You would be surprised. I interviewed at a place a couple years ago, where the CEO told me that "backups have no ROI, put it in CodeCommit." Needless to say, I didn't take that job.
Assuming the version control system is either backed up or distributed, that CEO was right. What's the point of having extra backups if every employee already has a backup?
In a true Free Market and the Libertarian Way, we should let any car into this country and if it's defective or unsafe, after a few people die, then folks would realize it's a crap car and stop buying it.
And then they'll change the name and the color of the paint, and start selling a different car that's just as bad as the old one. Rinse, repeat. The quality of goods designed in China doesn't seem to be improving at all, from what I've seen. If anything, it seems to be getting worse, on average, because instead of five or ten copies of the same product with different names, there are a hundred, all with the same serious flaws, and higher-quality products are impossible to find because they are buried in the ever-growing pile of crap.
Most of the good gear coming out of that country seems to be designed in other countries, with the companies that designed the product randomly inspecting the factories to keep them honest, and randomly inspecting the products to ensure that they aren't substituting lower-quality parts.
Well, the latter think "this will probably happen to me if I live long enough". A 1% annual chance over a 100-year lifespan works out to a 63% lifetime chance.
Which qualifies as "probably", as in "more likely than not".
I don't think random chance is a great model for evaluating the likelihood of your job being automated, though.
To some extent, it is. Lots of technological improvements are unpredictable, resulting from some unexpected innovation that suddenly makes some previously impossible task possible. Of course, some improvements are incremental, and highly predictable, so it's not pure chance. Some of the factors that make a job more likely to be automated are:
Repeatability. You do the same thing in response to the same situation.
Limited variation in inputs. You don't work with materials that are different every time.
Cost of labor. You get paid above minimum wage.
Consistent (fixed) location or location-independent: Your job is at a physical place of business, or can be relocated to a centralized place of business without adversely affecting performance.
For example, consider a burger restaurant. The people at the front counter are already being replaced. The burger flippers are next, because that's trivial to automate. Building the sandwich is harder because lettuce is inconsistent, but still possible, so that happens a little later.
On the flip side, plumbers deal with shoddy repairs done by other plumbers, so the input is highly variable. And a plumber can't do the work anywhere but on the customer site. So that fails two of the tests, which means it will take much longer to automate. However, construction can be automated, and for houses built with such automation, the design of the plumbing would be well understood, which eliminates one of those pain points. Those houses will be able to use auto-plumbers much sooner.
And so on. But none of those factors is a show-stopper for partial automation. For example, delivery robots that deliver to someone's apartment need to be smart enough to find a particular apartment, but you could build a dumb delivery robot that can't do that, and then have a person guide it to its final destination. And then, while the self-driving car drives the robot to the next location, that person can guide two or three more robots into apartments, and verified the street address for twenty more. And then you've reduced staffing by more than an order of magnitude even though the robots can't even fully do the job that the humans did. And it's pretty much random luck as to whether a given industry decides to try such an experiment.
Does anyone seriously think that a software program is going to replace an experienced and skilled surgeon? Just think of the requirements document for something like that. You'd better hope that the developers on that project know that they're doing and even then, how can they possibly write code to deal with every possible thing that could go wrong during an operation.
A lot of surgery is already being done by robots, under the guidance of a human. I could easily see machine learning algorithms watch the surgeons as they do those surgeries, until at some point you would turn the algorithm loose and let it do surgery on its own, using a small pool of shared humans who would take over if the robot encounter something that its programming can't handle. I wouldn't expect to see anything approaching full automation before the current crop of surgeons retire, but college students deciding whether to choose medicine as a career should at least keep in mind that they might be downsized before they would ordinarily retire, and manage their finances accordingly.
However.. There is this lithium ion battery inside.. Damage it somehow, leave something extra in accidentally (there's not a lot of space), have it puncture, and then burn the house down next time you charge the device.. Or even worse, have it catch fire next time in an airplane. You may have noticed that nowadays dropping a phone inside an airplane seat puts the cabin staff basically in panic mode. To avoid damaging the battery and causing a fire you are not allowed to move the seat at all.
Having replaced the lithium ion battery in my previous cell phone, I think you're grossly overstating the risk. Most of the risk is in accidentally puncturing the old one while you're removing it. Once you get the old one out, the odds of you doing anything that will cause it to later fail are pretty small unless the battery is defective.
Only because the materials used in clothing are flexible enough to make fully automated manufacturing challenging, and even that is likely to change in the very, very near future. Oh, and for shoes, it already has changed to a large extent.
And even in factories with low levels of automation, large parts of the work are still done by machine. Humans guide the material through the machines, but the sewing is still done by machine, not by hand stitching, which means orders of magnitude fewer people are involved than historically were. So when I say that manufacturing is mostly automated, that includes garments and shoes.
how is the work on your cars and trucks done? oh yeah, by mechanics.
The mechanics plug in a diagnostic machine, it figures out what part to replace, and a person replaces it. It's only a matter of time before that final step is automated. Once you train one robot to do the work, you can have a million robots doing that same task for the cost of building the hardware. The leap from robot manufacturing to robot repair is a lot smaller than you seem to believe. The minute one car company does it, they'll all rush to do it, because the labor cost on car repairs is downright insane. Frankly, if any industry is ripe for automation, that's it.
how is building inspection done? oh, by people.
Only because buildings are still built by people. When robot house builders take over that industry, the verification will be done by someone signing off on the wiring diagram, and inspections will be as unnecessary as the builders.
If you look at electronics, engineers design things, machines build things, machines package things up for delivery, and soon machines will handle the delivery, too. If you honestly believe that any other manufacturing industry is significantly different in some way that will make it impractical to automated, I have a bridge to sell you.
And although you are correct that there will still be people doing repairs for a long time to come, that is true only for the sorts of repairs that involve going to the customer site, such as plumbing, refrigerator repair, etc. Car repairs and electronic repairs are on the short list for automation. Apple is already doing cell phone screen repairs by automated machine. By 2030, the only people doing electronic repairs by hand will be the independent repair shops, assuming the manufacturers' zero-labor repairs don't undercut them and run them out of business.
this won't change anytime soon, because AI is mostly a farce with nothing fundamental new in decades.
This has already changed, and if you haven't noticed, it's no surprise that you still think AI is a farce with nothing fundamentally new in decades.
The difference between people who understand statistics and people who don't is that people who don't understand statistics see a 1% annual chance and think, "This will never happen to me," whereas people who do understand statistics think, "This will eventually happen to me if I live long enough," and plan accordingly.
It isn't a question of whether any given person's job will be replaced, but rather when. Eventually, nearly everything will be automated. Manufacturing is already mostly there. Retail and fast food will be next, replaced by touchscreen ordering, website-based ordering, delivery robots, etc. The trucking industry will follow shortly thereafter. Doctors likely will be replaced by a machine learning model within a couple of decades at most, though surgeons and nurses will hang around somewhat longer. Police will eventually be replaced by drones. Office workers will be slowly become unnecessary as the people they support cease to work.
At some point, the only jobs left will be writing software for the machines, designing the machines, jobs in arts/entertainment, and maybe firefighter robot drivers. The only real questions are how long it will take and whether the rate of redundancy significantly exceeds the rate of attrition.
This is why serious offensive drivers learn to turn their turn signals on randomly and then leave them on for extended periods of time without actually turning or changing lanes. Only by actively engaging in disinformation can you truly keep other drivers on their toes.
Which means that should be the default behavior. Actually, the default should probably involve caches measured in days, with some sort of automatic purge-on-modification, but I have no idea if WP supports that.
The problem with WordPress isn't that it can't be a good citizen, but rather that by default (which is how probably 99% of site admins leave it configured), it is an abusive nightmare.:-)
The situation is different with PCs with regard to older-style ports (COM, LPT); then again, there's an aftermarket, and a lively adapter and components market.
The situation with PCs is largely irrelevant. You don't carry your printer around with you in your pocket. You do carry your cell phone and earbuds in your pocket. And that's why dongles for mobile devices (including laptops) are bad—particularly when they are required for things that you typically do when away from home, like connecting your new MacBook Pro to a TV set (redacted swearing at Apple) or reading photos from SD cards from your camera.
Why is this comment magically at -1? This is what FB is really asking. Is it okay for us to continuously monitor private messaging on our platform?
It is at -1 because the poster got modded down so many times that his/her posts automatically start at -1.
They just use the "think about the children" loaded question so that you knee-jerk answer "Yes," when the answer should totally be, "If you're providing a private messaging service, then that should remain private. You should comply with any valid search warrant to monitor communications, but we don't do precrime, so we can't just monitor everyone, without a writ."
Facebook already allows parents to monitor their children's accounts, including, AFAIK, their private messages. So what they're really asking is, "Should we babysit your children so you don't have to pay attention to what they're doing online?" to which people's answers will vary widely.
1. Maybe FB should not allow children to register.
And they'll just go back to doing what they did before Facebook lowered the age to 13: lie about their age. Kids do that, you know.
2. Maybe FB should not allow children to add / request to be added as 'friends' with adults.
Like their parents, aunts and uncles, etc. So now, all the pedophiles can create accounts and pretend that they are 15, then court all the children they want, and the kids' parents will be none the wiser. Why do I have a feeling this won't work out as well as you think?
3. Maybe FB should be monitoring all conversations between men and children and then using various cues (as determined by some form of AI I suppose) to record conversations and then to pass information to a human/to authorities.
And this assumes that neither the would-be child abusers nor the children lie about their age. Otherwise, this fails completely.
4. Maybe FB should be monitoring all conversations at all times, record them and use AI to sensor/redact information and pass data to authorities.
That will work as long as it is secret. As soon as anybody finds out they're doing it, the first thing a child exploiter will do is send a "chat with me on [insert other service here]" message, and Facebook won't see anything interesting.
No, none of these things will work in practice. What might work is detecting sexually explicit images and, if uploaded via a child account, reporting them to the parents, but really, there's no substitute for proper parenting—teaching your kids not to do things like that—and any technology that tries to be a substitute for good parenting is doomed to failure.
In theory, yes. In practice, it's a bunch of Kickstarter projects, and if you happen to be there at the right time, you can order one. Otherwise, you're screwed. None of the major case manufacturers have gotten in on the act, unless I missed something, which is a shame, because that might actually be enough to make me consider upgrading my 6s before its last gasp.
Obviously, if this becomes law, the only way to comply is to automatically take down any reported content without review. My only hope is that they take them down just in the EU so those EU politicians can see for themselves what happens when you let their opposition have that kind of power.
This. I think we're all going to enjoy watching as every EU political speech get reported by somebody as terrorism and taken down, one right after the next.
It is very much an engineering decision, because you can injection-mould optically clear plastic very, very simply and very, very cheaply and in a normal environment I'd expect it to last ages.
You can also injection mould glass fairly cheaply, AFAIK. The metal mould doesn't really care if you're injecting plastic resin or molten glass, so long as the melting temperature of the glass is significantly lower than the melting temperature of the mould itself. There are companies making aspheric lenses with moulding all the time.
Either way, the point I was trying to make was not that all LIDAR lenses should be made of glass, but rather that if you're constantly having problems with lenses getting scratched, in the long run, it probably would makes sense to try to convince the manufacturer to sell a high-end version with a moulded glass lens, even if the resulting lens costs more and is heavier—unless, of course, you're constantly having things actually hitting the lens itself, in which case you have bigger problems than scratches.:-)
It shouldn't be that much more expensive to use glass. Even if you have some crazy lens design that can't easily be ground and must be moulded instead, you can always bond a thin layer of glass or sapphire to the exposed face. No, it isn't an engineering decision, or if it is, it's a terribly bad one.
More likely, it's a marketing decision. Why spend an extra buck to bond a thin layer of sapphire to the product when you can instead sell $5 replacement lenses for $50 each?
Of course, the workaround is for Redbox to add one extra line to their terms of use:
"By purchasing a digital license, you acknowledge that you have purchased a retail copy of the DVD from Redbox and resold that retail copy of the DVD back to Redbox for the purchase price minus the cost of this digital license."
And then account for each digital license sale as a $5.01 sale and a $0.01 purchase. And Disney still fails at that point.
You can't file an insurance claim without a police report, so when something gets stolen, you do need the police to respond, even though they probably won't be able to do anything about the theft itself other than documenting it....
Slashdot Mirror
Assuming the version control system is either backed up or distributed, that CEO was right. What's the point of having extra backups if every employee already has a backup?
And then they'll change the name and the color of the paint, and start selling a different car that's just as bad as the old one. Rinse, repeat. The quality of goods designed in China doesn't seem to be improving at all, from what I've seen. If anything, it seems to be getting worse, on average, because instead of five or ten copies of the same product with different names, there are a hundred, all with the same serious flaws, and higher-quality products are impossible to find because they are buried in the ever-growing pile of crap.
Most of the good gear coming out of that country seems to be designed in other countries, with the companies that designed the product randomly inspecting the factories to keep them honest, and randomly inspecting the products to ensure that they aren't substituting lower-quality parts.
Isn't the source... the movie/TV industry?
Ah. My bad. Well, it's still true, but only for a large value of "long enough". ;-)
Which qualifies as "probably", as in "more likely than not".
To some extent, it is. Lots of technological improvements are unpredictable, resulting from some unexpected innovation that suddenly makes some previously impossible task possible. Of course, some improvements are incremental, and highly predictable, so it's not pure chance. Some of the factors that make a job more likely to be automated are:
For example, consider a burger restaurant. The people at the front counter are already being replaced. The burger flippers are next, because that's trivial to automate. Building the sandwich is harder because lettuce is inconsistent, but still possible, so that happens a little later.
On the flip side, plumbers deal with shoddy repairs done by other plumbers, so the input is highly variable. And a plumber can't do the work anywhere but on the customer site. So that fails two of the tests, which means it will take much longer to automate. However, construction can be automated, and for houses built with such automation, the design of the plumbing would be well understood, which eliminates one of those pain points. Those houses will be able to use auto-plumbers much sooner.
And so on. But none of those factors is a show-stopper for partial automation. For example, delivery robots that deliver to someone's apartment need to be smart enough to find a particular apartment, but you could build a dumb delivery robot that can't do that, and then have a person guide it to its final destination. And then, while the self-driving car drives the robot to the next location, that person can guide two or three more robots into apartments, and verified the street address for twenty more. And then you've reduced staffing by more than an order of magnitude even though the robots can't even fully do the job that the humans did. And it's pretty much random luck as to whether a given industry decides to try such an experiment.
A lot of surgery is already being done by robots, under the guidance of a human. I could easily see machine learning algorithms watch the surgeons as they do those surgeries, until at some point you would turn the algorithm loose and let it do surgery on its own, using a small pool of shared humans who would take over if the robot encounter something that its programming can't handle. I wouldn't expect to see anything approaching full automation before the current crop of surgeons retire, but college students deciding whether to choose medicine as a career should at least keep in mind that they might be downsized before they would ordinarily retire, and manage their finances accordingly.
Having replaced the lithium ion battery in my previous cell phone, I think you're grossly overstating the risk. Most of the risk is in accidentally puncturing the old one while you're removing it. Once you get the old one out, the odds of you doing anything that will cause it to later fail are pretty small unless the battery is defective.
Only because the materials used in clothing are flexible enough to make fully automated manufacturing challenging, and even that is likely to change in the very, very near future. Oh, and for shoes, it already has changed to a large extent.
And even in factories with low levels of automation, large parts of the work are still done by machine. Humans guide the material through the machines, but the sewing is still done by machine, not by hand stitching, which means orders of magnitude fewer people are involved than historically were. So when I say that manufacturing is mostly automated, that includes garments and shoes.
The mechanics plug in a diagnostic machine, it figures out what part to replace, and a person replaces it. It's only a matter of time before that final step is automated. Once you train one robot to do the work, you can have a million robots doing that same task for the cost of building the hardware. The leap from robot manufacturing to robot repair is a lot smaller than you seem to believe. The minute one car company does it, they'll all rush to do it, because the labor cost on car repairs is downright insane. Frankly, if any industry is ripe for automation, that's it.
Only because buildings are still built by people. When robot house builders take over that industry, the verification will be done by someone signing off on the wiring diagram, and inspections will be as unnecessary as the builders.
If you look at electronics, engineers design things, machines build things, machines package things up for delivery, and soon machines will handle the delivery, too. If you honestly believe that any other manufacturing industry is significantly different in some way that will make it impractical to automated, I have a bridge to sell you.
And although you are correct that there will still be people doing repairs for a long time to come, that is true only for the sorts of repairs that involve going to the customer site, such as plumbing, refrigerator repair, etc. Car repairs and electronic repairs are on the short list for automation. Apple is already doing cell phone screen repairs by automated machine. By 2030, the only people doing electronic repairs by hand will be the independent repair shops, assuming the manufacturers' zero-labor repairs don't undercut them and run them out of business.
This has already changed, and if you haven't noticed, it's no surprise that you still think AI is a farce with nothing fundamentally new in decades.
The difference between people who understand statistics and people who don't is that people who don't understand statistics see a 1% annual chance and think, "This will never happen to me," whereas people who do understand statistics think, "This will eventually happen to me if I live long enough," and plan accordingly.
It isn't a question of whether any given person's job will be replaced, but rather when. Eventually, nearly everything will be automated. Manufacturing is already mostly there. Retail and fast food will be next, replaced by touchscreen ordering, website-based ordering, delivery robots, etc. The trucking industry will follow shortly thereafter. Doctors likely will be replaced by a machine learning model within a couple of decades at most, though surgeons and nurses will hang around somewhat longer. Police will eventually be replaced by drones. Office workers will be slowly become unnecessary as the people they support cease to work.
At some point, the only jobs left will be writing software for the machines, designing the machines, jobs in arts/entertainment, and maybe firefighter robot drivers. The only real questions are how long it will take and whether the rate of redundancy significantly exceeds the rate of attrition.
This is why serious offensive drivers learn to turn their turn signals on randomly and then leave them on for extended periods of time without actually turning or changing lanes. Only by actively engaging in disinformation can you truly keep other drivers on their toes.
Which means that should be the default behavior. Actually, the default should probably involve caches measured in days, with some sort of automatic purge-on-modification, but I have no idea if WP supports that.
The problem with WordPress isn't that it can't be a good citizen, but rather that by default (which is how probably 99% of site admins leave it configured), it is an abusive nightmare. :-)
The situation with PCs is largely irrelevant. You don't carry your printer around with you in your pocket. You do carry your cell phone and earbuds in your pocket. And that's why dongles for mobile devices (including laptops) are bad—particularly when they are required for things that you typically do when away from home, like connecting your new MacBook Pro to a TV set (redacted swearing at Apple) or reading photos from SD cards from your camera.
Since we are being pedantic: America is two continents, not a continent.
Nope. Facebook requires kids to be 13 or older, unless they are in a jurisdiction that imposes additional restrictions.
It is at -1 because the poster got modded down so many times that his/her posts automatically start at -1.
Facebook already allows parents to monitor their children's accounts, including, AFAIK, their private messages. So what they're really asking is, "Should we babysit your children so you don't have to pay attention to what they're doing online?" to which people's answers will vary widely.
And they'll just go back to doing what they did before Facebook lowered the age to 13: lie about their age. Kids do that, you know.
Like their parents, aunts and uncles, etc. So now, all the pedophiles can create accounts and pretend that they are 15, then court all the children they want, and the kids' parents will be none the wiser. Why do I have a feeling this won't work out as well as you think?
And this assumes that neither the would-be child abusers nor the children lie about their age. Otherwise, this fails completely.
That will work as long as it is secret. As soon as anybody finds out they're doing it, the first thing a child exploiter will do is send a "chat with me on [insert other service here]" message, and Facebook won't see anything interesting.
No, none of these things will work in practice. What might work is detecting sexually explicit images and, if uploaded via a child account, reporting them to the parents, but really, there's no substitute for proper parenting—teaching your kids not to do things like that—and any technology that tries to be a substitute for good parenting is doomed to failure.
Correction. I won't get what makes them the most profit, and they'll just stand around slack-jawed wondering why sales are soft.
In theory, yes. In practice, it's a bunch of Kickstarter projects, and if you happen to be there at the right time, you can order one. Otherwise, you're screwed. None of the major case manufacturers have gotten in on the act, unless I missed something, which is a shame, because that might actually be enough to make me consider upgrading my 6s before its last gasp.
This. I think we're all going to enjoy watching as every EU political speech get reported by somebody as terrorism and taken down, one right after the next.
You can also injection mould glass fairly cheaply, AFAIK. The metal mould doesn't really care if you're injecting plastic resin or molten glass, so long as the melting temperature of the glass is significantly lower than the melting temperature of the mould itself. There are companies making aspheric lenses with moulding all the time.
Either way, the point I was trying to make was not that all LIDAR lenses should be made of glass, but rather that if you're constantly having problems with lenses getting scratched, in the long run, it probably would makes sense to try to convince the manufacturer to sell a high-end version with a moulded glass lens, even if the resulting lens costs more and is heavier—unless, of course, you're constantly having things actually hitting the lens itself, in which case you have bigger problems than scratches. :-)
You mean the FBI isn't a secret society? Crap. They tricked me again. Now what am I going to do with a thousand gallons of goats' blood?
It shouldn't be that much more expensive to use glass. Even if you have some crazy lens design that can't easily be ground and must be moulded instead, you can always bond a thin layer of glass or sapphire to the exposed face. No, it isn't an engineering decision, or if it is, it's a terribly bad one.
More likely, it's a marketing decision. Why spend an extra buck to bond a thin layer of sapphire to the product when you can instead sell $5 replacement lenses for $50 each?
Of course there's a reason. Continuing to use the digital copy costs Disney potential sales for replacement discs. :-D
Oh, you mean a good reason. Never mind.
Of course, the workaround is for Redbox to add one extra line to their terms of use:
"By purchasing a digital license, you acknowledge that you have purchased a retail copy of the DVD from Redbox and resold that retail copy of the DVD back to Redbox for the purchase price minus the cost of this digital license."
And then account for each digital license sale as a $5.01 sale and a $0.01 purchase. And Disney still fails at that point.
You can't file an insurance claim without a police report, so when something gets stolen, you do need the police to respond, even though they probably won't be able to do anything about the theft itself other than documenting it....