I was diamond medallion on Delta for a while, and one of the perks that comes with that status is that the airline guarantees that they will sell you a seat on any flight, even if it's already excessively overbooked, and (implicitly) even if that means they have to get someone else off. This usually works out fine, because the high-status last-minute purchaser is paying full fare for the ticket and there are plenty of people who are willing to be bought off the plane for significantly less than that. But in rare cases complying with their commitment to the high-status passenger could require them to involuntarily bump someone.
This is exactly the kind of problem that exists in the system: that some passengers are more equal than others. Thus the outrage. The "random" selection of passengers who were asked to deplane presumably did not include any of those more equal people.
From the airline's point of view, some passengers *are* more equal than others. If you get in a situation where you have to pick someone to favor, it's much more important to serve a businessman who spends (well, whose employer spends) $100K per year on flights (many of them full-fare tickets) than a person who takes one deeply-discounted $200 bargain flight every three years. If you lose the latter's business, you haven't lost much. For that matter, the person with the deeply-discounted fare ranks below the person who isn't a frequent flier but bought a more expensive ticket. On many routes, airlines make basically all of their money from the repeat business travelers and they see the other travelers as just a way to avoid flying with unpaid seats. That's why those deep discount fares are often so ridiculously cheap.
I'd guess that if the random selection excluded high-status frequent flyers, it also excluded those who'd paid full fare. In practice, though, I doubt they excluded anyone from the selection pool. Heck, if I were in their shoes and had to pick, I'd pick exclusively people from the second-highest frequent flyer class. I'd want to pick experienced, regular travelers because they'll react much more calmly. That would make the highest frequent flyer class the ideal choices, except that they have that seat guarantee.
no stop signs & stoplights = underpass / overpasses for pedestrians.
Or walk signals for pedestrians, which tell the vehicles to stop. But, yeah, when intersections have free-flowing traffic except for pedestrians there will be a lot of incentive to build pedestrian overpasses or underpasses.
Actually, that should be "a few months' paid leave". "Several" to me implies 5-7, and it's not that much. I haven't looked at the details (I'm long past having kids so it's not relevant to me personally), but I think it's on the order of three months.
"extremely scientific and robust," relying on the employee's role, job level and location, as well as recent performance ratings"
So basically they are claiming that performance ratings are scientific, and that there's no possibility those are biased.
No, they said pay calculation was scientific, not performance ratings. Performance ratings are an input to the pay calculation, and one that inevitably has a subjective element, though Google does a lot to try to identify and eliminate bias -- vastly more than any other company I've worked with/for.
Women who do not have children get paid the same or more. But when you have to take several weeks or months off to take care of a child you slow your career.
Google offers several months' paid leave to both mothers and fathers, and all are strongly encouraged to take it. In part this is due to competitive pressure -- all the big tech companies are fighting over the same pool of employees and it's a really awesome perk, but it's also quite explicitly an attempt to eliminate this aspect of gender preference. Women who give birth do get a little extra time for "medical leave", but the actual maternity/paternity leave is the same. Oh, and it applies to employees who adopt, too (not the medical part, obviously).
If you want to have the biological and emotional fulfillment of giving birth and raising a child then realize you have to sacrifice your overall income.
Not at Google. Oh, I suppose it may slow you a little because the time you're on leave is time you're not doing promotion-worthy things, but in practice it doesn't seem to have any significant effect. I see people with families of all sizes at all levels of the career ladder. That even includes a few quite senior people who take maternity/paternity leave every year. I know one engineer who has had a child every year for six years, and taken all of the leave, and leads a large and important team. I know another who has negotiated a deal with management to accept a 60% salary in exchange for working only three days per week, and also to spread maternity/paternity leave over time, taking one day of it per week, with the net effect of a two day per week work schedule -- and just got a major promotion. That particular engineer is something of a rock star and I'm not sure that sort of deal is generally available (though fractional salary for reduced work schedule is, with management approval). Note that I didn't specify the gender of either of those examples. One is a man and one is a woman; their gender doesn't affect the options available to them.
As a Google employee, my reaction to the DOL claim was "WTF"? The claim is so utterly at odds with the way Google operates.
Here's my guess as to how the DOL came to their conclusion: They just looked at average male and average female compensation, without considering job role. Because women are underrepresented in engineering, and engineering jobs are better-compensated than most other categories, the average female compensation is probably lower. That women are underrepresented in engineering is something Google regularly and publicly discusses, and the company has a wide variety of initiatives aimed at improving that situation, mostly by trying to increase the number of women in the hiring pipeline.
nothing about city traffic will fundamentally change.
It'll change dramatically.
For one, it will largely eliminate parking. No need to park your car on the street in front of your home, have it park somewhere outside the city and have it drive in when you need it. When you go to the store or similar, no need to park your car at the store. It can circle the block, or drive a few miles away to where it can park. This will have the biggest effects not in dense urban areas but in moderately-dense suburban areas. Those huge mall parking lots will become unnecessary. Assuming car ownership stays high, parking will likely consolidate into enormous lots a few minutes' drive from merchants. It's also highly likely that many people living in moderately-dense areas that currently have a car will switch to using self-driving car services instead, since they'll be cheaper and almost as convenient, and they won't need parking at all since when they get out of the vehicle it will go off to service another customer.
In the longer run, cities will likely outlaw manually-operated vehicles and that will have an even more dramatic effect. Mutually-cooperating or centrally-controlled (could go either way) vehicles that communicate via RF have no need of things like stop signs and stoplights. Taken to the logical limit, there's no need for vehicles to stop at intersections at all as long as they can coordinate to arrange for gaps for cross-traffic to pass through -- and the gaps needn't be any larger than the time the crossing vehicle needs to pass through them (which will be terrifying until we get used to it). This will make city traffic smooth and predictable... which may in turn cause it to increase in volume until it's not.
On highways, such vehicles have no need of the "safe following distances" required by human situational non-awareness and slow response times. They'll close up into trains of vehicles inches apart, both to pack more vehicles into a given area of roadway and improve energy-efficiency by drafting. They will also be able to increase speeds significantly without danger, and without great loss of efficiency (due to drafting). I expect that it won't be too long after self-driving vehicles become commonplace that local governments will carve out self-driving only lanes so they can begin to take advantage of the higher efficiencies offered.
People also carpool. That's been around forever.
Sure, but people don't carpool very much, because unless you have a fixed group of people who go the same place every day at the same time, it's too hard to organize. It's often too hard to organize even with such a group of people. Self-driving cars with centralized dispatching systems can handle the organization automatically and dynamically. It seems entirely likely that in the future you'll put what time you want to go to work in your phone and your digital assistant will automatically notify your carpool service of choice. The service will take your requirements along with the requirements of thousands of others and construct routes that incur minimal overhead, and your phone will notify you when your ride is five minutes out. Note that it's not necessary that the vehicle be self-driving for this to work, but it will be a natural extension for self-driving car services, which will help mitigate the fact that ride demand is so bursty. They can purchase small vans which transport a half-dozen people during rush hour and make grocery store runs, etc., at other times. Demand-adjusted pricing will encourage people to take single-rider journeys at off-peak times, and to accept sharing at peak times.
$250 per hour is a pretty normal rate for an experienced professional, hardly outrageous at all. When I worked for IBM Global Services, they billed me out at $300 per hour. Granted that IBM commands a premium due to their marketing channels,
Its not the marketing channels. Its the organization behind you - technical support and resources so that even when you don't personally have the answers you have a direct line into the organization to get answers.
Yeah... no. None of that ever applied to anything I was doing.
The problem is, the technology isn't quite there... and the infrastructure is nowhere near there.
For most people not living in apartments, infrastructure is irrelevant as long as the car has enough range to get them to and from work, plus a bit of running around. If you can charge at home, infrastructure only matters when you go on long trips.
So an EV is fine for someone who pootles about town and keeps the car plugged in for 14 hours a day at home, but for the motorist doing 8000 miles a year
8000 miles per year? That is "pootling about town". I put over 24,000 miles per year on my LEAF (now that it's not artificially constrained by lease limitations). I almost never charge anywhere except home, because there's almost no place to charge where I live. I do have a level 2 charging station in my garage, so a charge from empty takes a little over three hours.
So I doubt the majority of cars on roads in the next decade will be fully electric.
Well, since the majority of cars on the roads ten years from now will be like those that are selling now, that's a very safe bet. On the other hand, I think that it's entirely reasonable to believe that the majority of sales ten years from now will be EVs. Barring regulatory forcing (which could happen), it'll take another decade for them to become the majority of cars on road.
I'm not saying Tesla will fail, but I am saying its over-valued and that will be corrected in time.
That, however, seems entirely likely. I think the Model 3 will be quite popular and that Tesla will do very well. That doesn't mean that Tesla will own the market in a decade, though they'll probably be a significant player. Basically, their current valuation is probably about what their valuation will be in ten years.
In places where you're talking to the owner, or at least a very empowered (and smart) manager or employee, you might be able to get them to knock 3-5% off for a cash payment, but it's rare.
It's also important to note that handling cash is not free. Let me say that again: handling cash is NOT free.
Very true.
Many years ago I designed a cash management system for a large grocery store chain. "Designed" because we didn't get the contract to build it (I don't recall why). It included counting machines that shrink-wrapped and barcoded stacks of cash, barcode readers at counting stations, safes, and in armored cars as well as at point of sale stations (where they already existed), a centralized cash-tracking system that knew where every bundle was and was integrated with the employee management system so we could identify every employee that touched a given bundle or group of bundles, or opened a safe, etc., and with the point of sale system so it could automatically check expected drawer contents against automated count results. It was also integrated with a network of CCTV cameras that oversaw cash-handling areas, so the handling of any bundle could be looked up and with the banks' cash ordering/pickup systems, to automate delivery of new cash and pickup of old cash.
The total cost of the system for a thousand stores was going to be on the order of $150M, plus another $10M or so in annual operational and maintenance costs... and it was expected to pay for itself easily. In two years, IIRC. The combined cost of all the man-hours that would be saved on counting and handling cash, and all of the savings from reducing "shrinkage" was huge.
That all sounds very reasonable. What are the odds that they walked up to the guy and explained it like you wrote ?
Impossible to say. There are a thousand ways it could have gone down, including the passenger beginning to foam at the mouth as soon as they said "Sir, we're going to need you to deplane", and the crewmember saying nothing but sending a cop to deliver a terse version of the message, backed up by force. I'm sure the actual event was somewhere in between.
But were you forcibly removed, hit, had your face bloodied? Were you a doctor who had patients to see? Was the flight not 30 minute later, but instead the next day?
No, because I wasn't foolish enough to refuse the orders of the crew, which carry the weight of federal law.
I hope all these COBOL programmers are smart enough to charge outrageous rates. Minimum $250/hr.
They've got the banks over a barrel, any time the situation is reversed the banks don't hesitate to screw us.
$250 per hour is a pretty normal rate for an experienced professional, hardly outrageous at all. When I worked for IBM Global Services, they billed me out at $300 per hour. Granted that IBM commands a premium due to their marketing channels, but I'm sure I could have gotten $200 per hour on my own and my expertise is far from as rare as deep COBOL experience. I'd expect people like the one mentioned in the article to cost more like $500 per hour, if they're actually aware of their own worth, and I wouldn't consider even that outrageous.
If I were in the position of those programmers, I would probably try to avoid quoting an hourly rate at all. Instead I'd do it all as piece-rate work, based on detailed specifications -- especially if I'm working on a system that I know, so I have a good idea of what sorts of obstacles I might run into. Then I'd try to set my price based on the value of the work to the business, rather than on the time it would take me to produce it. That could easily result in contracts that work out to many thousands of dollars per hour.
A friend who is a lawyer told me that the best piece of career advice he ever got from his father, also a lawyer, is "Take the money, son." He didn't mean to take money for unethical work, but just that one shouldn't balk at accepting high fees just because they seem too high. If the customer is willing to pay, take the money. That attitude should absolutely be applied to doing contract work for banks.
How the heck is it even possible to get involuntarily bumped?
There are lots of possibilities. In my case, they had screwed up and assigned two passengers the same seat. Another thing that can happen is when they bump a "normal" passenger in favor of a high-status passenger. I was diamond medallion on Delta for a while, and one of the perks that comes with that status is that the airline guarantees that they will sell you a seat on any flight, even if it's already excessively overbooked, and (implicitly) even if that means they have to get someone else off. This usually works out fine, because the high-status last-minute purchaser is paying full fare for the ticket and there are plenty of people who are willing to be bought off the plane for significantly less than that. But in rare cases complying with their commitment to the high-status passenger could require them to involuntarily bump someone.
In the case currently getting so much press, there was apparently a crew for another flight that immediately needed to get to the destination to crew another aircraft, due to some other snafu, so the airline decided to take some passengers off to make room for them, rather than canceling or further delaying another flight. Apparently they were unable to buy anyone off -- or possibly with all the passengers on board there was just no time to go through that process -- so they picked some passengers at random. The only other option I can think of would have been to get on the PA and announce an increasing sequence of cash payments until they got their four volunteers. From a financial perspective there's no point in offering more than the FAA's regulations define for involuntarily bumped passengers, though, so if they were already sure they wouldn't get a volunteer without going above that level (which they could know if they'd already had to offer high paymetns to buy people off even before the emergency need for four more seats cropped up), then from a strictly financial point of view it was best just to pick four.
Given the national PR that ensued after a bumped passenger refused to debark and ended up having to be hauled off by the police, that was a bad decision, but that's hindsight. What usually happens is what happened in my case, nothing much.
no, for an individual consumer, the best strategy is to, whenever possible, ask for a discount at least equivalent to the fees charged by the card to pay in cash, this way you recoup all the fees and not only the peanuts the bank throws back at you.
Yeah, walk up to the cashier in your local grocery store and ask for a cash discount. Good luck with that.
In places where you're talking to the owner, or at least a very empowered (and smart) manager or employee, you might be able to get them to knock 3-5% off for a cash payment, but it's rare. Everywhere else... get the best rewards card you can, and use it as much as possible.
I actually had the "cash discount" discussion last month when I bought a car (actually bought out my leased car). The dealership's policy is to accept plastic for purchases up $4K, so I offered to pay $4K with the card and the remaining thousand or so with cash. Once they said okay, I offered to pay cash if they knocked an additional $150 off the price, since that is about what they were going to pay in card fees. They refused. They also refused to knock $100 off. So, I paid with the card and got my 2% (a little under $80) discount that way. Unfortunately I noticed later that there was a promotion on my Discover card that I could have used to get 5% back, so I could have gotten $200 back (and they'd have undoubtedly been soaked for a bit more than that).
You're usually pretty safe from being bumped once you're actually on the plane though.
"Pretty safe", yes, but you've never been completely safe. That said, the FAA has some well-defined requirements about how they have to treat people who've been involuntarily bumped, which includes a hefty cash payment (equal, I believe, to the full round-trip fare) plus a seat on the next available flight (on any airline, in any class at or above the class you paid for). I've flown over a million miles, and I've been involuntarily bumped exactly once. I got a $600 check and a first class seat on another flight, on another airline, 30 minutes later. The seat they bought me was on a direct flight, so I actually got home before I would have if I hadn't been bumped. Oh, and they still gave me mileage credit for the flight they bumped me off of. All in all, I was quite happy with the arrangement.
Silly people, imagining that once they've paid for something they have any kind of rights.
You do have rights, but they don't include the right to refuse to exit the plane when the airline tells you to. Whatever the reason, whether it's a good one or not, if the flight attendants or captain tell you to get off, you get off or the police will be dragging you off. If they kicked you off involuntarily, and not as a result of anything you did, you do have a right to compensation and transportation.
BTW, if you don't want to watch the whole video (though I highly recommend it; Rosling does a great job of making dry statistics interesting), you can start at 10:15 for the most relevant part.
So global statistics show population growth is slowing even though we're still growing population faster than ever because population won't really stop growing until it stops growing?
No. Global statistics show that the annual total number of births has been steadily declining since the late 80s. Population is growing only because the global population skews young, and we're in the process of filling out the age groups. Hans Rosling explained it very well in this TED talk: https://www.youtube.com/watch?...
Or, instead of naively using a fixed percentage, they could develop a sophisticated algorithm which identifies the probability of a deviation and its likely magnitude. You know, like what they actually did.
We don't know what they did, but simply choosing a longer route does not identify the probability of a deviation. Depending on how the select the longer route, it may or may not identify likely magnitude.
In companies like Google, you have to ask for raises and promotions.
Promotions yes, raises no (except for the raises that come with promotions). At Google you have to apply for promotion. Raises are just allocated annually by management.
Global data pretty clearly demonstrates that we've reached a point where increasing wealth (and especially female education) slows population growth. Indeed, we've already passed the point of peak childbirth; fewer children are born every year and that trend has been continuing for the last 30 years. The only reason population is currently growing -- as wealth continues to increase -- is because we're filling out the age brackets.
All evidence shows that the wealthier we get, the fewer children we have.
Only some evidence shows that.
Until about 50 years ago, wealth meant more children.
So what changed?
1. Contraceptives
2. Rapidly falling infant and childhood mortality
3. Increasing urbanization
And female education. That seems also to play a very significant role.
"it's much better to simply not show ads on those channels"
This is what they are doing.
Yes, that's what I said. You said "It also cuts down the overhead of accounting for the money streams on small channels", and I said (at perhaps excessive length) that that doesn't make sense.
"I don't see how it's more difficult to account for money streams on small channels anyway"
I will thank you to note that I didn't say that it was.
My point was that the overhead is probably fixed but the revenue stream is not, so there need to be a minimum view rate for YouTube to cover their own costs. Less than that, it's cheaper just to show the video with no ad at all.
Meh. And I say that as a Google engineer who used to work on the accounting and payment-distribution systems, though I think I'd come to the same conclusion even without inside knowledge. You have to store a bunch of data about each account anyway, and you have to have regular map/reduce jobs over accounts to gather data and send it to the payment systems. It would be more work to distinguish channels based on viewer volume and decide whether to include them.
Come to think of it, part of the reason for this decision may have been analysis of engineering effort, precisely because you don't have to bother with those distinctions in the payment process. The ad-choosing software has to make a bunch of channel-specific (and user-specific) decisions anyway, so it's likely trivial to ad "if (total_views < 10000) return NO_AD;". That done, the accounting and payment systems can continue to operate completely unaffected; they'll still scan over all YouTube channels looking for ad views to revenue-share, but the little channels just won't have any. Easy peasy.
Google engineers typically don't think too much about the server-side cost of scanning a few million (or billion) extra records. Given the infrastructure, the cost of doing so is trivial. It would be easy to insert a little logic in the scanning process to check view counts, assuming that data is easily available at that point, but you've already done all of the I/O at that point, so it would gain nothing. To actually gain, you'd have to somehow identify beforehand which records to avoid scanning. I suppose you could create an additional table, populated by a pre-processing map/reduce, that tells you which ones to scan, but that (a) still wouldn't save any overall I/O or CPU and (b) could be error-prone in various ways. You could also update that table as views occur, but that would add latency and complexity to the up-front processes.
No... from both a user fairness perspective and an engineering effort perspective, it's much better to simply not show ads on those channels. Plus all of the other reasons I mentioned in my previous comment.
Slashdot Mirror
I was diamond medallion on Delta for a while, and one of the perks that comes with that status is that the airline guarantees that they will sell you a seat on any flight, even if it's already excessively overbooked, and (implicitly) even if that means they have to get someone else off. This usually works out fine, because the high-status last-minute purchaser is paying full fare for the ticket and there are plenty of people who are willing to be bought off the plane for significantly less than that. But in rare cases complying with their commitment to the high-status passenger could require them to involuntarily bump someone.
This is exactly the kind of problem that exists in the system: that some passengers are more equal than others. Thus the outrage. The "random" selection of passengers who were asked to deplane presumably did not include any of those more equal people.
From the airline's point of view, some passengers *are* more equal than others. If you get in a situation where you have to pick someone to favor, it's much more important to serve a businessman who spends (well, whose employer spends) $100K per year on flights (many of them full-fare tickets) than a person who takes one deeply-discounted $200 bargain flight every three years. If you lose the latter's business, you haven't lost much. For that matter, the person with the deeply-discounted fare ranks below the person who isn't a frequent flier but bought a more expensive ticket. On many routes, airlines make basically all of their money from the repeat business travelers and they see the other travelers as just a way to avoid flying with unpaid seats. That's why those deep discount fares are often so ridiculously cheap.
I'd guess that if the random selection excluded high-status frequent flyers, it also excluded those who'd paid full fare. In practice, though, I doubt they excluded anyone from the selection pool. Heck, if I were in their shoes and had to pick, I'd pick exclusively people from the second-highest frequent flyer class. I'd want to pick experienced, regular travelers because they'll react much more calmly. That would make the highest frequent flyer class the ideal choices, except that they have that seat guarantee.
no stop signs & stoplights = underpass / overpasses for pedestrians.
Or walk signals for pedestrians, which tell the vehicles to stop. But, yeah, when intersections have free-flowing traffic except for pedestrians there will be a lot of incentive to build pedestrian overpasses or underpasses.
Google offers several months' paid leave
Actually, that should be "a few months' paid leave". "Several" to me implies 5-7, and it's not that much. I haven't looked at the details (I'm long past having kids so it's not relevant to me personally), but I think it's on the order of three months.
"extremely scientific and robust," relying on the employee's role, job level and location, as well as recent performance ratings"
So basically they are claiming that performance ratings are scientific, and that there's no possibility those are biased.
No, they said pay calculation was scientific, not performance ratings. Performance ratings are an input to the pay calculation, and one that inevitably has a subjective element, though Google does a lot to try to identify and eliminate bias -- vastly more than any other company I've worked with/for.
Women who do not have children get paid the same or more. But when you have to take several weeks or months off to take care of a child you slow your career.
Google offers several months' paid leave to both mothers and fathers, and all are strongly encouraged to take it. In part this is due to competitive pressure -- all the big tech companies are fighting over the same pool of employees and it's a really awesome perk, but it's also quite explicitly an attempt to eliminate this aspect of gender preference. Women who give birth do get a little extra time for "medical leave", but the actual maternity/paternity leave is the same. Oh, and it applies to employees who adopt, too (not the medical part, obviously).
If you want to have the biological and emotional fulfillment of giving birth and raising a child then realize you have to sacrifice your overall income.
Not at Google. Oh, I suppose it may slow you a little because the time you're on leave is time you're not doing promotion-worthy things, but in practice it doesn't seem to have any significant effect. I see people with families of all sizes at all levels of the career ladder. That even includes a few quite senior people who take maternity/paternity leave every year. I know one engineer who has had a child every year for six years, and taken all of the leave, and leads a large and important team. I know another who has negotiated a deal with management to accept a 60% salary in exchange for working only three days per week, and also to spread maternity/paternity leave over time, taking one day of it per week, with the net effect of a two day per week work schedule -- and just got a major promotion. That particular engineer is something of a rock star and I'm not sure that sort of deal is generally available (though fractional salary for reduced work schedule is, with management approval). Note that I didn't specify the gender of either of those examples. One is a man and one is a woman; their gender doesn't affect the options available to them.
As a Google employee, my reaction to the DOL claim was "WTF"? The claim is so utterly at odds with the way Google operates.
Here's my guess as to how the DOL came to their conclusion: They just looked at average male and average female compensation, without considering job role. Because women are underrepresented in engineering, and engineering jobs are better-compensated than most other categories, the average female compensation is probably lower. That women are underrepresented in engineering is something Google regularly and publicly discusses, and the company has a wide variety of initiatives aimed at improving that situation, mostly by trying to increase the number of women in the hiring pipeline.
nothing about city traffic will fundamentally change.
It'll change dramatically.
For one, it will largely eliminate parking. No need to park your car on the street in front of your home, have it park somewhere outside the city and have it drive in when you need it. When you go to the store or similar, no need to park your car at the store. It can circle the block, or drive a few miles away to where it can park. This will have the biggest effects not in dense urban areas but in moderately-dense suburban areas. Those huge mall parking lots will become unnecessary. Assuming car ownership stays high, parking will likely consolidate into enormous lots a few minutes' drive from merchants. It's also highly likely that many people living in moderately-dense areas that currently have a car will switch to using self-driving car services instead, since they'll be cheaper and almost as convenient, and they won't need parking at all since when they get out of the vehicle it will go off to service another customer.
In the longer run, cities will likely outlaw manually-operated vehicles and that will have an even more dramatic effect. Mutually-cooperating or centrally-controlled (could go either way) vehicles that communicate via RF have no need of things like stop signs and stoplights. Taken to the logical limit, there's no need for vehicles to stop at intersections at all as long as they can coordinate to arrange for gaps for cross-traffic to pass through -- and the gaps needn't be any larger than the time the crossing vehicle needs to pass through them (which will be terrifying until we get used to it). This will make city traffic smooth and predictable... which may in turn cause it to increase in volume until it's not.
On highways, such vehicles have no need of the "safe following distances" required by human situational non-awareness and slow response times. They'll close up into trains of vehicles inches apart, both to pack more vehicles into a given area of roadway and improve energy-efficiency by drafting. They will also be able to increase speeds significantly without danger, and without great loss of efficiency (due to drafting). I expect that it won't be too long after self-driving vehicles become commonplace that local governments will carve out self-driving only lanes so they can begin to take advantage of the higher efficiencies offered.
People also carpool. That's been around forever.
Sure, but people don't carpool very much, because unless you have a fixed group of people who go the same place every day at the same time, it's too hard to organize. It's often too hard to organize even with such a group of people. Self-driving cars with centralized dispatching systems can handle the organization automatically and dynamically. It seems entirely likely that in the future you'll put what time you want to go to work in your phone and your digital assistant will automatically notify your carpool service of choice. The service will take your requirements along with the requirements of thousands of others and construct routes that incur minimal overhead, and your phone will notify you when your ride is five minutes out. Note that it's not necessary that the vehicle be self-driving for this to work, but it will be a natural extension for self-driving car services, which will help mitigate the fact that ride demand is so bursty. They can purchase small vans which transport a half-dozen people during rush hour and make grocery store runs, etc., at other times. Demand-adjusted pricing will encourage people to take single-rider journeys at off-peak times, and to accept sharing at peak times.
$250 per hour is a pretty normal rate for an experienced professional, hardly outrageous at all. When I worked for IBM Global Services, they billed me out at $300 per hour. Granted that IBM commands a premium due to their marketing channels,
Its not the marketing channels. Its the organization behind you - technical support and resources so that even when you don't personally have the answers you have a direct line into the organization to get answers.
Yeah... no. None of that ever applied to anything I was doing.
The problem is, the technology isn't quite there... and the infrastructure is nowhere near there.
For most people not living in apartments, infrastructure is irrelevant as long as the car has enough range to get them to and from work, plus a bit of running around. If you can charge at home, infrastructure only matters when you go on long trips.
So an EV is fine for someone who pootles about town and keeps the car plugged in for 14 hours a day at home, but for the motorist doing 8000 miles a year
8000 miles per year? That is "pootling about town". I put over 24,000 miles per year on my LEAF (now that it's not artificially constrained by lease limitations). I almost never charge anywhere except home, because there's almost no place to charge where I live. I do have a level 2 charging station in my garage, so a charge from empty takes a little over three hours.
So I doubt the majority of cars on roads in the next decade will be fully electric.
Well, since the majority of cars on the roads ten years from now will be like those that are selling now, that's a very safe bet. On the other hand, I think that it's entirely reasonable to believe that the majority of sales ten years from now will be EVs. Barring regulatory forcing (which could happen), it'll take another decade for them to become the majority of cars on road.
I'm not saying Tesla will fail, but I am saying its over-valued and that will be corrected in time.
That, however, seems entirely likely. I think the Model 3 will be quite popular and that Tesla will do very well. That doesn't mean that Tesla will own the market in a decade, though they'll probably be a significant player. Basically, their current valuation is probably about what their valuation will be in ten years.
Interesting. Thanks!
In places where you're talking to the owner, or at least a very empowered (and smart) manager or employee, you might be able to get them to knock 3-5% off for a cash payment, but it's rare.
It's also important to note that handling cash is not free. Let me say that again: handling cash is NOT free.
Very true.
Many years ago I designed a cash management system for a large grocery store chain. "Designed" because we didn't get the contract to build it (I don't recall why). It included counting machines that shrink-wrapped and barcoded stacks of cash, barcode readers at counting stations, safes, and in armored cars as well as at point of sale stations (where they already existed), a centralized cash-tracking system that knew where every bundle was and was integrated with the employee management system so we could identify every employee that touched a given bundle or group of bundles, or opened a safe, etc., and with the point of sale system so it could automatically check expected drawer contents against automated count results. It was also integrated with a network of CCTV cameras that oversaw cash-handling areas, so the handling of any bundle could be looked up and with the banks' cash ordering/pickup systems, to automate delivery of new cash and pickup of old cash.
The total cost of the system for a thousand stores was going to be on the order of $150M, plus another $10M or so in annual operational and maintenance costs... and it was expected to pay for itself easily. In two years, IIRC. The combined cost of all the man-hours that would be saved on counting and handling cash, and all of the savings from reducing "shrinkage" was huge.
That all sounds very reasonable. What are the odds that they walked up to the guy and explained it like you wrote ?
Impossible to say. There are a thousand ways it could have gone down, including the passenger beginning to foam at the mouth as soon as they said "Sir, we're going to need you to deplane", and the crewmember saying nothing but sending a cop to deliver a terse version of the message, backed up by force. I'm sure the actual event was somewhere in between.
But were you forcibly removed, hit, had your face bloodied? Were you a doctor who had patients to see? Was the flight not 30 minute later, but instead the next day?
No, because I wasn't foolish enough to refuse the orders of the crew, which carry the weight of federal law.
I hope all these COBOL programmers are smart enough to charge outrageous rates. Minimum $250/hr. They've got the banks over a barrel, any time the situation is reversed the banks don't hesitate to screw us.
$250 per hour is a pretty normal rate for an experienced professional, hardly outrageous at all. When I worked for IBM Global Services, they billed me out at $300 per hour. Granted that IBM commands a premium due to their marketing channels, but I'm sure I could have gotten $200 per hour on my own and my expertise is far from as rare as deep COBOL experience. I'd expect people like the one mentioned in the article to cost more like $500 per hour, if they're actually aware of their own worth, and I wouldn't consider even that outrageous.
If I were in the position of those programmers, I would probably try to avoid quoting an hourly rate at all. Instead I'd do it all as piece-rate work, based on detailed specifications -- especially if I'm working on a system that I know, so I have a good idea of what sorts of obstacles I might run into. Then I'd try to set my price based on the value of the work to the business, rather than on the time it would take me to produce it. That could easily result in contracts that work out to many thousands of dollars per hour.
A friend who is a lawyer told me that the best piece of career advice he ever got from his father, also a lawyer, is "Take the money, son." He didn't mean to take money for unethical work, but just that one shouldn't balk at accepting high fees just because they seem too high. If the customer is willing to pay, take the money. That attitude should absolutely be applied to doing contract work for banks.
How the heck is it even possible to get involuntarily bumped?
There are lots of possibilities. In my case, they had screwed up and assigned two passengers the same seat. Another thing that can happen is when they bump a "normal" passenger in favor of a high-status passenger. I was diamond medallion on Delta for a while, and one of the perks that comes with that status is that the airline guarantees that they will sell you a seat on any flight, even if it's already excessively overbooked, and (implicitly) even if that means they have to get someone else off. This usually works out fine, because the high-status last-minute purchaser is paying full fare for the ticket and there are plenty of people who are willing to be bought off the plane for significantly less than that. But in rare cases complying with their commitment to the high-status passenger could require them to involuntarily bump someone.
In the case currently getting so much press, there was apparently a crew for another flight that immediately needed to get to the destination to crew another aircraft, due to some other snafu, so the airline decided to take some passengers off to make room for them, rather than canceling or further delaying another flight. Apparently they were unable to buy anyone off -- or possibly with all the passengers on board there was just no time to go through that process -- so they picked some passengers at random. The only other option I can think of would have been to get on the PA and announce an increasing sequence of cash payments until they got their four volunteers. From a financial perspective there's no point in offering more than the FAA's regulations define for involuntarily bumped passengers, though, so if they were already sure they wouldn't get a volunteer without going above that level (which they could know if they'd already had to offer high paymetns to buy people off even before the emergency need for four more seats cropped up), then from a strictly financial point of view it was best just to pick four.
Given the national PR that ensued after a bumped passenger refused to debark and ended up having to be hauled off by the police, that was a bad decision, but that's hindsight. What usually happens is what happened in my case, nothing much.
no, for an individual consumer, the best strategy is to, whenever possible, ask for a discount at least equivalent to the fees charged by the card to pay in cash, this way you recoup all the fees and not only the peanuts the bank throws back at you.
Yeah, walk up to the cashier in your local grocery store and ask for a cash discount. Good luck with that.
In places where you're talking to the owner, or at least a very empowered (and smart) manager or employee, you might be able to get them to knock 3-5% off for a cash payment, but it's rare. Everywhere else... get the best rewards card you can, and use it as much as possible.
I actually had the "cash discount" discussion last month when I bought a car (actually bought out my leased car). The dealership's policy is to accept plastic for purchases up $4K, so I offered to pay $4K with the card and the remaining thousand or so with cash. Once they said okay, I offered to pay cash if they knocked an additional $150 off the price, since that is about what they were going to pay in card fees. They refused. They also refused to knock $100 off. So, I paid with the card and got my 2% (a little under $80) discount that way. Unfortunately I noticed later that there was a promotion on my Discover card that I could have used to get 5% back, so I could have gotten $200 back (and they'd have undoubtedly been soaked for a bit more than that).
You're usually pretty safe from being bumped once you're actually on the plane though.
"Pretty safe", yes, but you've never been completely safe. That said, the FAA has some well-defined requirements about how they have to treat people who've been involuntarily bumped, which includes a hefty cash payment (equal, I believe, to the full round-trip fare) plus a seat on the next available flight (on any airline, in any class at or above the class you paid for). I've flown over a million miles, and I've been involuntarily bumped exactly once. I got a $600 check and a first class seat on another flight, on another airline, 30 minutes later. The seat they bought me was on a direct flight, so I actually got home before I would have if I hadn't been bumped. Oh, and they still gave me mileage credit for the flight they bumped me off of. All in all, I was quite happy with the arrangement.
Silly people, imagining that once they've paid for something they have any kind of rights.
You do have rights, but they don't include the right to refuse to exit the plane when the airline tells you to. Whatever the reason, whether it's a good one or not, if the flight attendants or captain tell you to get off, you get off or the police will be dragging you off. If they kicked you off involuntarily, and not as a result of anything you did, you do have a right to compensation and transportation.
BTW, if you don't want to watch the whole video (though I highly recommend it; Rosling does a great job of making dry statistics interesting), you can start at 10:15 for the most relevant part.
So global statistics show population growth is slowing even though we're still growing population faster than ever because population won't really stop growing until it stops growing?
No. Global statistics show that the annual total number of births has been steadily declining since the late 80s. Population is growing only because the global population skews young, and we're in the process of filling out the age groups. Hans Rosling explained it very well in this TED talk: https://www.youtube.com/watch?...
Or, instead of naively using a fixed percentage, they could develop a sophisticated algorithm which identifies the probability of a deviation and its likely magnitude. You know, like what they actually did.
We don't know what they did, but simply choosing a longer route does not identify the probability of a deviation. Depending on how the select the longer route, it may or may not identify likely magnitude.
In companies like Google, you have to ask for raises and promotions.
Promotions yes, raises no (except for the raises that come with promotions). At Google you have to apply for promotion. Raises are just allocated annually by management.
Global data pretty clearly demonstrates that we've reached a point where increasing wealth (and especially female education) slows population growth. Indeed, we've already passed the point of peak childbirth; fewer children are born every year and that trend has been continuing for the last 30 years. The only reason population is currently growing -- as wealth continues to increase -- is because we're filling out the age brackets.
All evidence shows that the wealthier we get, the fewer children we have.
Only some evidence shows that. Until about 50 years ago, wealth meant more children. So what changed? 1. Contraceptives 2. Rapidly falling infant and childhood mortality 3. Increasing urbanization
And female education. That seems also to play a very significant role.
"it's much better to simply not show ads on those channels"
This is what they are doing.
Yes, that's what I said. You said "It also cuts down the overhead of accounting for the money streams on small channels", and I said (at perhaps excessive length) that that doesn't make sense.
"I don't see how it's more difficult to account for money streams on small channels anyway"
I will thank you to note that I didn't say that it was.
My point was that the overhead is probably fixed but the revenue stream is not, so there need to be a minimum view rate for YouTube to cover their own costs. Less than that, it's cheaper just to show the video with no ad at all.
Meh. And I say that as a Google engineer who used to work on the accounting and payment-distribution systems, though I think I'd come to the same conclusion even without inside knowledge. You have to store a bunch of data about each account anyway, and you have to have regular map/reduce jobs over accounts to gather data and send it to the payment systems. It would be more work to distinguish channels based on viewer volume and decide whether to include them.
Come to think of it, part of the reason for this decision may have been analysis of engineering effort, precisely because you don't have to bother with those distinctions in the payment process. The ad-choosing software has to make a bunch of channel-specific (and user-specific) decisions anyway, so it's likely trivial to ad "if (total_views < 10000) return NO_AD;". That done, the accounting and payment systems can continue to operate completely unaffected; they'll still scan over all YouTube channels looking for ad views to revenue-share, but the little channels just won't have any. Easy peasy.
Google engineers typically don't think too much about the server-side cost of scanning a few million (or billion) extra records. Given the infrastructure, the cost of doing so is trivial. It would be easy to insert a little logic in the scanning process to check view counts, assuming that data is easily available at that point, but you've already done all of the I/O at that point, so it would gain nothing. To actually gain, you'd have to somehow identify beforehand which records to avoid scanning. I suppose you could create an additional table, populated by a pre-processing map/reduce, that tells you which ones to scan, but that (a) still wouldn't save any overall I/O or CPU and (b) could be error-prone in various ways. You could also update that table as views occur, but that would add latency and complexity to the up-front processes.
No... from both a user fairness perspective and an engineering effort perspective, it's much better to simply not show ads on those channels. Plus all of the other reasons I mentioned in my previous comment.
efficiency improvements would normally lead to sharp population expansion
Huh? Humans are not rabbits. All evidence shows that the wealthier we get, the fewer children we have.