro or con, I think it was written before that. It does mention the dust-up, but only in passing. Sorry for also misleading you. -- Xpovos
[2018-11-08 12:57:20] - a: probably nothing significant. - mig
[2018-11-08 12:56:03] - "unbelievable stupid" so we at least agree on this much. but my guess is that there will be no repercussions for the stupidity. agreed again? ~a
[2018-11-08 12:49:43] - a: I was just pointing out the hyperbole of that statement. Acosta isn't really being accused of a crime either. Either way, Sander parroting that clearly doctored video was unbelievable stupid, but that's this administration in a nutshell. - mig
[2018-11-08 12:41:41] - a: they had already been engaged in back and forth for a while, so at some point, yes, the president does have to just decide to move on at some point, since there are other reporters who no doubt have questions. - mig
[2018-11-08 12:31:00] - "the President decides to move on" so it's settled then. the president has decided so it is so? why call on people for questions then. ~a
[2018-11-08 12:28:14] - I mean look, I want the press to challenge the president, but I'd rather they not sink to his level, which Acosta frequently does. - mig
[2018-11-08 12:25:08] - My issue with Acosta isn't the incident on video with the woman, but the context surrounding it where he's engaging the president in a debate about immigration. They both basically make their points and the President decides to move on, with Acosta basically demanding the debate continue, which then leads to the confrontation on the video. - mig
[2018-11-08 12:21:43] - mig: i don't think i suggested otherwise. the whitehouse is using falsified evidence. are there enumerated rights that prevent that? ~a
[2018-11-08 12:19:25] - xpovos: do you have a link to the new yorker article? i'm surprised they didn't talk about the doctored video. ~a
[2018-11-08 12:18:41] - a: there's no enumerated right to be a member of the WH press corp (you or I can't just invite ourselves to those briefings and be given time to ask questions). Acosta is still free to report and cover the WH in other capacities. - mig
[2018-11-08 11:42:30] - a: I was reading an article about it in the New Yorker (hardly a conservative venue) and it seemed to agree that revoking the credential was warranted, though it didn't agree with it, the way the conservative news pieces I read about it did. -- Xpovos
[2018-11-08 10:53:29] - paul: "the government is literally falsifying evidence to wrongfully accuse a private citizen of a crime in order to revoke his rights". at least the right to a free-press will only affect acosta. ~a
[2018-11-08 10:50:35] - you can have your credentials revoked. i imagine it isn't something done often, and it probably (politically) needs to be done with cause. ~a
[2018-11-08 10:48:47] - a: Naw, but I also wasn't even aware being removed from the press corps was something that was done. Maybe it really wasn't before Trump.
-Paul[2018-11-08 10:46:04] - paul: based on that video you saw, do you believe should jim acosta be removed from the press corps solely because of his actions touching (or being touched by) that woman? ~a
[2018-11-08 10:42:39] - a: I literally only saw like 5 seconds of video of the exchange, and it was from I think the local news (so not horrible biased in any way I don't think) and I had two thoughts based only on those 5 seconds: (1) The woman seemed pretty aggressive trying to take away his microphone (2) I was a little surprised he resisted at all. -Paul
[2018-11-08 10:40:16] - jim acosta got booted out of the press corps yesterday. i watched a video from a different angle and it's not surprising that someone was able to doctor a video to make him look bad. you don't go up to another human and try to grab something out of their hand. and doctoring a video like that is probably pretty easy. ~a
[2018-11-08 10:35:14] - xpovos: jon stewart is the only stewart i know. ah . . . corey stewart was on the ballot. i guess i should have known that. ~a
[2018-11-08 10:24:03] - I know Corey pretty well. So I guess that made it less exciting? But, yeah, it made a mess of things, too. -- Xpovos
[2018-11-08 10:22:59] - Daniel: Not often, I guess? I honestly don't often refer to somebody's gender either. -Paul
prev <-> next