The example I keep seeing used is self-driving vehicles, particularly trucks.
They've already proved themselves in mining contexts - they use less fuel, wear their tyres less, less maintenance, downtime, and of course, no wages to be paid.
People are falling over themselves to get them approved for road use. Truck driver is 3.5M jobs in the USA. There are about 285,000 HGV drivers in the UK.
The trend is already that middle class jobs are being eroded and replaced with low waged work. Truck driver is unglamorous but they probably count as middle class guys with the wages they get.
What jobs are we going to find for truck drivers? They're not all going to train up to be robot-truck mechanics (as above, robot trucks are pretty much existing trucks with a few extra sensors and something wired to the power steering etc - and require less maintenance). You'll have all those middle-class people with a lower disposable income, so the market for services and consumer goods will shrink, so what industry will expand or arise to employ them?
Tech changes that increase production, expand wealth and (eventually) the job market. Spinning Jenny and her ilk turned cotton shirts from a luxury into a mass-market commodity.
Tech changes that just do existing work with less human labour do not expand wealth. The automated trucks drivers, administrative assistants, warehouse pickers, burger flippers, etc, etc, that are coming down the pipe do not increase demand - if anything, they decrease it, by reducing the amount of wages entering the economy.
The scrutiny is the point. Blanket surveillance is shit for finding actual criminal / terrorist activity because the false positive rate means that your agents will all be tied up investigating bad leads forever.
As a tool to gain insight into a population and thus control over them, it's excellent.
CEOs operating for the long term good of their corporation instead of pumping their stock short term to get a stock option payout would be a marked improvement.
Wal-Mart isn't dumb. It recognises that automation is the backbone of it's ability to shift enormous volumes of goods. They're pioneers in a number of fields including RFID asset management, and they aren't going to start outsourcing their competitive advantage to contractors where they can't control who gets to use that competitive advantage next.
Spotted them doing this in White Collar : for a few episodes they couldn't stop* talking about the features of the new Taurus, principally the navigation and collision avoidance.
* had a short clumsy scene where they talked about it briefly and then got down to business.
Yeah, I'm a house-dweller, but I'd love a community kitchen that served healthy meals to be nearby. While in the UK, in a town you're never really very far from a pub that serves food, it's all too a la carte to be regarded as inexpensive.
Yeah, I've seen enough hideously expensive subcontracted projects that produce enormous bodies of crap code for multiples of my annual salary to form the opinion that the best option is in-house development. It's much easier to form a productive working relationship with someone that you know will be around in a few years to support the software you're developing for them, and you don't resent them for being paid a multiple of your annual salary.
But the point about waterfall sucking is true as well. Waterfall only works if you really are measuring twice, cutting once, and that only works for projects where the desired outcome is plain and simple. That does not describe modern software projects of any magnitude. Landing on the moon? You have a simple constrained goal, you only get to try once, you are willing to spend the budget required for the process to ensure that quality is baked in.
"Digitizing immigration forms" ought to have been simple, but I'm willing to bet it's all the hidden rules that live in institutional memory that sunk it. Agile is entirely appropriate for this kind of project - produce something that works on a basic level, keep making incremental improvements. Even just a plain version of the forms that you fill in with a keyboard instead of a pen would have been better than the paper ones because they would be guaranteed legible and easier/cheaper to copy.
An initial budget of $500M is utterly, utterly ludicrous. A budget like that is a problem for any project unless it's very very well planned out and very definitely necessary. Otherwise you have a huge cadre of middle management all struggling to justify their existence, inventing make-work. What we have here is a project that took that $500M and used it to get really good at inventing work to justify deadline extensions.
TFA discusses that it is exactly that which is the biggest problem. The robot chef being discussed takes prepared ingredients and cooks them.
Preparation is skilled work. As you say, there's removing bad bits of the ingredients, recognising which parts are unpalatable or tough, dealing with a variety of shapes, peeling, cutting, skinning of meat, removal of cartilage and bone and fat and tendons, scraping out seeds and pulp - the list of skills and their variations is large, and depends on not just visual recognition but also on touch and smell and sound and even things as subtle as regulating how much force you are using based on how much resistance your tools are meeting. There is a reason that a professional kitchen has chefs who specialize in food prep for maximum efficiency.
There are plenty of tough problems there that can benefit the robotics industry if they are solved, but they are so much harder than the problems the existing "chef" solves, which are mere cooking and assembly. Even though it nominally solves these problems, I still wouldn't trust the thing to do something as simple as cook a steak properly, a process that requires experience, judgement, and touch.
Exactly. People get so hung up on having a "perfect" system. I see this all the time in healthcare.
Start with something that's "no worse than paper". Just the fact that the form is guaranteed legible, can have basic field validation, and is capable of being copied and transmitted instantly pretty much gets you there straight away.
Most UK ISPs run a transparent web proxy ; the government could compel root CAs to provide a copy of their private keys so they they can dynamically sign man-in-the-middle certificates on those proxies. Of course that will lead to the keys inevitably being leaked and the collapse of internet commerce as we know it, but hey, something better will arise in it's place.
Pads become worthless once the entire entropy field has been mapped
Pseudo-scientific twaddle. The "rainbow table" you're talking about is not only infinite in size, by it's nature it also contains every possible plaintext message of any given length. Which means that you could just claim that the message was whatever you wanted it to be, within the size limit.
Who knows? Maybe this is what they'll claim they've done in the future to string up dissidents. But not until they've lobotomised a large swathe of the population.
If only there were some mechanism of disseminating and copying course materials that was virtually free and used equipment that everyone attending college can afford.
In this case the $180 reflects the number of courses the book applies to, and thus the amount of lock-in. If the exact same book wasn't mandated for so many courses, it would cost less, despite consisting of the same materials.
Didn't say "work". Said they wanted their low-bandwidth IM and email "to work", as in working communications (with your family and loved ones, as well as yes, your business associates).
Yes, it's self-limiting. So is herpes, but you don't want to go through that either.
If people weren't trying to do something futile and stupid (and rude, because of it's impact on others), then everyone's day would just run better.
Well, OK, it's not fair, but it's more fair than not having any metering.
Mobile wireless is a limited and non-fungible resources. On any given cell at any given time, there's a limited amount of bandwidth to be shared around. If that cell has enough bandwidth to serve everyone using it, they should be able to fill their boots. If that cell is congested, you need some mechanism to prevent all the bandwidth being consumed by a few users. When I'm in a setting like a railway station with a thousand people all wanting some bandwidth, it makes me angry to see the people streaming video on their phone like their entertainment is more important than everyone else who wants their email and IM to work.
What we have now does not really work ; it's a pricing mechanism of deterring excessive use, but it doesn't assign prices correctly, it just bills your bandwidth use at a flat rate. Bandwidth that is uncontested should be cheap, or free. Bandwidth that is contested should cost more. What really needs to happen in these circumstances is a kind of Dutch auction, where every handset has a notion of what price it's prepared to pay for bandwidth, and the cell sets the price according to how congested it is. But this would be too complicated to sell to most people.
Unmetered wireless bandwidth is not fair, because it's only "too cheap to meter" when the cell has very few clients. Part of the drive for 5G is to push things to this point - 4G too (or else why would they provide such tiny caps with such huge bandwidth that you can burn through it all in minutes - it's not about giving your more bandwidth, it's about having enough bandwidth to serve more customers). As long as we keep up thinking of new media that consume more bandwidth, wireless will not be good enough to serve crowded cells.
Not quite : public domain is not entirely open source and doesn't have quite the same list of positive points.
Not least, public domain works can be taken and used without attribution. If you're a software developer in a gig economy, your reputation matters.
And as you qualify : public domain in the USA.
And while the "PERMISSIVE LICENSING!!!" crowd will jump on me for this, there are advantages to copyleft licensing that you can't have if the work is in public domain, because these licenses require copyright to have force.
For example, software designed for interoperability is probably best licensed GPL or LGPL (to allow closed software to link to it), because this prevents one company making it's own "extensions" to it, thus rendering everyone else's implementation inferior. If you were a large player you could then exploit network effects to make your product the de-facto standard, eliminating the benefits of having interoperability in the first place. Copyleft licensing ensures that such improvements are available to everyone.
Yeah, my no.1 peeve with the UI - having panels you HAVE to visit to play the game being to the left and right of your cockpit window, accessible by keys - becomes just another bit of immersion with a headtracker on. The most natural thing in the world to turn your head to look at your target panel.
I also get what all those flight sims were doing with all those keys to "padlock view" your targets, etc. I never got on well with them and usually just stuck to a straight ahead view. Head tracking is just seamless and immersive compared to that.
My first wife earned far more than I did for many years and the tables only turned recently (and not by much).
We parted amicably, our only common financial interests were the house (we split evenly because while she earned more, I was more frugal and paid substantial chunks of the mortgage off), and our daughter.
My current GF earns about a quarter what I do but still insists on paying her way equally, which works fine because we both have pretty modest lifestyle demands.
The only kinds of women who are expensive are the ones who are into you for the money.
Slashdot Mirror
The example I keep seeing used is self-driving vehicles, particularly trucks.
They've already proved themselves in mining contexts - they use less fuel, wear their tyres less, less maintenance, downtime, and of course, no wages to be paid.
People are falling over themselves to get them approved for road use. Truck driver is 3.5M jobs in the USA. There are about 285,000 HGV drivers in the UK.
The trend is already that middle class jobs are being eroded and replaced with low waged work. Truck driver is unglamorous but they probably count as middle class guys with the wages they get.
What jobs are we going to find for truck drivers? They're not all going to train up to be robot-truck mechanics (as above, robot trucks are pretty much existing trucks with a few extra sensors and something wired to the power steering etc - and require less maintenance). You'll have all those middle-class people with a lower disposable income, so the market for services and consumer goods will shrink, so what industry will expand or arise to employ them?
Tech changes that increase production, expand wealth and (eventually) the job market. Spinning Jenny and her ilk turned cotton shirts from a luxury into a mass-market commodity.
Tech changes that just do existing work with less human labour do not expand wealth. The automated trucks drivers, administrative assistants, warehouse pickers, burger flippers, etc, etc, that are coming down the pipe do not increase demand - if anything, they decrease it, by reducing the amount of wages entering the economy.
The scrutiny is the point. Blanket surveillance is shit for finding actual criminal / terrorist activity because the false positive rate means that your agents will all be tied up investigating bad leads forever.
As a tool to gain insight into a population and thus control over them, it's excellent.
They actually want to make it illegal to even discuss the surveillance, so that could be interpreted as being in breach of the law.
CEOs operating for the long term good of their corporation instead of pumping their stock short term to get a stock option payout would be a marked improvement.
Wal-Mart isn't dumb. It recognises that automation is the backbone of it's ability to shift enormous volumes of goods. They're pioneers in a number of fields including RFID asset management, and they aren't going to start outsourcing their competitive advantage to contractors where they can't control who gets to use that competitive advantage next.
Spotted them doing this in White Collar : for a few episodes they couldn't stop* talking about the features of the new Taurus, principally the navigation and collision avoidance.
* had a short clumsy scene where they talked about it briefly and then got down to business.
Yeah, I'm a house-dweller, but I'd love a community kitchen that served healthy meals to be nearby. While in the UK, in a town you're never really very far from a pub that serves food, it's all too a la carte to be regarded as inexpensive.
Yeah, I've seen enough hideously expensive subcontracted projects that produce enormous bodies of crap code for multiples of my annual salary to form the opinion that the best option is in-house development. It's much easier to form a productive working relationship with someone that you know will be around in a few years to support the software you're developing for them, and you don't resent them for being paid a multiple of your annual salary.
But the point about waterfall sucking is true as well. Waterfall only works if you really are measuring twice, cutting once, and that only works for projects where the desired outcome is plain and simple. That does not describe modern software projects of any magnitude. Landing on the moon? You have a simple constrained goal, you only get to try once, you are willing to spend the budget required for the process to ensure that quality is baked in.
"Digitizing immigration forms" ought to have been simple, but I'm willing to bet it's all the hidden rules that live in institutional memory that sunk it. Agile is entirely appropriate for this kind of project - produce something that works on a basic level, keep making incremental improvements. Even just a plain version of the forms that you fill in with a keyboard instead of a pen would have been better than the paper ones because they would be guaranteed legible and easier/cheaper to copy.
An initial budget of $500M is utterly, utterly ludicrous. A budget like that is a problem for any project unless it's very very well planned out and very definitely necessary. Otherwise you have a huge cadre of middle management all struggling to justify their existence, inventing make-work. What we have here is a project that took that $500M and used it to get really good at inventing work to justify deadline extensions.
TFA discusses that it is exactly that which is the biggest problem. The robot chef being discussed takes prepared ingredients and cooks them.
Preparation is skilled work. As you say, there's removing bad bits of the ingredients, recognising which parts are unpalatable or tough, dealing with a variety of shapes, peeling, cutting, skinning of meat, removal of cartilage and bone and fat and tendons, scraping out seeds and pulp - the list of skills and their variations is large, and depends on not just visual recognition but also on touch and smell and sound and even things as subtle as regulating how much force you are using based on how much resistance your tools are meeting. There is a reason that a professional kitchen has chefs who specialize in food prep for maximum efficiency.
There are plenty of tough problems there that can benefit the robotics industry if they are solved, but they are so much harder than the problems the existing "chef" solves, which are mere cooking and assembly. Even though it nominally solves these problems, I still wouldn't trust the thing to do something as simple as cook a steak properly, a process that requires experience, judgement, and touch.
Exactly. People get so hung up on having a "perfect" system. I see this all the time in healthcare.
Start with something that's "no worse than paper". Just the fact that the form is guaranteed legible, can have basic field validation, and is capable of being copied and transmitted instantly pretty much gets you there straight away.
Or at least, a different design of fleshlight.
Most UK ISPs run a transparent web proxy ; the government could compel root CAs to provide a copy of their private keys so they they can dynamically sign man-in-the-middle certificates on those proxies. Of course that will lead to the keys inevitably being leaked and the collapse of internet commerce as we know it, but hey, something better will arise in it's place.
Pads become worthless once the entire entropy field has been mapped
Pseudo-scientific twaddle. The "rainbow table" you're talking about is not only infinite in size, by it's nature it also contains every possible plaintext message of any given length. Which means that you could just claim that the message was whatever you wanted it to be, within the size limit.
Who knows? Maybe this is what they'll claim they've done in the future to string up dissidents. But not until they've lobotomised a large swathe of the population.
If only there were some mechanism of disseminating and copying course materials that was virtually free and used equipment that everyone attending college can afford.
not much, though, even for a $180 text
In this case the $180 reflects the number of courses the book applies to, and thus the amount of lock-in. If the exact same book wasn't mandated for so many courses, it would cost less, despite consisting of the same materials.
Or there's MIT's Opencourseware on Linear Algebra, zero dollars.
Jeez, they are INFORMATION.
Information by nature is plentiful. Like a hole, the more you take, the more it grows.
MIT give away their course on Linear Algebra. If it's good enough for MIT, why is not not good enough for Fullerton?
Oh, right, because it's not a $180 textbook that pays royalties to the chair and vice-chair of maths.
Clearly you don't follow the news about our PM.
A better haiku would be
The Turgid Member,
Slipped into a dead pig's mouth,
Like he fucks the poor
Didn't say "work". Said they wanted their low-bandwidth IM and email "to work", as in working communications (with your family and loved ones, as well as yes, your business associates).
Yes, it's self-limiting. So is herpes, but you don't want to go through that either.
If people weren't trying to do something futile and stupid (and rude, because of it's impact on others), then everyone's day would just run better.
That is what I meant by "non-fungible" : data now, when you need your email, is not the same as data tomorrow.
I think metering is fair on mobile wireless.
Well, OK, it's not fair, but it's more fair than not having any metering.
Mobile wireless is a limited and non-fungible resources. On any given cell at any given time, there's a limited amount of bandwidth to be shared around. If that cell has enough bandwidth to serve everyone using it, they should be able to fill their boots. If that cell is congested, you need some mechanism to prevent all the bandwidth being consumed by a few users. When I'm in a setting like a railway station with a thousand people all wanting some bandwidth, it makes me angry to see the people streaming video on their phone like their entertainment is more important than everyone else who wants their email and IM to work.
What we have now does not really work ; it's a pricing mechanism of deterring excessive use, but it doesn't assign prices correctly, it just bills your bandwidth use at a flat rate. Bandwidth that is uncontested should be cheap, or free. Bandwidth that is contested should cost more. What really needs to happen in these circumstances is a kind of Dutch auction, where every handset has a notion of what price it's prepared to pay for bandwidth, and the cell sets the price according to how congested it is. But this would be too complicated to sell to most people.
Unmetered wireless bandwidth is not fair, because it's only "too cheap to meter" when the cell has very few clients. Part of the drive for 5G is to push things to this point - 4G too (or else why would they provide such tiny caps with such huge bandwidth that you can burn through it all in minutes - it's not about giving your more bandwidth, it's about having enough bandwidth to serve more customers). As long as we keep up thinking of new media that consume more bandwidth, wireless will not be good enough to serve crowded cells.
Not quite : public domain is not entirely open source and doesn't have quite the same list of positive points.
Not least, public domain works can be taken and used without attribution. If you're a software developer in a gig economy, your reputation matters.
And as you qualify : public domain in the USA.
And while the "PERMISSIVE LICENSING!!!" crowd will jump on me for this, there are advantages to copyleft licensing that you can't have if the work is in public domain, because these licenses require copyright to have force.
For example, software designed for interoperability is probably best licensed GPL or LGPL (to allow closed software to link to it), because this prevents one company making it's own "extensions" to it, thus rendering everyone else's implementation inferior. If you were a large player you could then exploit network effects to make your product the de-facto standard, eliminating the benefits of having interoperability in the first place. Copyleft licensing ensures that such improvements are available to everyone.
Yeah, my no.1 peeve with the UI - having panels you HAVE to visit to play the game being to the left and right of your cockpit window, accessible by keys - becomes just another bit of immersion with a headtracker on. The most natural thing in the world to turn your head to look at your target panel.
I also get what all those flight sims were doing with all those keys to "padlock view" your targets, etc. I never got on well with them and usually just stuck to a straight ahead view. Head tracking is just seamless and immersive compared to that.
It's very good ; the magnetometer counters the drift from the gyroscope in the accelerometer board.
I think it's main advantage over the webcam is that it consumes minimal CPU and has virtually no lag.
The downside is you have something with a wire on your head.
It's an EDTracker (invented by Elite : Dangerous fans for their favourite game...)
http://edtracker.org.uk/
My first wife earned far more than I did for many years and the tables only turned recently (and not by much).
We parted amicably, our only common financial interests were the house (we split evenly because while she earned more, I was more frugal and paid substantial chunks of the mortgage off), and our daughter.
My current GF earns about a quarter what I do but still insists on paying her way equally, which works fine because we both have pretty modest lifestyle demands.
The only kinds of women who are expensive are the ones who are into you for the money.