I have a bunch of not necessarily self-coherent opinions on this: If the local government is going to parcel out a public good, and is doing so presumably because they've been captured by monied interests, they could at least take the effort to inform everyone of it. Not everyone checks, or even has the ability to check, the app. And establishing but not enforcing such a rule is shitty and confusing.
The bourgeois part of me that craves order does think that the guys who were already there should have left--it's not exclusively their field (it does really suck when some group aggressively hogs a public place), and the Dropbox guys did reserve it--but they were at least okay with sharing it, which is more decent than the Dropbox guys were. If there's a continued pattern of these conflicts, they could at least try to reach out to the city or parks service that there's no local support for reservations. But this is probably a more effective way to make that clear.
The Dropbox guys behaved dickishly. They could have at least waited until the guy with the permit showed up before they tried to get the other guys to leave. I do kind of love how it looks like Valleywag changed the name of the article, since the URL has "douchebags" while the article says "dudes."
The obvious correct solution is to have them each play to two goals and whoever wins gets the field for the next hour, whoever loses goes home in shame.
The bourgeois part of me that craves order does think that the guys who were already there should have left--it's not exclusively their field (it does really suck when some group aggressively hogs a public place)
It appears, from poking around, that the community norm at that field is that when there are players waiting for the field, the existing players play until a goal is scored, and then the losing side vacates and a new team challenges the scorers. This seems fair and would prevent one group hogging the field forever. Anyway I'm strongly in favor of honoring community norms over having dudes with money and papers demand that people vacate the field because they downloaded an app and paid a fee or whatever.
Also, "disastrously awkward"? Omg what a douche, seriously.
2: Oh--that's a pretty good norm, and should be respected. Sucks if you want to have the field to yourself for a bit, though, which I think is a legitimate desire, but if this has been working for them so long they should continue with it. (I'm used to sports fields that are time-shared between e.g. college teams for practice and the public.)
I like 2.2, because then you grow up and understand the quarters on the pool table.
I like 2.2, because then you grow up and understand the quarters on the pool table. Street Fighter II cabinet
4: they have quarters on pool tables? I thought that was a Street Fighter arcade machine thing.
Oh my god that's the most pwned I've ever been. Sorry, everybody. Sorry, dalriata.
I'm just more impressed that we went with the same reference. Solidarity!
8: I assume this is a joke because there are actually somewhere between five and six thousand versions of Street Fighter 2.
You can assume it's a joke because I'm well-known to be hilarious.
there aren't any finishing moves in street fighter 2 don't you guys know anything gosh.
up up down down left right a b a b select anachronism
15: OMG! BLACK BLOOD OF THE EARTH!
Double Dragon is kind of like a Big Trouble in Little China game without the things that make that movie awesome.
There's way more dynamite in double dragon than in BTiLC.
BUCKAROO BANZAI defeats JOHNNY CASTLE with RED FALCON
Oh, yeah. There are different things that made it awesome.
Anyway, really sorry for the early threadjack and too many guys in tech are entitled douchebags. (For those of us who are, we need to police our own better.)
(For those of us who are, we need to police our own better.)
Well, that explains the Double Dragon fact-checking, at least.
The update to the story says the app in question wasn't used to book the field. Did the app get a bad rap?
Oh god I was just talking about that with my girlfriend. The whole situation's a clusterfuck; as the updated Gawker says, the pickup soccer group that was charging to play there gave up because folks in the neighborhood were upset about it.
Meanwhile, the city's been trying to build new (artificial-turf) fields out by the Beach Chalet for years now, to address the pent-up demand for fields, only to have to fight a bunch of NIMBYs. The NIMBYs already lost their court case to stop the project, but now they've gotten a proposition on the November ballot.
On our own reserved soccer fields, the leagues that have had the same reservations for years have had to kick off Quidditch players.
While not saying that they look like perverts running around holding a stick between their legs.
22: Yeah. I belonged to the app's mailing list and they gave up on it at least a year ago. Meanwhile the city's Parks and Rec department has a reservation site which lists the field as one of the choices.
21: These are the folkways of my people. It's important that these stories be passed on to future generations. Or else how will they ever learn about the hidden 1up before the first pit in world 1-1 of the original Super Mario Bros.?
(But really, video gamers are dicks.)
I've kicked kids off fields we reserved easily half a dozen times. We had paid for the permit months in advance and I called the city ranger-person to enforce it. Go home and cry, little soccer kids. Your coach should have put in the hours I spent to reserve fields, which they would have gotten, since kids' sports get first priority in the reservation process.
I can't tell from the story how the app interacts with the city reservation process. It is fine if the city doesn't include a known pick-up game time in the reserve-able slots. But given that it was included, I have a fair amount of sympathy for the people who got a permit and reserved the fields.
If you have reserved an hour or two of field space and you don't know how long it will take the city ranger to show up to back up your permit (I always brought the papers to the field), it does suck for the people already there to be slow to vacate. I have ten minutes of field set-up to do! You are costing us daylight and playing minutes.
I wonder if this is a good place to enact my ambivalence about sharing this driverless-car-hype article that has got to be kidding. Obvious tell:
"I could sleep in my driverless car, or have an exercise bike in the back of the car to work out on the way to work," he told the Times. "My time spent in my car will essentially be very different."
31: In fairness, the people in that article don't seem to be thinking of the driver-less car as something that would ease congestion so much as something that would make sitting in traffic more tolerable because you wouldn't have to be driving.
Currently, Google's self-driving cars don't seem to do much driving outside of an extremely mapped out area in and around Mountain View, require someone to be ready to take the wheel at any moment, struggle with heavy rain, which can be hard to distinguish sensorily from a wall, and can't make unprotected left turns on their own. So it might be a while.
31: One of the comments is great -- proposes self-driving apodments and people just live in their cars. Of course, this was a Dr Who horror plot.
I'm looking forward to self-driving cars so that car2go can rebalance itself, or I can summon one from afar. Really the sweet spot for urban travel in Seattle -- with two people, often as cheap as taking the bus. (They park for free on city parking. This turns out to be a spectacular savings in time and money. I am curious if the city wrote a general description of services that get this perk, or have a Shoup-style number for how to raise their parking rates if they start to outnumber private cars. Maybe the car2go company pays.)
I'm in the Bay Area at the moment, and I was kind of surprised that things don't feel like they're all that much more gentrified than 10 years ago. I'd somehow gotten the idea that it'd be totally different.
(But really, video gamers are dicks.)
The Gamer[-I-can't-bring-myself-to-use-the-suffix-"gate"] stuff is showing some of them to be far worse than I would have guessed, given that I've had the fortune to not have been subjected to that kind of abuse.
To be fair, clearly no one involved with that article put measurable amounts of thought into whatever they said or wrote, but it struck a chord of terror anyway. I guess maybe the dude could ride a recumbent exercise bike? Oh! They could have workout cars where you pedal to generate energy to charge your laptop! If you drive a pedal-car, your health insurance premium goes down, but the car will track your pedaling and transmit data to your employer... am I doing the tech fabulist thing right?
It is funny that they panic in heavy rain. I also recall a story about some target of demonstrations who lives (?) in Berkeley and commutes to/from Google via self-driving car. The question mark is because I have completely lost sight of what constitutes life, evidently. The cars should all be printed with the slogan "My commute is better... for ME!"
36: Absolutely, but I'm kind of stuck at the tautological level of "well, sociopaths DO like games."
35: I'm not in SF and the only part of SF I lived in was the Richmond, only for a summer in 2010. That area didn't seem hugely different when I was up there last week, but I bet its relative isolation up in the corner of SF gives it a buffer against rapid gentrification.
Down here in the south bay there are lots of places where I feel like you know there's a bunch of wealth around but things still have that post-WWII to 1990s ex- and suburban sprawly look. And you might not know there's a tech boom at all if you spent most of your time in downtown San Jose with the exception of a couple isolated blocks. But there are also office and residential developments around here that are clearly the product of the boom, including luxury-style apartment buildings going up or recently finished in downtown San Jose, and stuff looming in a non-threatening non-high-rise manner over parts of El Camino that formerly were just one-story whatever buildings. And also blocks of two-to-three bedroom apartments and condos distributed at various locations within the sea of office parks from northern San Jose up to at least as far north as Sunnyvale.
recumbent exercise bike?
Or an elliptical bike. I've seen a few of those around.
33.2: yeah, it's going to be a long, long time. None of those problems are trivial (arguably they have the resources to map larger areas that intensely, I guess. But that's what it takes.)
37- It would be cool if the pedals were connected directly to the drive mechanism, and you could make it more realistic if you let the exercise bike's handlebars control the direction of the vehicle. Controlling the drive mechanism might be hard with that much weight so you want to make the shell as minimalist as possible, maybe reducing it to just the frame of the bike in the extreme.
yeah, it's going to be a long, long time.
Till touchdown brings you once around to find you're not the man they think they knew before?
That sounds kind of like a song I know except for several words that are totally out of place.
Roads have to be retrofitted for auto-automobiles? Doesn't that defeat the purpose?
45.1: the problem isn't the roadway. The problem is that the car has no ability to identify or gauge the seriousness of obstacles. So a Google driverless car on a highway would come to a squealing halt if a plastic bag blew in front of it.
Highway driving is actually less of an issue, though. It's roads with other uses besides cars (and with traffic signals, construction, driveways, etc.) that are going to be problematic.
The way they solve this in Mountain View is by mapping the entire road network to millimeter scale or something with LIDAR, and assuming that any deviations from the existing map are meaningful. This is a pretty good 80% solution for a small scale test but... see above about left turns and heavy rain. Designing a transportation system for an operator -- human or non-human -- that can't "see" in any meaningful sense is a big and not-very-close-to-solved problem.
Now I'm curious what the magical 75-100 miles are. Some strip between LA and SF? The beltway? The LIE?
Also, a driverless car where the "driver" doesn't have to pay pretty close attention is even farther off; the likely nearer-term scenario is a car where you sit and stare at the road but mostly don't do anything unless it's an emergency, which is fucking terrifying for reasons closely related to the crash of Air France 447.
47.1: That's similar to a version the trolley problem. You just need to program the content of recent philosophy into the car.
Of course, in ten years we'll be able to upload our own brains into the cars and teach them how to drive.
"Give me a stopwatch and a map, and I'll fly the Alps in a plane with no windows."
"If the map is accurate enough."
52: you're kinda being a lot flippant. It is an engineering problem, but it's an engineering problem that has been actively researched for decades without a terribly huge amount of progress. I'm not saying it'll never be solved, but a computer that can pass the Turing test is also an engineering problem.
52: well, listen, adaptive cruise control plus automatic lane changing are essentially already here. I'm sure steering assist for highways won't be far behind. So if that's what you want, you're in luck. Until you're stuck behind somebody who's blind drunk but doesn't get pulled over because they aren't weaving, but anyhow.
But what Google is after is something very different: an end-to-end driverless car that works on surface roads.
It's certainly possible. But the current state of Google's car doesn't give us a ton of information on that score.
Commercial fusion reactors! Machines to suck excess greenhouse gases out of the atmosphere! Pot-crapping dogs! #mereengineeringproblems
I mean, for one thing I imagine the demand is great enough that driverless cars would get released and sell very well even if they didn't work particularly well and were extremely dangerous (especially for pedestrians and cyclists). So the difficulty of solving the technical challenges might not be dispositive.
54: I was hoping for something less shitty.
I know people who have taught their dog to crap in a pot. SOLVED!
A lot of the business case depends how the liability shakes out, though. Who is at fault when a driverless car plows into a baby carriage because it thinks it's a plastic bag?
What is the race and income level of the baby's parents?
I was going to say that the answer was probably "the baby", but anyhow.
It's worth noting that the major auto manufacturers -- who obviously have a lot to gain, and who have been working on this kind of thing for a long time -- are not generally on board with google's strategy and are instead focusing on the limited highway approach I mentioned above. It's also quite possible that the computer vision aspects of the problem are not easily amenable to the "bigger data and deeper networks" approach Google is go good at.
Did you know that you can actually look at google's LIDAR data? If you put Google street view in anaglyph 3D mode the LIDAR ranging is how they build the depth map. It's rather eye-opening in its glitchiness.
They could just bring the NSA* in on it, and classify all the documents relating to how these decisions are made in order to make it harder to sue the companies involved.
The enthusiasm for driverless cars really makes me suspicious of the people involved, because honestly every time they describe why it would be a good thing to have them it sounds like "it would be like public transportation only really expensive and fashionable because poor people couldn't have one!" Well, that or the claim that computers would be safer because they wouldn't be subject to the dangers of human error, which is about as convincing as the version in the linked video.
*"Threat to national security detected. Swerving to increase chances of collision."
This is a pretty good article about the Google cars.
72: LIDAR may be, but living in areas with poor pedestrian infrastructure isn't.
74: the best (that is, most legit) argument for self-driving cars is that even relatively low numbers of them would make a huge positive difference in highway congestion, just because they would actually keep at an appropriate following distance and avoid pointless lane changes.
67: in order to do that kind of back-of-the-envelope cost estimate you're probably assuming that we're just outfitting each car with cameras and computing, right? I guess that wasn't clear to me before; I was thinking that the engineering solution we were envisioning might involve somehow instrumenting roadways everywhere into some kind of network that keeps track of all sorts of detail outside the scope of what an individual car's cameras can see. In which case it seems like costs could potentially be exorbitant.
77 cont'd: of course, the people driving those cars might find them infuriating for the same reason.
79 cont'd: which will presumably lead to a robust subculture of hacking self-driving cars to drive like assholes.
Maybe for Massachusetts Google will market a special version of the car that honks needlessly and incessantly.
What is the criminal penalty for splashing paint onto the sensors of a self-driving car as it passes by? (Asking for a friend.)
I have to say I'm somewhat pleased that there's a real world computing problem that is not easily solved in the near future. My industry suffers from a number of tech bros who think they can cross over and snap their fingers and tell us if we just engineered the problem correctly we'd solve this whole disease thing. See Grove, Andy.
I think self-driving cars will come by way of a whole bunch of incremental advances, like this:
The Tesla D includes an advanced "lane changer" feature that moves your car into the next lane when the coast is clear, following your initial trigger -- turning on the turn signal.
Another semi-autonomous feature of the Tesla D is that is reads road signs and keeps you within the speed limit. (Though, given that traffic often flows ~10 miles above the speed limit, I wonder if that is or will eventually be as advanced as Google's system.)
The Tesla D will also keep itself within the lane by following the lines on the road, and can very gently stop itself by keeping an eye on where the car in front of it is.
Elon Musk also noted the Tesla team was working on "self-parking technology that will let you get out of the car and before it parks itself, as well as a feature with the calendar app that will have the Model S drive itself to a pre-determined destination to pick you up when you need it."
83: there are lots and lots of them, mostly within the domain "things people do easily or automatically".
86: such as being assholes. It's hard to get a computer to do that right.
WHAT'S YOUR FUCKING PROBLEM?
tech bros who think they can cross over and snap their fingers and tell us if we just engineered the problem correctly we'd solve this whole disease thing.
It's been fun over the years watching one computer scientist after another hear about the protein folding problem and proclaim "Why that's trivial!", and then checking back with those same people after a few years.
87: I think that one might be pretty easy, actually. If you define it as "design a computer program that will reliably get people to scream 'FUCK YOU' at it" Microsoft nailed that down decades ago.
It's been fun over the years watching one computer scientist after another hear about the protein folding problem and proclaim "Why that's trivial!"
There are hard problems whose hardness is surprising, but this doesn't seem like one of them. I can't imagine what sort of miscalibrated intuition someone would have to have to not think that protein folding is a really, really hard problem.
way they solve this in Mountain View is by mapping the entire road network to millimeter scale or something with LIDAR, and assuming that any deviations from the existing map are meaningful
And in the current design, the map has to include the locations of all traffic lights. There's a camera that reads green/yellow/red, but it doesn't detect light placement, so it needs to know where to look. You could map out every traffic light in the US, or retrofit every light to emit signals for self-driving cars, but it will take a while.
I don't know how hard the weather problem is. The LIDAR reads heavy rain, and presumably snow, though I don't know if they've gotten to snow conditions, like it's a wall. In the early days it had trouble distinguishing groups of pedestrians as groups of individuals rather than some block of something, and it had trouble identifying motorcycles/bikes between cars, but all of that has improved. They've got a safety precaution that makes the car aware of and more cautious around anything that gets identified as matching the shape of a child. The default when the car's systems can't make a safe decision is to slow down and stop with the hazard lights on, so it's not going to run anyone over unless they're blocking the LIDAR somehow.
91: It seems to be specific to computer science, for some reason. I haven't seen similar attitudes from, for example, theoretical physicists who decide to take up protein folding.
The attitude seems to be bound up with some vague notion that it's "just" a pattern recognition problem.
94.2: and computers are typically so terrific at those!
I've asked a few people this question over the years and never got a clear answer: is it even clear that protein folding is a well-specified problem? If the problem, as it usually seems to be framed, is "given the amino acid sequence of a protein, specify the three-dimensional conformation it takes on in its biologically relevant form", it seems like you might need to know more than just the sequence, e.g. whether there are other enzymes around that might destroy unwanted conformations, or whether other aspects of the chemical environment change the thermodynamically preferred state.
94.1: I know at least one string theorist who uses protein folding as his go-to example for explaining why it's so damn hard to get any predictions out of the theory. "We have this huge set of possible ground states, and no clear rule for figuring out which one we happen to be in--it's like protein folding!" I guess he never tried that one out on a computer science audience.
Oh, now I see that there's a whole Wikipedia entry about the question in 96 and biologists reassuringly refer to it as a "dogma".
Well, no one is ever going to get to close highways to driven vehicles in advance of driverless vehicles being a big big thing.
This is a silly vanity project. I predict that actual driverless cars will be no more common than Deloreans at any point in my lifetime.
It's pretty clear that biologists don't know what "dogma" means. See"central dogma of molecular biology"
99: But imagine how great it would be if we did that. A dedicated system of corridors connecting cities to each other (and various locations in cities as well) covering large distances where people can just get on it and sit back in a seat while they are automatically transported more efficiently and at higher speeds to the closest exit point to where they want to go than you could ever get from people driving there themselves. TrainsDriverless cars are the way of the future.
96/7: To answer why it appears simple seems pretty straightforward: You have a known sequence of animo acids, with known bond lengths and distances. The thermodynamic properties appear to be mostly additive. We can predict things like DNA annealing pretty well for a variety of tertiary structures. Proteins should be predictable by the same approach - hydrogen bonding, freedom or limitation of bond rotation, etc.
I'm sure you understand why this is naive (since you just read the wiki article), and most people who have the difficulties explained and still don't get it are just arrogant. That said, it really doesn't ruffle me when people want to tackle longstanding problems in any field. The more eyes, the better. I don't expect them to succeed, but I think that varying the standard approach isn't a bad thing. It's when the nonstandard approach becomes trendy enough that everybody jumps ship and forgets how to do the painstaking work that went into advancing the field to date that makes me cranky.
Oh right that was the thread where GB told us a bunch of awesome stuff. Which somehow didn't sink in enough for me to remember much of it.
Maybe because my memory of things I read but don't ever use again is bad, and maybe because that was right before I had a week of very little sleep including some all-nighters for the first time since college.
103: I guess the problem is that my intuition runs exactly the opposite way, so that I'm really surprised to hear that the problem is much easier for DNA than for proteins (if that's what you're saying?). It seems like no matter how well you understand the building blocks the combinatorial explosion of possible structures to consider is going to be overwhelming. So maybe I'm naive in the opposite direction.
It's not just the end state but the kinetics of how natural proteins get there. They obviously don't sample all the states but have some crazy efficient path it seems like we should be able to determine but even that is a huge mystery.
89: And ex-NASA engineers who figured on taking over medicine in a few years in the Sixties.
96: yeah, it's "sequence plus environment". Even then, though, sometimes a single sequence can encode 2 distinct folds that are near enough in (free) energy that the protein in question will morph between them (I work on one of those).
I think the classic way of posing the question - can we unambiguously "read" the structure from the amino acid sequence - is becoming less interesting because the answer emerging seems to be "sometimes but even in the best scenario there will be enough exceptions that you wouldn't want to bet your life on the resulting prediction".
Even the most successful prediction folks are no longer really wedded to the idea that we will be able to read sequences like words.
|| Speaking of vanity, I'm teaching a law school class for a day later this month, and have to decide now what I want the students to read ahead of time. This is actually kind of hard. |>
110: I can give you some references on protein folding.
Do law school students reliably do the reading?
This is everyone's go-to example of how topology can be applied, but I'm not really clear on how extensively topology can solve stuff. I get that you can say "no, you can't go from state A to state B MATHEMATICALLY!" but math can't actually handle that much input.
106: I just typed and lost a big comment, so I'll just say that when predicting DNA structure, each base is rigid, large, and pairing options are limited. The bases are hydrophobic and the backbone is hydrophilic, and ion binding is limited to a few ions. Also, in general, a sequence's crystal structure generally matches reasonably well with its solution state structure. (DNA oragami is an example of using a sequence to build a particular structure.) Proteins have 20+ amino acids, with varying properties, many more freely rotating binds, all kinds of ion binding (not to mention substrates or other proteins). Crystal structures don't necessarily match solution state, which is environment dependent. It's possible to design an amino acid sequence to have a particular secondary structure, but not generally a teriary structure.
Also, take with a grain of salt since it's not really my field.
114 makes sense. After re-reading GB's old comments which nicely explained how to count the combinatoric complexity of proteins in terms of the number of rotational configurations of the single bonds in the backbone, I can kinda see how what I remember about DNA's structure from high school biology would tend to take most of that freedom away by requiring that too many pieces fit together in too many places.
Back to the OP, I just watched the video, and I'm not seeing the guys with the permit as particularly awful here. There's an argument that maybe the city shouldn't have been allowing reservations at all, but once they are, it's an orderly process and I don't see the problem with abiding by it.
(I'm also kind of suspicious of 'community norms' in terms of who gets to play. The coverage is describing the community norms here as utopian and inclusive to any team who calls next, but if it really works like that it's unusual in what I know (anecdotally, unreliably, second-hand) of heavily used public sports infrastructure, which in the absence of permits tends to get captured for exclusive use by insiders. (Like, say, the guy with dreads who was talking about having lived in the neighborhood for twenty years rather than one, as if that should be relevant to who gets to play. I mean, the insiders here are 'the community' in a way that guys with tech jobs aren't eligible to be, but still.)
What I don't get is that the Dropbox guys seem to say a couple of times that they would be happy just to play against the neighborhood kids, 7 on 7, according to the neighborhood system, but they never do it. Instead there's a jump to when the guy with the paper shows up. Why not just follow the local custom? It seems a perfectly good compromise.
That said, it really doesn't ruffle me when people want to tackle longstanding problems in any field. The more eyes, the better.... It's when the nonstandard approach becomes trendy enough that everybody jumps ship and forgets how to do the painstaking work that went into advancing the field to date that makes me cranky.
I think interdisciplinary research ends up like this quite often, because people in field A get interested in field B, but continue publishing their work in field A, where they're presenting it to people who are even less familiar than they are with field B. So they can say a lot of crazy stuff, and build a whole community of people in field A who think they're solving an important problem in field B, but actually it's just a problem they made up themselves. It's very annoying, but it pays the bills gets people tenure.
118: Um, being suspicious here? But the linked article seemed pretty focused on calling the Dropbox guys assholes, when they seemed really pretty calm for people who thought they had a permit (again, you could argue that the city shouldn't have started issuing permits where there was a thriving pickup scene, but that doesn't make the guys who got the permits assholes.) I wonder if there was an interpersonal breakdown that wasn't unambiguously driven by the Dropbox guys being assholes that kept the compromise solution from happening, and it got cut out of the video as not supporting the narrative.
I saw this guy give a talk Friday where he said a lot of crazy, made-up stuff about my field, but he said a lot of crazy, made-up stuff about like fifty fields so I didn't feel singled out or anything.
I saw something advertising that talk and was confused because there's a guy with almost the same name in my department. Different middle initial, though.
What 102 said plus engineering/tech continues its impressive run of being unable to succeed in negotiate everyday human encounters, see story in OP.
TrainsDriverless cars are the way of the future.
Have I mentioned my idea of rebranding "bus rapid transit" as "very light rail"?
ydnew's 103 is a good description of why it seems like it should be an easy problem to someone with a computer scientist's intuition. heebie's 113 is a good point in general, but computer science has had good success in figuring out how to extend math to handle lots of input. (I kind of want to make a CS:math :: chemical engineering:chemistry analogy, but I don't have enough examples to pull it off, and also banned) I do think this is a genuinely hard problem, and a good thing to point to when any CS person thinks that we'll have arbitrary precision physics simulators (of anything more complex than a handful of atoms) soon.
102: I wonder if intercity trains would be more popular in American if they did car-carrying, so you could drive your own car at your destination. Yes, that would defeat some of the advantages of trains.
119.2: I share your frustrations (or possibly enact them, it's not for me to say).
||
People fascinated by knitting should go look in the Flickr pool.
|>
130: Very pretty. How long did that take?
Shut up, she explained. I'm not a fast knitter.
Um, since March?
I don't knit so I don't know what counts as fast, but it looks intricate and as a personal project that's probably pretty low on the to-do priority list, that sounds very reasonable.
It's a big complicated intricate thing, yes, but for a good knitter that means a few weeks, not six months. Nonetheless, I'm pleased with myself.
I'm not seeing the guys with the permit as particularly awful here. There's an argument that maybe the city shouldn't have been allowing reservations at all, but once they are, it's an orderly process and I don't see the problem with abiding by it.
Isn't the problem that the app preys on the digital divide, and that the Dropbox guys most likely knew they were making a reservation at the middle of prime pick-up time, and would disrupt a bunch of people who had no idea the app existed? I think the assymmetry of knowledge is what makes them douchey - they knew perfectly well they were going to come in and kick unsuspecting people off, and so they're not going to get riled up.
All that opinion without watching the video!
I'm pleased with myself and I've never even knitted at all.
If I could remember by yahoo password to look at the knitting, I might be even more pleased.
136: but they didn't make the reservation using the app
136: Huh. I would not attribute that asymmetry of knowledge to them at all. First, I thought the app turned out to not be actually relevant, it was a detail thrown in to make the douches sound douchier, if you look at the update to the linked post. They just got a permit through whatever the normal permit process was.
But generally, if the guys with the permit aren't part of the pre-existing pickup scene, why would they know either that it's pickup prime time, or that pickup play isn't regularly interrupted by permitted play? And of course, how do we know that pickup play isn't regularly interrupted by permitted play? Is this really the first permit that was ever issued for this field?
139: The same arguments apply to using the website.
Plausibly. I actually don't really have a dog in this fight (and am not going to watch the video). But holy shit the way those articles are written is tendentious!
140.2: apparently it recently reopened with an actual soccer field, where before it had not had an actual soccer field but people had played soccer anyhow?
But generally, if the guys with the permit aren't part of the pre-existing pickup scene,
Do we know this? He's lived in the neighborhood for a year. Probably they at least know it's not an idle field.
I'm imagining that they went to the website, filled out some clunky sign-up procedure, and saw acres and acres of unreserved time slots, and knew perfectly well that the fields were generally used far more than the time slots represented.
The blame falls with the city - if you want a reservation system, you need to go to great lengths to inform the current users that you're shaking things up. It would even be possible to solicit their input!
So the douchey guys aren't to blame, but I think they qualify as smirking douches for knowingly ambushing the established scene.
Still guessing! Haven't even clicked through.
I actually don't really have a dog in this fight
Where "this fight" is the up-its-own-assholes battles that have characterized San Francisco local politics since ever but which have now, for reasons that aren't entirely clear to me, entered the national consciousness.
Did you guys hear the one about the coffee cart that launched a petition campaign to get the other coffee cart shut down because it was a chain except the other coffee cart was actually owned by noted blogg...
...
... whoops. Just fell asleep for a second.
It is also worth nothing, since the one guy says he's lived there for "twenty years" that twenty years ago is barely long enough ago to have moved to the Mission without yourself being a gentrifying asshole. Maybe not even.
146: Well, in his case, he looks twenty-ish. So I'm guessing he means he was born there.
(No, I guess that's different. But that means that for his entire life there have been lots of people working in tech living in SF.)
144: None of that's impossible, I guess. But attributing smirking assholishness to people because they signed up for a permit seems to import a lot of information not clear on the face of the situation.
(Really, I should be on the other side of this one. The sports teams at Sally's school seem to do all of their practicing by sneaking onto fields that no one's using at the moment and then lying about having permits if someone calls them on it. It's not clear to me why they can't successfully get space somewhere properly, but (1) it's not my problem, I'm not the coach and (2) field space is really tight, so maybe they honestly can't. It's still a pretty new school, and they didn't get teams going for a few years, so it's possible that all the slots are held by users who've been there for longer.)
But attributing smirking assholishness to people because they signed up for a permit seems to import a lot of information not clear on the face of the situation.
I dunno. The sports scenes and rec scenes that I've been a part of are absolutely rife with unadopted technology, where they make some big effort to implement some terrible new system, contracted out, it's confusing and requires all sorts of procedural hoops, and most of the people never really recognize that there's a website whatsoever, and it falls by the wayside within a few months. To me, this smells like using the existence of the new website as cover to be disruptive and obnoxious.
Did everybody miss the latest Google car prototype? The idea is that it's just for in-town driving, so detailed mapping is more feasible. And it's light and slow so it's not very likely to mow down pedestrians. It won't replace inter-city rail, but it could replace Uber. (To keep it on topic: Uber, what a bunch of douchebag tech bros!)
And it just plays into the whole racial/ethnic component of the digital divide in a particularly obnoxious way. In Heebie Town, if you had a primarily Hispanic game, probably half the people playing don't have easy access to a computer whatsoever, let alone an email address, ease with English, and general familiarity with how the internet works to register for the newfangled system.
(and everybody knows this. This is no secret. The Dropbox guys can be expected to recognize this.)
Amen to 145, although I keep falling back on the explanatory power of "media organs want the (reputedly numerous and literate) rich tech workers to look at their ads," because it seems plausibly ham-fisted.
152: I wonder if they could replace in-city buses cheaply. I wonder if that would be a good thing. Probably. Probably the benefit of most people using the lightest possible vehicle, and everywhere having transit-ish access, and minimizing the amount of the city given over to parking, would be better than buses empty part of the time and the routes reinforcing stable neighborhoods.
My kid was invited to a (permitted) birthday party on that soccer field last month. That kid from his preschool class: such a douche.
There's actually lots more non-snarky information in this blog post. Apparently the field has "open play" times and reserved times. I would expect regulars to be familiar with the general rules.
153, 154 completely right. Also the prop re new playing fields out near ocean beach is basically motivated by the downwardly mobile overeducated white surfer bro community trying to keep Hispanic soccer players out of "their" part of GG park. Someone who attended a community meeting on the plans for the fields told me the now proponents of the measure said "those people travel in packs" - charming.
Probably the benefit of anything up to and including dogcarts drawn by shelter pitbulls would be better than Muni, to put this back in tedious local context.
I seem to have screwed up the link in 152: http://recode.net/2014/05/27/googles-new-self-driving-car-ditches-the-steering-wheel/
I love Muni, but I recognize that it's delusional.
156: Are you deranged? Where is the app for dealing with the aftermath to Friday night's party overlapping with the onramp to Saturday night's party on the 19 Polk? And btw what the fuck is wrong with a decent paying union job?
161: As I understand it, Prop A would authorize a bond to pay for about 5% of Muni's infrastructure investment needs. Whoop!
Somebody should do the Clifford Geertz-thing on an American soccer field. Ideal, somebody from Bali.
153: I'm not wholly convinced Heebie Town is a good model for the mission.
Hey, Minivet: http://www.avclub.com/article/archer-drops-isis-name-suddenly-finds-itself-bunch-210378
(Kindly interpolate pause-play around 166.)
165: I bet there are some faint differences if you stare long enough. But surely there's a digital divide thing going on?
165: The intuition that there may be unequal familiarity with/access to apps and the Internet is probably sound, but yeah something strikes me as a bit condescending about 153.
Hard to believe but there truly are a lot of poor people still living in SF. A propos - recent immigrant from Mongolia turned up for mock trial this year!
It is par for the course to work with junior high and high school students here who have no access to a computer (at junior high not even in school), parents don't speak any English, I know a decent number of kids who couch surf through high school as home life too unstable, nonexistent ir traumatic. No phone totally common for younger ones.
171: You're jumping to the conclusion that the largely Hispanic community that uses the soccer fields can't be expected to understand the park's rules, schedule and reservation system. I understand what you're getting at but you're assuming facts not in evidence.
And sometimes the live within blocks of twitter hq! Imagine fucking that!
Imagine fucking that!
The whole headquarters? Sounds exhausting.
I'm using some halfway jumps, specific to my town: that the Hispanic community is a very high poverty community, and that there's a big problem with people in poverty not being able to develop basic fluency with technology.
I definitely don't know that being Hispanic is correlated with being poor in the Mission (district? neighborhood? arena?) but certainly working at Dropbox is correlated with not being poor.
You're jumping to the conclusion that the largely Hispanic community that uses the soccer fields can't be expected to understand the park's rules, schedule and reservation system.
Specifically, they're perfectly able to understand rules and reservations - as long as they can access them. If the city wanted to implement such a thing, they should have posted plenty of signs and had the system be on a paper sign-up sheet, on the kind of bulletin board with the little glass weather-shielding door.
DQ: You seem to be working yourself into a pretty high dudgeon against a straw man. I apologize if something I said seemed glib about people living in poverty in SF.
No easy access to the web is a bit different from "the very idea of reserving this field is alien to me." To be sure, having a web-only reservation system (if that's what they have) seems like a bad idea.
To be sure, having a web-only reservation system (if that's what they have) seems like a bad idea.
This is the assumption that I've been coasting on! Out of laziness and desire to argue? Perhaps.
I've also been assuming that the system is 1) new, and 2) poorly advertised.
And that the Dropbox guys knew they'd lose if they tried to get in via winner-stays-on, because white guys just can't jump, am I right?
SF parks and rec traditionally really poorly managed but that has changed a lot in recent years. Currently demand for soccer playing fields so far outstrips demand that it appears inexhaustible. Every time they open up some marginal strip for reservations it is instantaneously oversubscribed. As far as I know organized play signs up for everything everywhere, and little kids/youth get priority. Young adult / adult pickup play is going to be squeezed out in these circumstances. Some park and rec failure may well lie behind this.
None of that detracts from the the social dislocation currently having a disproportionate impact on recent immigrants and poor people in general in SF.
There's a sign explaining when the permitted hours are.
The thing I don't recognize from the description of Heebie-town is the presumed lack of English fluency in the Latino community. Here, while there are certainly plenty of recent immigrants with weak or no English, you'd be surprised to find a bunch of Latino guys hanging out where no one spoke pretty good English, enough to manage the reservations system just fine.
And the sign gives a number to make reservations by phone.
162 has got to have been to 159, and I think it's just that Muni, like more or less all transit systems here very much including the auto infrastructure, seems to fall into this deadly zone of necessity-but-not-sufficiency where genuine improvements (that don't have serious costs) are very difficult to implement. I'm sure the bond measure would do some good, and I am having a hard time envisioning an app that would improve on the dogcarts. I guess it's just the can't-improve/ can't-replace quality of not just Muni (cheap shot) but every poorly-interlocking system here that I find maddening.
you'd be surprised to find a bunch of Latino guys hanging out where no one spoke pretty good English, enough to manage the reservations system just fine.
Wait, I definitely did not say this. I said half of them would have shaky English, and I was picturing a reservation system where all the players have to register.
Oh, then I misunderstood you. It never occurred to me that each individual player would have to sign up individually -- I assumed that one organizer gets the permit for the hour.
(Which doesn't necessarily make any sense; it's just what I was picturing. Most likely you just need one captain to sign up.)
Anyway, the article in 183 paints a more measured picture. It's not clear to me that there are free hours or if it's all pay-to-play.
If you read the linked sign, it's open play M, W, F -- permits are Tuesday and Thursday only during the week, looks like.
Sounded like there were only two blocks a week (on Tuesday and Thursday) where the permitting was in effect.
That is, from 6:30 to closing. I don't know what happens before that time.
It just says that MWF is "all age play", not necessarily whether it costs money to reserve it or not. At the bottom it says that permit holders have priority at all times, and that permits will only be issued during open play for special events. But the Dropbox guys apparently paid $27? So they qualify as a special event?
I read elsewhere that the permitting was only Tuesday and Thursday. The parsimonious reading seems to me to be that this happened on a Tuesday or a Thursday.
Yeah, what Tweety said. If the issue was that the permit guys were showing up on an open-play day, I think the day of the week would have come up in the video -- that the argument would have been "WTF you have a permit, this is Wednesday." That was not how it went down.
I added value by pointing out that there was a sign, for people who didn't click through on 157. So pbbbtbt.
Guys buy something. Expect to get what they bought. Internets outraged.
162 was to 156 and the was least conclusion that greater automation that would deprive people of unionized employment would of course be an improvement.
Muni's work rules need a serious overhaul but the correct "app" for that is political skill, coalition building and hard negotiations. The level of service provided by the system in relation to resources available and numbers served is extremely high.
198: The kids on the ballfield were not representatives of Parks & Rec*, is the problem.
* Actually Rec & Parks here, I believe.
I'm probably a bit dense here, but I don't see what that has to do with anything.
162:
I can't figure out what the first half of this means, unless it's literally that you think transit should be optimized for the drinking districts? Why, bread and circuses? Of all the things that should pay for themselves...
Anyway, I'm not in SF, and don't remember Muni being all that great for people not living in the right place anyway, and the city I'm in has even more horribly-underserved almost-can't-commute-by-transit neighborhoods. If tiny lightweight self-driving cars will get more people where they need to be for less cost, that's my vote; I don't think they should just replace buses, but most single-user cars. Obviously the jobs involved should be unionized; no reason it can't be a city utility.
Also the prop re new playing fields out near ocean beach is basically motivated by the downwardly mobile overeducated white surfer bro community trying to keep Hispanic soccer players out of "their" part of GG park. Someone who attended a community meeting on the plans for the fields told me the now proponents of the measure said "those people travel in packs" - charming.
What kills me about this is that I see people slagging on the Dropbox guys and then turning around and saying "vote against the new field project by the Beach Chalet". I understand worrying about the privatization of the parks, but as you say in 182 there's such ridiculous demand for fields that *something* has to be done.
And this gets at one of the other things that drives me to despair about the future of the Bay Area: this is a regional problem, not a municipal one. Additional fields in SF benefit me even though I live in the East Bay, but I can't vote on Props H and I.
* Actually Rec & Parks here, I believe.
Leslie Knope strikes again. Remember when she couldn't figure out who the coaches were for the competing boys and girls soccer teams who tried to use the same field at the same time? I'm glad they recalled her.
204: You have it backwards. The prop I an attempt to stop a previously approved project to renovate and expand access to the fields, eg by installing lights to allow nighttime play. Among other things the picnic facilities because they don't want families from the Easter neighborhoods making a day of it.
206: Right, that's what I meant. Substitute "vote to block" for "vote against" if you prefer.
Is the Bay Area unique among major metropolitan areas in having a central city with so little of the region's total population and so little of the region's physical space?
209: I think so, yes, especially the physical space part.
209: the Boston area is a bit like that but doesn't have remotely same geographic constraints.
I blame the 1856 California legislature.
In fact, the city of Boston has much less land area and a much smaller population, even though the MSA is slightly bigger than the SF MSA. So more similar than I thought, but the nature of the peninsula makes things seem much more stark in the SFBA.
Does the SF MSA you're looking at include the south bay? Lack of paying much attention to the area south of Fremont/Palo Alto into while planning seems to be a real problem for the Bay Area, historically and currently.
Looks like the OMB has decided to say, "Nine county Bay Area? Fuck it, we're going with 12 counties."
209: Pittsburgh is a city of 300,000 at the center of a metro area of over two million. In terms of land area, the city is probably a smaller part of the region.
214: not mostly. The SFBA CSA is bigger than the Boston (really eastern MA plus most of Rhode Island and the populated parts of NH) CSA but not by that much (about 400k).
216: I said major!
This list of US metropolitan areas makes me think DC, Bay Area, and Boston are roughly similar. All have tech sectors outside of the central city.
... but again, the fact that you can easily walk into Boston from a number of the most populated neighboring municipalities really changes things.
Right, SF is more of an island than the inaccurately-named Monster Island.
206, 208: Because our proposition system is completely fucking insane (surprise!), "yes" on Prop H is "no" to the ballfields and "yes" on Prop I is "yes" to the ballfields. You may wonder, "what happens if both propositions pass?" From what I understand, Prop I trumps Prop H; "yes" on both is "yes" to the ballfields.
But "no" on Prop I is not "no" to the ballfields. "No" on both is also "yes" to the ballfields. The only combination that kills that ballfields is "yes" on H and "no" on I. Phew.
221: if they both win, whichever has more votes wins.
It's weird to me that the SF MSA doesn't include San Jose. The NY and LA MSAs seem more all-encompassing.
225: San Jose is a much bigger city than SF, and they didn't used to have much to do with each other, I think?
223: That's right, but I'm not sure I understand why. The two propositions are plainly in conflict, but Prop I doesn't explicitly mention the GG fields; it just says "the City should be allowed to make reasonable renovations to park amenities, if usage and environmental studies back up their plans." I would think that mean the City could ignore Prop H.
But whatever. I don't even want to understand how this system works.
Yawnoc is just going to use the private soccer fields on campus.
226: It's larger by population, but only by about 20%: 1 million vs. 800k. Given that their economies are linked but those are two very separate economic hubs, having them be in a CSA but separate MSAs is reasonable to me.
226: Yes, but I don't think there's any clear drop-off in density between one and the other, so who's to say who is the master and who is the slave? There is a CSA covering the Greater Bay Area, so I guess that's good enough.
230: I assume the artificially low density on the peninsula is part of the story.
I think up until about the 1950s, San Jose/Santa Clara County didn't have much of an impact on the larger region. Then it kind of exploded, population and development wise.
Miami is a city of 420,000 on 36 sq miles of land in a metro area of 5.5 million. But the explanation for that is very different.
even relatively low numbers of them would make a huge positive difference in highway congestion at which point "drivers using alternative routes begin to use the expanded highway; those previously traveling at off-peak times (either immediately before or after the peak) shift to the peak (rescheduling behavior as defined previously); and public transport users shift to driving." Another bright single-car idea foiled yet again by triple convergence theory.
234: ah, gee, that's wonderful. How did I never know about that before?
So it won't decrease the net commuting time but it will allow more people to use the existing infrastructure.
236: except in practice it doesn't really seem to work that way.
Travel in towns: jam yesterday, jam today and jam tomorrow?
New mouseover?
I think the correct argument is that driverless cars would make car ownership drastically less necessary. Imagine a robot Uber that could quickly and cheaply take you wherever you need to go; you never have to pay for parking or overhaul the transmission. You might have more-or-less exactly the same traffic volume, but it might be more orderly. And you could do with far fewer parking lots, which would bring about the Yglesian Utopia.
Have you seen the load of shit people have in their car? I don't think a robot Uber is going to be a perfect substitute.
It's weird that I always think of it as "robot Uber" instead of "robot taxis" but the part where you summon them with a computer and they quickly and reliably actually show up is salient.
Do people keep loads of shit in their car because they really need to or because who gives a fuck?
Alternatively: "I'd sell my car but where would I keep all my shit?" said no one, ever (who wasn't already sleeping in their car).
Right, but I bet I'd never, ever remember to bring the reusable grocery bags if I didn't keep them in a giant pile in the car.
I only keep car-related stuff in my car. Mostly jumper cables and tires.
239, 241: why's it gotta be robots?
I keep some spare Zardoz puke in the car, lately. That might not be a good idea.
You keep four tires in your car?
I maintain that the in/on distinction is crucial here.
I maintain a poise of surprise and pwnededness.
Is this a thing in Alaska or did it turn you into a redneck?
Your apartments don't come with storage for snow tires?
Mine doesn't, at least. I'm sure some do.
And or course most people have garages.
It's a thing for people in Alaska who don't have a place to store their snow tires?
I mean, I assume it is. I admittedly don't know anyone else in this situation.
Or it's a thing for everyone for some other reason? If so, why?
234: This looks like one of those a priori economic arguments that attempts to do one thing will have the opposite effect. It like the argument that goes "Trying to get people to conserve will lower the price of the resource and make people use it more."
The difference is just that what you are trying to get people to conserve is efficient ways of getting from A to B. "Oh if you relieve congestion on the roads, people will use them more, so the congestion will come back."
I distrust these arguments when libertarians use them to claim that any effort to promote environmental consciousness is self defeating. I will also distrust them in this context.
I also have questions about how you fit four tires in a car, assuming it's not a wagon or SUV and that the tires are mounted on wheels.
264: These sorts of arguments really have to be evaluated empirically on a case-by-case basis. This particular one has lots of empirical support.
265: It's a wagon; there's plenty of space.
This reminds me that I got new tires a month ago but haven't gotten the wheels aligned/checked.
239 because we would charge Shoup -level prices for parking, or the actual opportunity cost, whichever is higher. Robocars could connect poorly bussed neighborhoods to transit stops w/o the parking lots. One could volunteer to share rides to save money.
City Carshare provides reusable bags in its cars. Baby seats would be trickier.
There's a guy at work who leaves a car, covered with a dust cover, in the parking lot. He has a different car where he lives and uses the two car strategy to be able to drive if he needs to when he's near work, but he rarely needs to since he commutes by bike + train. The car he leaves at work is a 1968 something-or-other station wagon that he might have had since 1968.
It seems like you'd have to have a massive fleet of cars of many varieties to make the robocar idea have an impact at peak times. Or get employers and schools to accept people arriving and departing in a slow trickle throughout the day. It wouldn't take much hassle to tip people back to taking their own cars.
Also, the cars should be called clown cloud cars, since that seems to be the model we're working off of.
Or wait, what about Amazon Car Services? Ride with the packages! That's an innovation worthy of rural postal routes dating back to the early 20th century.
The Amazon Glacial Car: it never gets in the carpool lane, but it's cheap and it will get you there eventually. Unfortunately, you cannot leave the car within one hour of entering it without paying a penalty.
When our old car died, I just got a Zipcar membership in case both of us needed a car at the same time. So far, I've never used a Zipcar, but figure the option is worth the $25.
I have only heard triple convergence theory brought up in response to freeway widening projects, e.g. LA's beloved the 405. It seems to fit for the self-driving car benefit, but it could be that there's an appropriate caveat.
274 shows a real commitment to the use of articles with freeways.
Anyhow, I knew about the thing where additional freeway capacity just produces more traffic, but I didn't know it had a name and hadn't thought about it in the context of self-driving cars and congestion reduction. Thinking about it, I'm not completely sure it applies, but it easily might.
Anyhow the congestion argument is only an argument for self-driving cars over non-self-driving cars. It says nothing about whether self-driving cars are a better solution than these obviously to good to be true "trains" people keep claiming are a real thing.
My brat to parent at soccer: They seem to be building a house near the golf driving range, that's a strange place for a house, what if someone hooks the ball badly and it goes through the window? I think they should build robots to teleport in when balls are going to go through the window and catch them.
Parent: Oh right, but robots are from the fuuutuuure, aren't they?
Myself, mentally: no you idiot teleporters are from the future robots have been selling you books and killing people in Pakistan for years now.
Myself, verbally: hey it looks like the game is starting, come on brat!
Good lord soccer people are supposed to be the damn ruling classes or at least their technocratic servants aren't they?
Oh, conflated. It's the servants' job to know which technologies are from the future, not the masters'.
Doesn't that leave them vulnerable to disruption by another elite that picked a different horse? Or not if you are good enough at picking servants?
Do elites ever really disrupt each other? It seems to me that whenever a new elite emerges, the old elite gets a cut.
Robots are from the future. Just, so are we.
I think thanks to steampunk we all understand that robots are from the past. The future is all slight variations on Yik Yak.
280: As a class yes, but individuals can get thrown off the gravy train and under a guillotine.
If you use it to chop things before putting gravy on them, it's a mandoline.
282: It's robots on Yik Yak making undergraduates cry by telling them they are showing too much fat thigh.
Re: storing stuff in the car.
All of my boxing/kickboxing gear lives in the car. Not heavy but two biggish sports bags. Nowhere in the flat to store it.
Re: storing stuff in the car.
All of my boxing/kickboxing gear lives in the car. Not heavy but two biggish sports bags. Nowhere in the flat to store it.
Do you have room in the flat for four extra tires?
276.2: Trains are inconvenient because you (a) have to get to/from the station on both ends, (b) service may not be frequent or align with your schedule and (c) it's nearly always slower than the alternative* (partly due to (a)). Robotaxis are (a) door-to-door and (b) on-demand. Assuming they're cheap and plentiful and a car trip would be competitive anyway, it seems they beat trains hands-down.
* Obviously, this doesn't apply inside New York City or [your favorite example of a great public-transportation city]. Obviously, it does apply between San Francisco and San Mateo.
Further, to 289.last: This does apply to New York and it's exurbs, actually. I used to reverse-commute to Summit and it was Not Fun.
Walking is good for you, even though people keep trying to avoid it.
289: So upgrading the train infrastructure in the Bay Area to New York standards is so impossible that it's easier to invent a science-fiction solution instead?
Do you have room in the flat for four extra tires?
If they're flat.
292: Actually, given American politics, that sounds about right.
292 is what always gets me about the self-driving car discussion. We have a dirty, inefficient mode of transport that wreaks havoc on the landscape and communities, and a relatively safer, more efficient, less polluting mode, and we decide to pour all our brainpower into committing ourselves more tightly to the former. Not to mention that upgrading the rail infrastructure (and coming up with ideas for better rail) would employ people for decades. And then we act like, oh, if those poor benighted people in the 50s could have seen what car culture would become, they would surely have done things differently.
292: Yeah, that's right. Getting from where we are, after 50+ years of under-investment, to something sensible is genuinely hard, even setting aside all the crazy and deeply-entrenched NIMBYism. I honestly think you'll have a self-driving car long before the Peninsula BART reaches San Jose.
The good news, o, is that the technological advances necessary to create cheap, plentiful, and safe robotaxis will also make a very substantial percentage of the people now commuting redundant. So they won't need to commute.
(They could go to the cities to panhandle, except that the door-to-door robotaxis will deliver the rich from one bit of private property to another, so there won't be places where that makes sense either.)
If they were asking me how society's resources ought to be invested, I'd talk up substantial and continuing advances on the technology infrastructure for telecommuting. And some incentives: anyone who commits to stay home gets a free scanner, with replacements/upgrades at decent intervals; employers get subsidized for allowing telecommuting, etc.
BART to San Jose plans are all via Fremont.
SF employers ought to be paid not to have their employees come into the City in the same way we pay farmers not to grow wheat.
298: Yeah, I know. That in itself is proof of the problem. It also won't be getting to Palo Alto, or Redwood City, or San Mateo any time soon.
296: safe and efficient self-driving cars are far harder. Not-terribly-safe self-driving cars that exacerbate current problems are still very hard, but arguably the easiest path. They solve very few -- if any -- of the current problems with our infrastructure, but they are arguably the easiest path.
And I'll note that in lieu of self-driving cars the big employers in the valley are attempting to ameliorate commuting issues with... busses! Amazing technology, the bus.
We will see self-driving cars on city streets when we learn how to build human-level artificial intelligence. At which point "commuting to work" will become a superfluous activity.
And I'll note that in lieu of self-driving cars the big employers in the valley are attempting to ameliorate commuting issues with... busses! Amazing technology, the bus.
I was just talking to a friend who specializes in transit/transportation infrastructure about the fucking disaster that is regional planning here, and she said that Apple's new campus is a catastrophe transit-wise. Apparently they would rather pay for private buses for their employees than make the new offices transit-accessible.
I have no idea what motivates that other than a "fuck you I've got mine" attitude.
We will see self-driving cars on city streets when we learn how to build human-level artificial intelligence.
We mostly agree, but I don't think this is a terribly meaningful or necessary rubric.
If you add in the (variable) time it would have taken to park, I'm far from sure that driving to Stanford from the lower Haight would have been faster, when I lived there, than biking to and from Caltrain and taking the baby bullet.
when we learn how to build human-level artificial intelligence.
A mere engineering problem...
Maybe this is a pointlessly complicated way to think of it, but I think of it in analogy to complexity classes in the theory of computation. There are problems that are the same order of difficulty of human AI, such that if we solve one we solve them all, vaguely in the same way that if we find a polynomial time algorithm for any one NP-complete problem, we've found one for all of them. I suspect safe driverless cars on city streets is one such problem.
308: oh, I see. I'm still not completely sure I agree, but that seems like a reasonable way to think about it.
I don't think self-driving cars solve (or are meant to solve) traffic and infrastructure problems. They're theoretically safer and could lead to less real estate dedicated to parking.
They're theoretically safer
For whom?
I can't wait for the first sci-fi story from the perspective of a self-driving car, end its day late and heading back to the giant car boarding house only to rise at 4am (the boarding house being far from the city) for the day to begin again, then ferrying unappreciative jerks around all day.
On twitter t'other day I suggested that amazon could take the wind out of the sails of both uber and google by launching mechanical turk for cars.
312: Asimov had one from the perspective who ran a sanctuary for retired self-driving cars, if that's close enough.
mechanical turk for cars
A fleet of cars driven by people who are paid based on the amount of driving they do? Is this somehow different from "taxis"? Or is that the joke?
No, it's way different from taxis, in that taxis are regulated and restricted in number by law. But it's the same thing as "Uber".
Oh wait we could do cars that are driven by humans, but remotely. Like somewhere in India there's a warehouse employing ten thousand people to drive cars that are actually, physically, in San Francisco.
Man, can you imagine how fun 318 would be? Live your Bullitt fantasies with no danger to (your) life or limb!
Customers in SF can be offered a scale of rates depending on how many vehicles their 'driver' is running at any one time.
They're theoretically tautologically safer
Because if they're not, you can't have them on the road, and the fantasy collapses. If, on the other hand, we assume they'll be cheap, plentiful, and safe, then guess what: they'll be cheap, plentiful and safe. Cool.
Ah, I see. I'm less sanguine than that, actually. The nightmare I have is that enough money gets thrown at developing them that the manufacturers essentially have launch them to recover their investment, and push through laws that dramatically limit manufacturer liability, so we end up with inattentive and/or semi-conscious drivers behind the wheels of glitchy self-driving cars prone to copious buggy behavior.
Could be great for my back-up career plan of "expert witness for plaintiffs in suits against driverless car manufacturers", though. Just need to work on that "expert" part.
Yeah 324 seems to exhibit a lot less bloody-mindedness than I tend to expect from Charley.
Well, are the passengers in self-driving cars going to be libertarians? If so, some bloodshed just might be a sad but necessary step.
A properly libertarian self-driving car would refuse to drive anyone else.
A lot of the anxiety about self-driving cars assumes that humans are actually good at driving whereas the car majority of accidents are presumably cause by driver error [citation needed]. A self-driving car will never get angry or tailgate or change lanes without signaling or turn off its sensors to answer a phone call. A lot of the scary cases are essentially trolley problems where (a) if you drive cautiously it won't happen and (b) a human being is not going to make a "rational " choice between mowing down granny or the baby carriage either.
329 doesn't describe my concerns, for what that's worth.
329(b) would probably be more persuasive to someone who's never owned a computer, or a smart phone, or any other device controlled by software.
You don't think a huge proportion of accidents are probably caused by somebody doing something illegal or or otherwise careless? And another big chunk couldn't be avoided with the use of omnidirectional sensors not limited to human perceptual ranges?
omnidirectional sensors not limited to human perceptual ranges?
omnidirectional sensors not limited to human perceptual ranges?
This is begging the question and a half, man. Human perception is an astoundingly, exquisitely well-tuned system for extracting salient information from the local environment. No computer system -- whatever its range, whatever its directionality -- comes remotely close to being able to build as rich or informative a picture of the world as the sensory apparatus of, like, a month-old baby wearing sunglasses. And we don't let babies wearing sunglasses drive.
You're talking about, like, video processing. I'm talking about 360-degree video, plus sonar, plus lidar. Just last week, we were talking about how easy it is to turn right into a crosswalk while your head is turned left. Computers don't have to choose which way to look.
Yeah, but computers can only figure out there's a crosswalk 40% of the time unless they have a map which already tells them it's there. And they don't know the walk sign is lit up unless they get a seperate signal. And they don't know if somebody walking towards the curb is likely to cross the street against the light. And they don't know that somebody walking out into the crosswalk is likely to speed up if they think they're not going to make the light. And they don't know that a person on a bicycle riding on the sidewalk is the same category as a person. Like, I get it, the sensors are neat, but first of all LIDAR (and video) don't have remotely the temporal or spatial resolution of human vision and, more importantly, sensors without understanding (which, make no mistake, the google cars do not currently have in any kind of meaningfully comparable-to-humans way) give you nothing but a lot of incomprehensible data.
I mean, I know analogies are banned, but for instance I could say "computers are much better at language because they don't have the capacity limits that people do, and can parse written text far, far faster and also know every language simultaneously", and I wouldn't be wrong, exactly, but I would also be exactly wrong in most of the most important ways.
This XKCD seems to get at what Sifu is saying.
338: But it's almost like sensation and perception are two different things! Whodathunkit?
It's also worth nothing that one of the reasons we have problems with human drivers going right across crosswalks without checking the other direction is that humans have a really sophisticated set of expectations about where things will come from and how they'll behave. Those expectations fail sometimes in important ways but it's pretty easy to make the case that without an equally robust set of expectations you'll never get anywhere. So the question is not whether a self-driving car can be made that will not make the mistakes humans make, it is what mistakes the self-driving car will make, and whether they will be predictable enough for humans in the vicinity to avoid disaster.
I should probably really shut up about now, but I'll just say, "there is a thing that is moving and I should try not to hit it," is not _such_ a hard problem and there's a bit of perceptual goal-post-moving in 339.
I just want to be able to get home from drinking without have to break the law. Public transit, robot cars, or greater density all work equally well for me.
Speaking of buses to go to bars and programing tasks that are now possible, this is pretty cool.
344: it really is a very hard problem, especially if figuring out what the item is is important. But even, frame to frame, saying "there is an object that is X size and moving in Y direction" is neither an easy nor a solved problem.
But, sure, if you're comfortable with cars that can be catastrophically DOS'd by leaves blowing in the road, that is probably medium-term achievable.
I really don't agree with the assumption in 343. You don't want to to train a computer to say, "Based on my experience, I don't have to check my blind spot in this situation." It is probably overall safer for everybody if the computer says, "I'll believe what my sensors tell me," and doesn't sneak in any extra assumptions. (Tempting the analogy ban-hammer, I'm playing Sam Wang to your Nate Silver.)
I feel like watching the 58 do a complete circuit because its route has never made the slightest bit of sense to me.
You can't have perception without assumptions.
So which one is easier, figuring out protein folding or going from raw camera/lidar/whatever data to deciding whether you're about to hit an object you don't want to hit?
Also, 348 seems very beside the point. The point is that "believe what my sensors tell me" won't work by itself.
The assumption is, "I shouldn't hit other things that are in the road." Figuring out what they are, exactly, isn't so important.
The other things are moving. If you don't make assumptions about plausible trajectories, you can't move.
And figuring out what the other things are is a very good way of figuring out what they are likely to do next.
Maybe I'm the opposite of the guy who thinks he can waltz into some other field and everything is trivial; to me, all of these things sound impossibly hard! I mean, if I were to try to break down the problem of "given a camera image of the road ahead of me, how do I figure out if a pedestrian is about to cross into the street?", it just seems to be made out of lots of different hard problems. I have to parse the image into objects, and not just as a static problem, but I have to identify that the different parsed components are somehow continuous from one instant to the next. If I want to know which objects are moving, I have to correct for my own motion, because even a static object is moving across my visual field. And if I want to know if they're moving toward the street, I have to know something about perspective: if they're just walking along the sidewalk then when I project onto a two-dimensional image it still looks like they're moving toward the road just because parallel lines are converging toward the horizon. I also have to have some gauge of how far away they are, which is probably going to rely on lots of priors about the sizes of different kinds of objects, which requires not just that I've parsed the scene into objects but that I've categorized some of them into examples of things whose size I know. A lot of those pieces sound like hard problems in their own right. Fucking brains, how do they work?
And which ones of them you do want to hit as opposed to doing anything unexpected (raindrops, leaves, paper bags).
But what is "a thing"? That's Sifu's point about leaves, right? Or rain, or maybe a plastic bag, or maybe a cardboard box, or maybe a sheet of corrugated iron that's just flown off a truck, or whatever.
Fucking brains, how do they work?
And why can't I persuade mine to focus on the grant proposal I'm trying to finish this weekend?
359: Sounds like we need some robot grantwriters.
In fact, you know what's not a completely trivial problem, as evidenced by lots of video games? You're given some representation of the three-dimensional positions and shapes of a bunch of objects in a scene, and some of them are moving. Your task: figure out when two objects have collided. Even that turns out to be kind of a pain! Now imagine: you're not actually given all of that three-dimensional information, just some weird two-dimensional projection of it. Oops! Just got harder.
cars that can be catastrophically DOS'd by leaves blowing in the road, that is probably medium-term achievable.
Easy solution, we just cut down all the trees within a mile of our robot-car designated roads.
Last time I submitted an application for this particular grant I got a really snarky review about how I was trying to do so many different things that I would probably fail at all of them. (It's supposed to be a five year plan.) So this time I'm trying the opposite extreme of writing in detail about a bunch of things that I'm already 50% or so finished with, all of which I can probably finish within two years. I assume it'll get trashed for the opposite reason.
Seriously, how many non-flat objects do you come across where you have to exercise your finely-evolved human senses to determine, "it's OK to plow through that"? Do you encounter so many floating plastic bags in the road that slowing down for them out of an excess of caution would be an unconscionable waste of your time?
What? On a fairly frequent basis I encounter and drive, run, or walk through leaves, branches, high grasses, and piles of snow.
Especially snow. My wife and I have different rules about the branches. My rule is I won't miss a good parking spot just for some overhanging tree limbs. Her rule gives a higher priority to the paint.
She also drives over grass less often than me, but I think I'm still within the plausibly not-dangerous driver range on that.
I don't know about you, but when I drive, I have to look not only at what's currently in the road but what might be in the road momentarily.
I live in what some people consider a major city.
Also, what if a bird flies low across the road somewhere ahead of you, or a squirrel is darting across? Would you have the self-driving car screech to a halt even if it's clearly going to be nowhere nearby by the time you get to its location? If not, you have to make some nontrivial inferences about what you're seeing, right?
The problem deep down is that sensors don't actually tell you anything at all: that's what all the assumptions are for in the first place. And human beings have a really sophisticated bit of processing gear built to work efficiently and quickly, to minimize the amount of raw data needed, to override the general procedure in certain specific, dangerous cases (objects moving towards us at rapid speed; snakes; etc.), to maintain and track objects in the visual field even though we aren't actually getting data on them the majority of the time, and so on. And with that massively complicated apparatus assembled through millions of years of strong evolutionary pressures we fuck up constantly.
With the cars we're talking about something that we're building over a very brief period of time and giving at best very limited testing, out of things that as often as not we don't actually understand that well in the first place, and for which there are massive highly motivated* groups acting to subvert exactly what you're doing for their own purposes*. And we're going to use it to move multi-ton pieces of metal through densely populated areas at incredible speed. And there are just so many different problems that would need to be solved: what about conflicting data from different sensors, caused by odd lighting, or something reflective in a way that the engineers didn't predict, or a smudge on the side of the car? Or what happens when there's a minor impact on the front of the car when the visual sensor doesn't pick anything up (did the sensor glitch, or the light hit it wrong, or some kid move in a way the apparatus that predicts kid movements got wrong, or a bird poop on the fender? (The car would have to decide well before it had enough data to distinguish between the two.)
*Mostly lolz.
Seriously, how many non-flat objects do you come across where you have to exercise your finely-evolved human senses to determine, "it's OK to plow through that"?
This is hardly uncommon on freeways.
Realistically I think the major ethical question wouldn't be something like "hit old lady or hit minivan full of children?" or "intentionally collide with car in other lane to knock it out the way before it hits a pedestrian?". It would be way less fancy sounding things like "should the car slam to an emergency stop every fifteen minutes and leave all the occupants with neck injuries, or should it run over people, animals, and various solid obstructions constantly, killing people and probably also injuring all the occupants?*". (The correct answer is probably "no".)
For some reason 375 has put me in mind of the following, from §33 of Intention:
Consider the following:
Do everything conducive to not having a car crash.
Such-and-such will be conducive to not having a car crash.
Ergo: do such-and-such.… But this syllogism suffers from the disadvantage that the first, universal, premise is an insane one, which no one could accept for a moment if he thought out what it meant. For there are usually a hundred different and incompatible things conducive to not having a car crash; such, as, perhaps, driving into the private gateway immediately on your left and abandoning your car there, and driving into the private gateway immediately on your right and abandoning the car there.
Topically, most of the time when I drive into leaves or grass, it's because I'm parking at the soccer field.
Maybe I'll get one of those self-parking cars. I mostly suck at parallel parking.
334 No, I agree that most vehicle accidents are human, rather than mechanical, error. What I think needs citation is the proposition that most of the objections to robotcars is that they'll be mowing people down. It seems to me that people don't think they'll work well enough to be used, and that even if they do, it'll be a bad solution to the problems they are supposed to solve.
I stand by my Delorean comment above.
(I'd have commented sooner, but was out riding my bike. Got a pretty nice picture of a rainbow butt with my phone, but it's not uploading to FB for some reason. Compared to the world of 1979, the whole thing is a miracle. Compared to perfection, it's one frustration after another.)
I missed "rainbow" the first time I read that.
Self-driving multiton vehicles at high speeds: nah, I don't want to make long fast SOV commutes easier. Smart Fortwos (1900lb curb weight) and drop city speed limits to 20mph and build mass transit for the carlets to feed, my city gets much nicer over a wider area. Maybe they only self-drive to load balance.
Trains feeding the metropolis worked with separate drivers at the periphery - hence kiss-and-rides and the 'station wagon'.
If they drop city speed limits to 20 mph (which I would think is great), I'm getting an electric golf cart and using that for all in-town trips.
Golf carts, bicycle surreys, segways, kick scooters. Make most streets safer for bodies and ride ponies. (Chip them to spend their downtime grazing municipal berms. There must be blood, it must be innocent.)
There is a woman who goes around Oakland by Segway.
382, 384: all of that is great, for sure, but I'm not sure how it requires self-driving cars.
Anyhow, essear gets it right. That Google has gotten as far as they have is genuinely impressive, but this is the kind of problem where getting 80% of the way there doesn't mean that much.
I ate three bowls of chili and now I feel sort of bloated.
A lot of the scary cases are essentially trolley problems where (a) if you drive cautiously it won't happen and (b) a human being is not going to make a "rational " choice between mowing down granny or the baby carriage either.
The trolley problem stuff is a load of old bollocks; Sifu's points about this are much more important and interesting. Self-driving cars won't get angry or make phone calls, sure, but they will do the equivalent of saccading through someone pulling out, just we don't know what the equivalent issue would be, because machine vision is difficult. It's actually much more difficult and interesting whether or not it would detect and acquire the fat Einstein/three future lawyers/whatever crossing the street than what it did after that.
I put it to you that you wouldn't get on an aircraft that was going to navigate based on a map with all the flight plans of all the other planes on it, and no way to look outside, query TCAS, or call air traffic control. That's actually easier than the Google approach to cars (because there is so much less stuff in the sky, which is big).
386: Balancing vehicle distribution, a reward to drivers for not speeding (also enforcement), make the heavy transit useful to more people (drunk, old, tired) in the whole city.
391: I guess? Seems somewhat marginal, compared to all the rest of the (excellent!) changes you propose.
In my experience, riding in a Google self-driving car back in May, the car is far more likely to slow or stop - often in ways that human drivers wouldn't expect - than it is to hit anything. When it doubt, the default is to slow down and put the hazard lights on, unless a human overrides the controls. Which they always did during my ride; it wouldn't look good if your sci-fi advanced cars were always holding up traffic. I later read some article where people mentioned that sometimes you get people trying to mess with the car by deliberately cutting it off, slowing down, etc. It would be hard to imagine any wide-scale implementation of the technology that didn't involve changing how people not in self-driving cars act around self-driving cars.
There was also some glitch during my ride that led the car to repeatedly drop out of self-driving mode. The Google employees in the front were pretty frustrated by the end, and the last part of the drive was pretty much all in human-driving mode. I can't say it was the most impressive display from a this-technology-will-be-ready-soon perspective, but it was impressive from a "hey, it actually drives!" perspective.
often in ways that human drivers wouldn't expect
If you would like to know my actual particular hobbyhorse, this is it. Unless the cars can be made to act meaningfully similarly to humans -- especially in emergency situations -- they're always going to be disastrous in mixed road-use situations. Navigating roads in the context of other road users is a theory of mind task, and introducing a set of road users for whom all of our theory of mind is useless is going to cause total chaos.
394: we already have this problem between drivers, bicyclists & pedestrians.
395: to a limited degree. But that's still people making judgments about other people's intentions, and that's something that people are absurdly good at, and do automatically. What happens when your automatic, anthropomorphizing intuitions about goal-directed behavior intersect with an automated system that is doing things accoring to the output of an almost completely alien value function? It's going to be confusing as hell, like trying to figure out why a Roomba is doing what it's doing except without any knowledge of whether the Roomba is actually under human control or not.
Can't we just take the programming from a cruise missile and add a NOT at the front of every line?
The scenario fake accent mentions (where people try to mess with the car by cutting it off and so on) is really interesting, in this context; another way to read that is other drivers trying to probe the limits of an unknown system so they can get a sense of how it's going to behave. It is beyond foolish to assume that people will treat self-driving cars the way they treat human-driven cars; conflating the task driving-with-other-drivers and driving-with-other-drivers-plus-algorithms is super shortsighted and dangerous.
Anyway, 386: 382.2 doesn't seem relevant?
399: I dunno? If people want to live at the periphery, wouldn't they just keep driving regular cars? If it's a low-traffic outer belt suburb what's the cost?
And devoting enormous amounts of time and energy to self-driving cars so that things are mildly more pleasant for people living out at the edges of regional rail transit seems... like a weird way to allocate resources.
Maybe we'd only need to put a NOT on the final line of the cruise missile program.
Most of the purported weirdness of how self-driving cars drive is that they're unusually cautious in situations where humans ought to be more cautious. It's not exactly an unbridgeable chasm of otherness.
An easily identifiable car that you can reliably cut off with no adverse consequences seems like a good thing.
403: we don't -- can't -- know that yet. My personal sense is that you're wrong, but, yeah, we don't have that data.
(Incidentally the people at google working on this are perfectly aware of -- and working to address -- this issue, by measuring and integrating the behavior of drivers when they shift into self-drive mode.)
Anyhow I don't think 403 is even particularly plausible. Just to come up with one example: there's a railroad crossing near here where the road on either side of the railroad crossing gets backed up. Because the trains don't run very often, (idiot) drivers will often not stop clear of the tracks. Occasionally (I've seen it once) somebody will stop partway on the tracks when the gates come down, and will have to back up. A human driver who is stuck behind said idiot will generally figure out a way to back up, or maybe honk angrily to indicate that they can't back up. What would a google car do? Honking angrily is presumably out of the question. Would the driver who was stuck on the tracks figure out "wait, that's a driverless car, I have no recourse" quickly enough to come up with another solution or get out of their car? What, in general, does the google car do when a car in front of them unexpectedly starts reversing? Note that this is a situation where what the self-driving car has done is perfectly in accordance with the rules of the road, and the other car is the bad actor. But the outcome is far worse than if all the drivers involved had been humans.
This particular situation, having been described, would be perfectly easy to deal with by hard-coding behavior into a self-driving car -- maybe this has already happened! But the set of situations like this is essentially unbounded.
In general the 'more cautious' approach is good, though the overlap between "cases where drivers will overrule the car and drive themselves" and "cases where the driver will overrule the car because it's driving too safely" is likely to be pretty massive to say the least. I think part of the trouble is that for basically every other mechanized/highly computerized thing we basically accept that it's not going to be particularly flexible/is going to be idiosyncratically unreliable. And so we always put in a bit where the human being steps in to take control in emergencies. (Trains are practically automated by now, and run on rails and we still put conductors on them.) But here the selling point is just supposed to be that the car does the emergencies better. And when it comes to doing something complicated, outdoors in all seasons*, and with a massive degree of risk in even not-especially-tricky cases it's hard to see how that would work out well.
*Does anyone else think that California is maybe not the best place to be developing something that's supposed to work outside year round? Even with built in ice/snow routines there's still the problems that come along with having complicated machines that sit outside in -10 degree weather, and undergo rapid heating and cooling on a daily basis. A massively complicated and sophisticated computer working off of multiple very sensitive sensors seems like the sort of thing that might have trouble with that.
we always put in a bit where the human being steps in to take control in emergencies
Not that that's such a great plan, cf. 51.last.
"Dave. I can't allow you to do that."
What would a google car do?
I assume it would be something like: given enough data, the car could leave room for the car in front to move back if needed (by calculating one car-length from the stop line and not going beyond that, probably). Until the car had enough data? Manual override.
Or just look up local body shops while remaining stationary.
Honestly, I'm less skeptical about getting a road-ready self-driving car in the near future than I am of the existence of said cars helping much with traffic congestion, since traffic congestion is much more of a collective problem.
410: well, sure, right. The new in-city google car doesn't have manual mode, though. (My impression is that they were moving away from having a self-driving mode because it caused more problems than it solved?)
It could look up and broadcast appropriate quotes.
"Here I stand. I can do no other."
Anyhow, my point was not "here is a problem that self-driving cars will be unable to surmount". My point was that driving in real environments is a massively complex, importantly social undertaking where the right course of action is only incompletely described by the (formal and informal) rules of the road.
I've only seen the in-city car from a distance, driving around an empty parking lot. It seems like its main problem, in the south bay and on the peninsula at least, if finding a city with low enough speed limits on the main streets.
If it apparently impossible for me to type the correct two letter words beginning with 'i'.
414: "I think you're going to need a bigger dataset."
I don't know what surprises me more. That somebody put the Digimon theme on the jukebox or that somebody paid a dollar to play it.
Is the jukebox connected to that service that you can control with an app? That's the ultimate for trolling people, since you don't even actually have to be at that bar to do it.
If that's you who put in the Power Rangers song, I could probably figure out where you live.
Or you could ask me, if you're that interested. I have beer.
But no. I'm sufficiently milquetoast that the idea of trolling is more enjoyable than actually trolling. I mean, gosh, you could annoy someone and that'd be a little sad. But if I were to troll, and do it via an embarrassing millennial childhood memory, I'd use this.
Never mind. Probably just a hipster or an asshole. Now it's Britney Spears.
I love that Moby recognizes the Digimon themesong, which may be the best part of this story.
Alas I think you'd figure out what it is, too. It isn't a subtle song.
I have Netflix and an eight-year old with shitty taste in cartoons.
Now everybody can pretend they're at the bar with Moby!
I kinda dig it. A little bit '90s breakbeat-y.
Somebody who is at the bar with me just played Barry Manalow at a bar in Shadyside. I did not suggest this.
Playing here is a song that starts, "Sucking on my titties." This probably isn't trolling since it plays fairly often.
Does the song continue with "HOORAY?"
It's a strange song. About sex and staying in school.
This song?. I like that song. I like this remix, too.
That's it. I thought she was saying, "Fuck the same way."
This particular situation, having been described, would be perfectly easy to deal with by hard-coding behavior into a self-driving car
I'm a bit more optimistic when I think about what the process of "hardcoding" is in this situation. The process is 1) hand over the controls to a human. 2) Watch what they do. 3) Remember that process. 4) Next time you get plugged into the internet, share what you have observed with every other car. Given enough data points across enough different cars, a new model should emerge for how to deal with a similar situation.
So when a car experiences a novel situation, but can determine that the parameters were similar to the case where human trainers performed a certain action 95% of the time, its going to perform that same action. And maybe that action will be wrong thing to do 5% of the time. And in some proportion of that 5%, there will be an accident as a result. But this thing doesn't have to be perfect, it just has to have a lower percentage of fuck-ups than humans. And, humans being what they are, that's a far easier goal to reach than 100% safety 100% of the time.
437.2 isn't reallly true, but to be completely honest, I'm pretty self-driving-car-discussion'd out. Do you have any thoughts on the Digimon theme, Spike?
I can tell you that my kid has spent an unfortunate amount of time this weekend playing Pokemon through a Game Boy Advance emulator on his Nexus tablet, and he insists on not turning the volume down, so now I have the goddamn Pokemon game music earworm going on and it won't go away.
So, not the Digimon theme song exactly, but close.
You could implement 437 by having a bunch of different people in India all with remote control of the same car, and the car decides what to do with majority-wins.
Star Wars XII: Attack of the Drone Cars
I'm not sure if 437 would work, I mean, does google even build the things with the ability to give other drivers the finger?
320 to 443? I don't even have a twitch.
The digimon theme is a bit of an earworm, but neither is this. Both are pretty close though.
D'oh.
On the cheap-remote-drivers thing: that was a plot point in one of Charles Stross's near-future books, probably Halting State.
I don't know about robot cars, but at this point I think The Good Wife writers can drive me wherever they want. A show in its sixth season can't be that good, right? It must be hypnotizing me or something.
446: I love the embedded quiz.
"Have you ever had a bug stuck in one of your orifices? YES NO"
There are a surprising number of animals that are into the idea of hanging out inside human beings, so maybe it's not that rare?
453: I think for the purposes of this discussion, yes, but oh god let's not have this discussion anymore. Let's talk about trolling your remote-from-India taxi driver with the theme song to monster hunters, which you have stored in your dropbox space, or really anything else.
I just want a train system that actually goes places and doesn't share track with freight rail and can go the speeds it might as well go and does go in enlightened, carless Europe. That's truly all I want in my life of sometimes going to other places. It strikes me as insanity that SF-LA is going to take fifteen years if it happens at all.
And don't get me started about the Bering Tunnel! Well ok not really about that one.
449 is torment because we have a tv that is not a tv, and I won't see TGW until probably Wednesday. At this point though I don't think Season 6 is as crazy good as Season 5.
And don't get me started about the Bering Tunnel! Well ok not really about that one.
Good, because it's definitely never going to happen.
I agree with you about the American passenger rail system in general, though.
I thought it was going to be a bridge.
It's too far for a bridge. It's probably too far for a tunnel, too.
Plus there's the thousands of miles of rail you'd need to lay on each side just to reach the strait.
I just want to be able to go to Europe without leaving the ground, is that so wrong? Even if it takes like nine weeks!
456 was also me. Of course it was!
Oh, sure, it's a stupid idea all around, no obvious economic reason for it, massive cost, faster and more energy efficient shipping via sea, just really idiotic.
That being said, still awesome and I could imagine a future Russian autocrat deciding on it as a prestige project. ("I could imagine" is the lowest burden of proof I can think of.) It is far for a bridge, but I don't think it'd be "too far"; most of it could be causeway. The difficulty would be dealing with all the hazards of an arctic sea's freeze-thaw cycle on the piers.
Think of how it would stimulate Little Diomede Island's economy!
Yeah, I guess there are actually longer bridges, all in China of course. (There are also longer tunnels, but only for water supply rather than transportation.)
466: They could sure use it, too. Now there's a community with a lot of problems.
Of those my favorite would have to be the Jiaozhou Bay Bridge, because it's completely extraneous. A tunnel was built simultaneously in a much more sensible location.
More so than the usual isolated Alaskan Native village?
I guess having your community and land split in half by an international border is worse than normal.
470: Yes. In addition to the inherent problem that it's so difficult to get there, they've had a lot of challenges recently with operating and maintaining their infrastructure. I'm not sure what the exact underlying issues are that are driving this, but the small size of the community and its isolation surely play a role.
471: Little Diomede is just the one on the American side. Big Diomede is a separate (but closely related) community that no longer exists in its original location.
400, 401: but lots of actual city is dense, close to dense jobs, and not served by transit. Isn't NYC only now building a subway in the last quadrant of Manhattan? We never had a meetup at Cafe Leila because it was just too far to walk. The first line in a city is often under-utilized for years. Etc. Letting people in those neighborhoods use transit or at least not need parking @ both ends: win. Others will take the train if there's an easy cheap car for unpredictable errands, kid pickup, etc.
473: oh, I assumed they were formerly the same community but grew apart due to the sale of Alaska/Stalinist deportations/the Cold War. I should read up more on it.
475: No, they've always been separate villages, but with close connections because they were historically so close to each other.
One example of the strength of the connection: people from there always refer to it as "Little Diomede" whereas everyone else in the state who has reason to refer to it (not many people!) refers to it as just "Diomede."
So where is Big Diomede these days? I mean, I suppose the island is more or less where it's been since Beringia was flooded, but the people?
474: sure, but, again, already-existing carshare/zipcar and bikeshare and taxi/uber-is-a-shithead seem to accomplish this. At least, they accomplish it in my neighborhood, which fits the approximate criteria you mention (until they build the train in a few/ten years).
We don't have a bike share but we do have new bike racks that are shaped like squirrels.
That doesn't seem terribly useful. I don't think of squirrels as having a particulary convenient or high-security shape.
That's a lot more useful than most "arty" bike racks, actually.
The squirrel is the most secure of all the rodents.
Morphologically speaking. Very socially anxious.
The Bering Straight is the perfect location to build the Hyperloop, because cheap land. It would be best to build it now, before everyone moves up there when the globe heats up. You could build it all the way from Seattle to Vladivostok.
Robotaxis are (a) door-to-door and (b) on-demand. Assuming they're cheap and plentiful and a car trip would be competitive anyway, it seems they beat trains hands-down.
You know what's cheap and plentiful? Taxi drivers.
The Bering Straight is the perfect location to build the Hyperloop, because cheap land. It would be best to build it now, before everyone moves up there when the globe heats up. You could build it all the way from Seattle to Vladivostok.
The tectonics might be problematic.
The tectonics might be problematic.
And yet they still want to build one in California.
As I understand it, if they build one in California and wait long enough, it will be in the Bering Straight.
You know what's cheap and plentiful? Taxi drivers.
Clearly you are not a Londoner.
477, 478:
Of course, there was Old Gregor, and his son, Young Gregor. Young Gregor's son was older than Old Gregor. Nobody could figure out how that happened.
||
This seems like a good enough place to note that I've been hanging out with someone who attended the infamous University of North Florida. I haven't yet told her that a forty-three year old balding man living in his parents' basement has been inadvertently landing on her alma mater's web page for years.
|>
492: I'm not the only person who does that? Oh, man, I thought I was so alone.
(Well, a google search that brings up results for it, not the landing page itself. But close enough.)
If my timing's a little more off, I end up at the UN page.
478: The Soviets moved them to somewhere on the mainland and made the island a military base. I don't know where exactly they ended up.
455: LA to SF is starting in the valley, because land is cheap. But the locals suspect it's a boondoggle. Fresno turned down $1 million in funding to plan, twice, before finally accepting the free money.
Honestly, I think a solid "we're going to build this, come hell or high water" would swing a huge group of fence sitters and mild-opposition voters. We just don't expect that it's really going to come true... and what does a length of starter track that goes nowhere do for anyone?
449: Ok this week's is really good.
For the record, the kids at the ballfield knew exactly what was going on and planned this confrontation: http://missionlocal.org/2014/10/planned-soccer-field-standoff-came-after-weeks-of-frustration-with-pay-to-play-policy/
499 doesn't surprise me. It is intelligent on their part, but sure makes the valleywag post more epically tendentious.
