That's fine for you, but if you're building phones for a large base of consumers it's better to serve their needs, even those that you think are silly.
Sorry, the "none of them" bit should have been quoted. That was from the post I replied to, not from me. I included the link to the court ruling allowing the emoluments suit to go forward.
True. And, they may be right!! See also, "I want to invest in the housing market in 2008". If enough people do a dumb thing, the government sometimes makes it not dumb retroactively.
Even without government involvement, if, for example, every single job applicant for a position you're trying to fill has posted videos of themselves doing crazy, stupid or illegal things on Facebook, then you have to ignore that, or hire no one. There's a notion of "radical transparency" that says that if we can see all the details of everyone's lives, we'll realize that everyone is a screwup at least some of the time and stop paying attention to those screwups.
That said, I don't post my own stupid mistakes on the web, and I discourage my kids from doing it either.
I see a lot of selfies floating around that clearly aren't about the background. Many, taken in bathrooms and whatnot, would definitely be better with the background de-emphasized.
There are three. Two of them apply to the president, one barring him from receiving foreign emoluments, one barring him from receiving anything other than his salary from the federal government or the states.
Regardless, Trump is in violation of none of them.
If you have an iPhone it already does this with two cameras working in tandem on the back.
Pixel 2 and Pixel 3 do this with a single camera The sensors have "split pixels", so when you take one photo with one camera you're actually getting two photos, with a tiny separation. This separation is enough, it turns out, to compute a depth map. Not good enough for actual 3D photography, but more than adequate for doing what most two-camera phones do: make portraits with a blurred background. And since two cameras aren't required, it works with the front camera, too. Bokeh for your selfies.
because they won't release those stats. They use contractors to hide it. The published stats are for FTEs. A huge part of the reason to use contractor firms for that is to hide those stats.
Anecdotally the last 4 places I've worked at are pushing 80%.
From what I can see at Google there are almost no contractors doing engineering work. I'd be surprised if it was even 1%.
Bitcoin uses SHA-256(SHA-256(nonce+data)) < target value. They don't actually look for collisions.
And if they did look for collisions, at the peak hash rate thus far achieved (about 2^69 hashes per second) it would still take about 2^161 years to reach a 50% probability of finding one. This would make BTC transactions a bit slower than they are now.
Google allows users to export all of their data. Hasn't seemed to destroy Google.
A) Google definitely does NOT permit exporting all their data or even accessing all of it. Some yes but definitely not all.
Cite? Looks like all of it to me. They can't include anything they infer is *probably* about you, such as browsing history while not logged in, because if it's not actually yours giving it to you would violate the privacy of whoever it was.
D) Meta-data about group tendencies are as valuable as personal information and they don't (and won't) give you access to that.
Woz also suggested that Facebook should allow users to export their data so they could upload it onto competing social networks.
You know, Woz is seemingly a very decent but this is just almost weapons grade stupid. How about Woz tell Apple to drop all their patents and open source their software? Because that's the functional equivalent of what he is suggesting. He's telling Facebook to hand over the crown jewels of their empire which is data about their customers.
Google allows users to export all of their data. Hasn't seemed to destroy Google. I'll grant that Google has assets other than user data, but user data is still pretty big.
The essential fiduciary duty of a publicly traded company is to maximize profits for its shareholders without regard to any moral considerations.
This is untrue.
The essential duty of a publicly-traded company is to work towards the goals outlined in its articles of incorporation and IPO letters, and perhaps the will of the shareholders if votes are held to alter those documented goals (not common). The goals nearly always include generating profits, but that's often not the only goal, and sometimes it isn't even the primary goal. It is always a goal, because all publicly-traded corporations are for-profit (US law bars non-profit corporations from selling shares).
And even for corporations that do have profit generation as their primary, or only, goal, it's still not true that directors and executives will be held legally accountable for failing to maximize profit. In theory that's possible, in practice it only happens with the most egregious of mismanagement.
There are characteristics of large organizations that tend to dilute moral concern and enable otherwise reasonably-moral individuals to make extremely amoral decisions. But it has nothing to do with fiduciary duty. Please, let's kill this tired old trope.
With precise control of the cars, it should be feasible to get inter-car gaps down to only a few feet, or even to zero, locking them together into an ad-hoc, very temporary train. If we assume only one car per second, that's 480-960 people per minute, in the same ballpark as a subway. At 4.4 cars per second it could carry 2100 to 4200 people per minute.
Ah, but this is science fiction. Currently we don't know how to do this in a safe manner.
Actually, quite a lot of research has already gone into doing this for highway vehicles. One company (Peloton Technology) even has a system on the market for "platooning" pairs of semi trucks. NIST and US DOT both have active research programs, as do several organizations in the EU. Demonstrations of safe and effective platooning have been done with inter-vehicle distances as small as 0.6 seconds at speeds up to 70 mph, including automatic handling of unexpected merges from vehicles not part of the system. Nearly all of the safety issues in such systems arise from the presence of human-operated vehicles on the road. Those wouldn't be present in Musk's system, allowing the tolerances to be closed considerably, but even with a very conservative 0.6 second gap, that's still 72 cars per minute.
Musk himself only projects 30 second headways , a far cry from 1 second or 0.23 second headways like you suggest.
That was specific to the Chicago project, where only ~2000 passengers per hour were thought to be needed. The article said nothing about "30 second headways", it said a car would leave every 30 seconds... that's technically the same thing, but the focus is on passenger departure rates, not safe vehicle packing rates. There's no indication that anyone thinks those 30-second gaps are necessary for safety. Indeed that would be ridiculous, equivalent to maintaining half-mile separations between cars on a highway. If an uncontrolled environment like a highway doesn't need 30-second gaps, a controlled tunnel certainly does not.
There is precise definition of "profit" in accounting, and Tesla is not doing it.
The precise definition is embodied in GAAP. Tesla turned a >20% GAAP profit in the third quarter. It remains to be seen if Tesla can sustain that, but teardowns show Tesla's vehicles to all have healthy per-unit profit margins and they seem to have turned the corner on their production issues, so it's reasonable.
What made SpaceX's first test launch worthy of praise? The first Falcon 1 flew for 41 seconds and landed (impacted is more accurate) 250 feet from the launch site.
You've got to start somewhere. Musk is excited about getting started, that's all. Clearly there's nothing particularly special about this tunnel, other than it's the beginning of what he has planned. Yes, that means any praiseworthiness is based entirely on the assumption that more is to come. But that seems like a reasonable expectation to me.
Musk proposes that each vehicle carry only 8 to 16 passengers. A full subway train, in contrast, carries over 1000 passengers. Musk plans for a vehicle every 30 seconds, compared to every 90 seconds for a modern subway line. So Musk's system will be able to carry 16-32 people per minute, compared to a subway which carries around 700 people per minute.
You're assuming Musk's system uses the same layout as subway tunnels, which it wouldn't.
In Musk's system, cars would not generally accelerate or decelerate in a transit tunnel (there may be a tiny bit as the control system accelerates or decelerates cars a bit to make room for a merging car), so you could have an almost continuous stream of cars going by. Cars would enter or leave a transit tunnel through an entrance or exit ramp tunnel. Think of a divided highway with entrance and exit ramps and overpasses... but with all vehicles centrally-controlled.
If we assume a 60 mph tunnel speed (88 fps), and individual cars that are 20 feet in length (that's generous), then we have 4.4 car lengths passing a fixed point each second. With precise control of the cars, it should be feasible to get inter-car gaps down to only a few feet, or even to zero, locking them together into an ad-hoc, very temporary train. If we assume only one car per second, that's 480-960 people per minute, in the same ballpark as a subway. At 4.4 cars per second it could carry 2100 to 4200 people per minute.
For that matter, if you reach the maximum throughput of a single transit tunnel, based on the minimum inter-car gaps and maximum speed dictated by safety, turn sharpness, etc., you can always do what highways do to increase capacity: add another lane by digging a parallel tunnel.
The other limiting point for throughput is stations. Stations would have to be sized appropriately. Suppose a subway station has a 1000-passenger train every 90 seconds, and that each train loads/unloads at most 100 people. With Musk's system the station would therefore have to load/unload, say, 12 cars every 90 seconds, or one every 7.5 seconds. Assuming it takes people 30 seconds to get on or off (that's generous, I think, except for disabled people), that means you need to have four cars unloading and four cars loading at once, so you'd need a loading/unloading bay of perhaps 10 car lengths, or maybe two bays of five car lengths. Or maybe you'd need a bit more because there would also be a sorting process going on; people going to the same destination would be assigned to the same car. I'm imagining that people would be informed by their phones which bay they should to go and which car they should get on. So you might want to have more bays to allow same-destination groups to "accumulate".
Actually, for routing of foot traffic, it might make more sense to separate loading and unloading entirely, since people exiting a car would be at their destination and exiting the system (transfers don't make sense in Musk's system). Street entrances/exits should probably be separated so all foot traffic is always moving the same direction. You might end up with station designs where each station is split into a pair of disjoint stations, one for people entering and another for people leaving.
And, of course, with a subway the 900 people who don't get on or off at the station still have to stop and wait for those who do. In Musk's system they'd just whiz right past the station.
Secondly, I don't know about USA regulations, but in the UK the railway construction regulations do not permit more than a cetain modest amount of cant; AFAIR is is about 6 degrees. The reason is to avoid standing passengers falling over or merely being discomforted if the train has to stop at those places for signals or any other reason. You might think that rule is too cautious, but that is how it is and I have no doubt there are similar regulations concerning roads, although not for fairground rides.
If you get rid of the rails and use wheeled skates, then you don't have to have a fixed cant at all. Just have the skates ride as far up the side as necessary at the given speed so the passengers don't feel any side acceleration. They'll experience an increase in "downward" acceleration during turns, of course, so it will make sense to set an upper limit on allowable gee force.
Major OS updates? Yes. That's acceptable. But it'll also receive security updates for much much longer. That's standard for all but no-name-company phones.
I wish it were standard. It's not. Unfortunately, it's much better than the standard, which is no commitment to updates at all.
By it's very definition a balance in independent of gravity. The balance will remain the same and that 1Kg *of mass* will work just as well if you try this experiment on Jupiter.
But a Kibble Balance isn't that sort of a balance. It's a "single pan balance" which balances gravitational acceleration against acceleration caused by a magnetic field. So the Kibble Balance is very sensitive to changes in the gravitational field.
Luckily, it's possible to measure the local force of gravity with extreme precision, without reliance on the definition of the kilogram. It's done with dropping-mass gravimeters that measure the deflection of a laser beam, so it only relies on standard units of distance and time, and the speed of light as measured in terms of those units, not on the definition of mass. Obviously this is crucial or else you'd need a definition of a kilogram in order to calibrate your Kibble Balance.
So you can do this on Jupiter just fine, but you first have to measure the local gravitational field and adjust the amount of current you feed the Kibble Balance to balance against your kilogram test mass.
The biggest downside of this new method of defining the kilogram is that turning the definition into a measurement is incredibly precise and difficult work. It's so expensive to do correctly that for the foreseeable future there will probably only be a handful of wealthy countries who bother to do it. This means that for practical work, the definition will just be used to calibrate the exemplars that are used today, and everything else will continue as always. But it does mean that we now have a definition which is independent of those exemplars and guaranteed to be perfectly unchanging as long as the Planck constant remains constant.
Slashdot Mirror
That's fine for you, but if you're building phones for a large base of consumers it's better to serve their needs, even those that you think are silly.
Sorry, the "none of them" bit should have been quoted. That was from the post I replied to, not from me. I included the link to the court ruling allowing the emoluments suit to go forward.
True. And, they may be right!! See also, "I want to invest in the housing market in 2008". If enough people do a dumb thing, the government sometimes makes it not dumb retroactively.
Even without government involvement, if, for example, every single job applicant for a position you're trying to fill has posted videos of themselves doing crazy, stupid or illegal things on Facebook, then you have to ignore that, or hire no one. There's a notion of "radical transparency" that says that if we can see all the details of everyone's lives, we'll realize that everyone is a screwup at least some of the time and stop paying attention to those screwups.
That said, I don't post my own stupid mistakes on the web, and I discourage my kids from doing it either.
I'm curious then why I get at least one recruiter per month - for years and years - soliciting me to work as a contractor for Big Brother Google.
Me, too. Do they give you any information about what part of Google?
I see a lot of selfies floating around that clearly aren't about the background. Many, taken in bathrooms and whatnot, would definitely be better with the background de-emphasized.
Yeah like Google is a realistic average employer.
Read the thread, starting with AmiMoJo's post.
HTTPS does three things
a) Identify the remote site
b) Encrypt the traffic
c) Ensure the integrity of the traffic
FTFY. And the item you forgot is just as important as (a), and generally more important than (b).
There are several clauses relating to emoluments.
There are three. Two of them apply to the president, one barring him from receiving foreign emoluments, one barring him from receiving anything other than his salary from the federal government or the states. Regardless, Trump is in violation of none of them.
Federal District Judge Peter Messitte thinks there might be.
Further, even if a sitting President is found to be in violation, Congress will just grant an exception.
That seems much less likely than it did 15 days ago.
If you have an iPhone it already does this with two cameras working in tandem on the back.
Pixel 2 and Pixel 3 do this with a single camera The sensors have "split pixels", so when you take one photo with one camera you're actually getting two photos, with a tiny separation. This separation is enough, it turns out, to compute a depth map. Not good enough for actual 3D photography, but more than adequate for doing what most two-camera phones do: make portraits with a blurred background. And since two cameras aren't required, it works with the front camera, too. Bokeh for your selfies.
because they won't release those stats. They use contractors to hide it. The published stats are for FTEs. A huge part of the reason to use contractor firms for that is to hide those stats. Anecdotally the last 4 places I've worked at are pushing 80%.
From what I can see at Google there are almost no contractors doing engineering work. I'd be surprised if it was even 1%.
Bitcoin uses SHA-256(SHA-256(nonce+data)) < target value. They don't actually look for collisions.
And if they did look for collisions, at the peak hash rate thus far achieved (about 2^69 hashes per second) it would still take about 2^161 years to reach a 50% probability of finding one. This would make BTC transactions a bit slower than they are now.
Google allows users to export all of their data. Hasn't seemed to destroy Google.
A) Google definitely does NOT permit exporting all their data or even accessing all of it. Some yes but definitely not all.
Cite? Looks like all of it to me. They can't include anything they infer is *probably* about you, such as browsing history while not logged in, because if it's not actually yours giving it to you would violate the privacy of whoever it was.
D) Meta-data about group tendencies are as valuable as personal information and they don't (and won't) give you access to that.
Google sells such aggregated statistical data.
Woz also suggested that Facebook should allow users to export their data so they could upload it onto competing social networks.
You know, Woz is seemingly a very decent but this is just almost weapons grade stupid. How about Woz tell Apple to drop all their patents and open source their software? Because that's the functional equivalent of what he is suggesting. He's telling Facebook to hand over the crown jewels of their empire which is data about their customers.
Google allows users to export all of their data. Hasn't seemed to destroy Google. I'll grant that Google has assets other than user data, but user data is still pretty big.
The essential fiduciary duty of a publicly traded company is to maximize profits for its shareholders without regard to any moral considerations.
This is untrue.
The essential duty of a publicly-traded company is to work towards the goals outlined in its articles of incorporation and IPO letters, and perhaps the will of the shareholders if votes are held to alter those documented goals (not common). The goals nearly always include generating profits, but that's often not the only goal, and sometimes it isn't even the primary goal. It is always a goal, because all publicly-traded corporations are for-profit (US law bars non-profit corporations from selling shares).
And even for corporations that do have profit generation as their primary, or only, goal, it's still not true that directors and executives will be held legally accountable for failing to maximize profit. In theory that's possible, in practice it only happens with the most egregious of mismanagement.
There are characteristics of large organizations that tend to dilute moral concern and enable otherwise reasonably-moral individuals to make extremely amoral decisions. But it has nothing to do with fiduciary duty. Please, let's kill this tired old trope.
With precise control of the cars, it should be feasible to get inter-car gaps down to only a few feet, or even to zero, locking them together into an ad-hoc, very temporary train. If we assume only one car per second, that's 480-960 people per minute, in the same ballpark as a subway. At 4.4 cars per second it could carry 2100 to 4200 people per minute.
Ah, but this is science fiction. Currently we don't know how to do this in a safe manner.
Actually, quite a lot of research has already gone into doing this for highway vehicles. One company (Peloton Technology) even has a system on the market for "platooning" pairs of semi trucks. NIST and US DOT both have active research programs, as do several organizations in the EU. Demonstrations of safe and effective platooning have been done with inter-vehicle distances as small as 0.6 seconds at speeds up to 70 mph, including automatic handling of unexpected merges from vehicles not part of the system. Nearly all of the safety issues in such systems arise from the presence of human-operated vehicles on the road. Those wouldn't be present in Musk's system, allowing the tolerances to be closed considerably, but even with a very conservative 0.6 second gap, that's still 72 cars per minute.
Musk himself only projects 30 second headways , a far cry from 1 second or 0.23 second headways like you suggest.
That was specific to the Chicago project, where only ~2000 passengers per hour were thought to be needed. The article said nothing about "30 second headways", it said a car would leave every 30 seconds... that's technically the same thing, but the focus is on passenger departure rates, not safe vehicle packing rates. There's no indication that anyone thinks those 30-second gaps are necessary for safety. Indeed that would be ridiculous, equivalent to maintaining half-mile separations between cars on a highway. If an uncontrolled environment like a highway doesn't need 30-second gaps, a controlled tunnel certainly does not.
There is precise definition of "profit" in accounting, and Tesla is not doing it.
The precise definition is embodied in GAAP. Tesla turned a >20% GAAP profit in the third quarter. It remains to be seen if Tesla can sustain that, but teardowns show Tesla's vehicles to all have healthy per-unit profit margins and they seem to have turned the corner on their production issues, so it's reasonable.
What makes this short tunnel so worthy of praise?
What made SpaceX's first test launch worthy of praise? The first Falcon 1 flew for 41 seconds and landed (impacted is more accurate) 250 feet from the launch site.
You've got to start somewhere. Musk is excited about getting started, that's all. Clearly there's nothing particularly special about this tunnel, other than it's the beginning of what he has planned. Yes, that means any praiseworthiness is based entirely on the assumption that more is to come. But that seems like a reasonable expectation to me.
Musk proposes that each vehicle carry only 8 to 16 passengers. A full subway train, in contrast, carries over 1000 passengers. Musk plans for a vehicle every 30 seconds, compared to every 90 seconds for a modern subway line. So Musk's system will be able to carry 16-32 people per minute, compared to a subway which carries around 700 people per minute.
You're assuming Musk's system uses the same layout as subway tunnels, which it wouldn't.
In Musk's system, cars would not generally accelerate or decelerate in a transit tunnel (there may be a tiny bit as the control system accelerates or decelerates cars a bit to make room for a merging car), so you could have an almost continuous stream of cars going by. Cars would enter or leave a transit tunnel through an entrance or exit ramp tunnel. Think of a divided highway with entrance and exit ramps and overpasses... but with all vehicles centrally-controlled.
If we assume a 60 mph tunnel speed (88 fps), and individual cars that are 20 feet in length (that's generous), then we have 4.4 car lengths passing a fixed point each second. With precise control of the cars, it should be feasible to get inter-car gaps down to only a few feet, or even to zero, locking them together into an ad-hoc, very temporary train. If we assume only one car per second, that's 480-960 people per minute, in the same ballpark as a subway. At 4.4 cars per second it could carry 2100 to 4200 people per minute.
For that matter, if you reach the maximum throughput of a single transit tunnel, based on the minimum inter-car gaps and maximum speed dictated by safety, turn sharpness, etc., you can always do what highways do to increase capacity: add another lane by digging a parallel tunnel.
The other limiting point for throughput is stations. Stations would have to be sized appropriately. Suppose a subway station has a 1000-passenger train every 90 seconds, and that each train loads/unloads at most 100 people. With Musk's system the station would therefore have to load/unload, say, 12 cars every 90 seconds, or one every 7.5 seconds. Assuming it takes people 30 seconds to get on or off (that's generous, I think, except for disabled people), that means you need to have four cars unloading and four cars loading at once, so you'd need a loading/unloading bay of perhaps 10 car lengths, or maybe two bays of five car lengths. Or maybe you'd need a bit more because there would also be a sorting process going on; people going to the same destination would be assigned to the same car. I'm imagining that people would be informed by their phones which bay they should to go and which car they should get on. So you might want to have more bays to allow same-destination groups to "accumulate".
Actually, for routing of foot traffic, it might make more sense to separate loading and unloading entirely, since people exiting a car would be at their destination and exiting the system (transfers don't make sense in Musk's system). Street entrances/exits should probably be separated so all foot traffic is always moving the same direction. You might end up with station designs where each station is split into a pair of disjoint stations, one for people entering and another for people leaving.
And, of course, with a subway the 900 people who don't get on or off at the station still have to stop and wait for those who do. In Musk's system they'd just whiz right past the station.
Secondly, I don't know about USA regulations, but in the UK the railway construction regulations do not permit more than a cetain modest amount of cant; AFAIR is is about 6 degrees. The reason is to avoid standing passengers falling over or merely being discomforted if the train has to stop at those places for signals or any other reason. You might think that rule is too cautious, but that is how it is and I have no doubt there are similar regulations concerning roads, although not for fairground rides.
If you get rid of the rails and use wheeled skates, then you don't have to have a fixed cant at all. Just have the skates ride as far up the side as necessary at the given speed so the passengers don't feel any side acceleration. They'll experience an increase in "downward" acceleration during turns, of course, so it will make sense to set an upper limit on allowable gee force.
Major OS updates? Yes. That's acceptable. But it'll also receive security updates for much much longer. That's standard for all but no-name-company phones.
I wish it were standard. It's not. Unfortunately, it's much better than the standard, which is no commitment to updates at all.
I bet it'll come with two whole years of updates!
Two years!!!
I'm sure it'll be three years, same as the other Pixel 3s.
I guess if you only ever spend $15 on ear buds that may be fine.
Meh. I use these. $20. They sound fine. Great battery life, too. And wireless >> wired. Wires are a PITA.
infinite number of significant figures
First thanks for the explanation, but are you sure your meant 'infinite'?
Unlimited is probably a better word.
By it's very definition a balance in independent of gravity. The balance will remain the same and that 1Kg *of mass* will work just as well if you try this experiment on Jupiter.
But a Kibble Balance isn't that sort of a balance. It's a "single pan balance" which balances gravitational acceleration against acceleration caused by a magnetic field. So the Kibble Balance is very sensitive to changes in the gravitational field.
Luckily, it's possible to measure the local force of gravity with extreme precision, without reliance on the definition of the kilogram. It's done with dropping-mass gravimeters that measure the deflection of a laser beam, so it only relies on standard units of distance and time, and the speed of light as measured in terms of those units, not on the definition of mass. Obviously this is crucial or else you'd need a definition of a kilogram in order to calibrate your Kibble Balance.
So you can do this on Jupiter just fine, but you first have to measure the local gravitational field and adjust the amount of current you feed the Kibble Balance to balance against your kilogram test mass.
The biggest downside of this new method of defining the kilogram is that turning the definition into a measurement is incredibly precise and difficult work. It's so expensive to do correctly that for the foreseeable future there will probably only be a handful of wealthy countries who bother to do it. This means that for practical work, the definition will just be used to calibrate the exemplars that are used today, and everything else will continue as always. But it does mean that we now have a definition which is independent of those exemplars and guaranteed to be perfectly unchanging as long as the Planck constant remains constant.
Wrong, the pound was defined as that in 1959. nothing changed here
Since the pound is defined in terms of the kilogram, changing the definition of the kilogram implicitly changes the definition of the pound as well.