The $1000 item vending machine of the future: Apple.
In 10,000 years, aliens visit a strangely composite rock. Tiny intricate compartments of hydrocarbons and rare earth metals cover the surface almost entirely, the rectangular compartments are in such a formation as to make it infeasibly difficult to usefully recover any of the materials short of throwing them into a star - the alien archaeologists will find that the extinct inhabitants converted 99% of the useful earth material of the planet into these strange little formations - they drew the only possible conclusion: the inhabitants must have had blindingly strong religious connections to these small rectangles to have dedicated all resources and life on their planet into irreversibly manufacturing these small objects before quickly burying them in the ground.
Let me guess: the fucking page was trying to push 1MB of HTML, 2MB of CSS, 5MB of javascript and 10MB of images for each page hit?
That's a pretty fucking terrible guess... anyone who bothered to follow the link would see one of the most spartan sites on the web - then it's just a button (12) and a mouse click (network tab) away to find out that it's a whopping 67 KiB... so no, your completely and utterly wrong.
Also serving static content is almost never the bottleneck in any website, pipes are enormous these days and static content is the easiest think in the world to distribute among a cluster.
...Analysis has already shown a large number of signatures from outside the UK, including North Korea and Russia.
Last time I checked there were shown to be 25 from Russia... The petition site is far from infallible but this is one of the most signed petitions since the sites inception, and clearly the one with the highest signing rate - Aren't you missing the point.
the AC is on Max cold, the radio is as loud as it goes, and then BAM you are wide awake because you've turned the trunk of a Corolla into an accordion
You were aware enough to try to use loud noises and cold air and you didn't think to pull over... do you think external stimulus replace sleep? was it not obvious to you that your decision to do those things instead of stopping was going to increase your chances of uncontrollably falling asleep?
I think it's far more likely she consciously made a decision to rest her eyes while yielding to the autopilot. Tesla's autopilot is not fully autonomous... in which case anything less than pulling over to get some sleep or letting someone else drive is irresponsible.
How is this trolling? They are advertising a supposed feature of their phone, and supposing their competitors lack that feature then it's automatically trolling? I really hate Apple and think they are evil, but my view is not so warped as to think advertising advantages is trolling.
We actually don't call them stores any more, we call them town squares.
Heh, fuck, I could not better this in trying make a cringe-worthy capitalistic statement - It's so perfect I almost suspect it's astroturfing - Yes Apple! please do come and run our town, I for one welcome our new technological overlords.
Because Apple like to obsolete hardware quickly both officially and by making old hardware run slow on their latest OS version while EOLing older versions of their OS... I have a 10 year old MBP that is decent Linux machine, it doesn't matter that it's Linux specifically, it matters that you have the freedom to continue to boot other OS on your hardware... Apple have recently come to the conclusion that it is not your hardware, but it's theirs, even if you pay them.
This is officially goodbye Apple, you truly disrespect your users.
because the claims that people suffered damage were not supported by the facts advanced by the campaign group
Well of course... damage caused by violating certain human rights such as privacy tend to be intangible and not even possible to directly draw causal relationships between, just because you can't easily quantify it in monetary terms doesn't make it invalid or worthless.
Moore's law doesn't even take into account the physical size of the IC. So even if transistors stayed the same, simply throwing double the cores at the problem and making the chip twice as large still is very much Moore's original observation which had everything to do with manufacturing cost when he defined his law.
You are completely ignoring the context it was observed in... in an era where shrinking process size was the driving force in doubling transistor count - a phenomenon that improved a processor threefold: increases transistor count, switching speed, reduces local propagation delay, increases power efficiency and most importantly had a reasonably sustainable future to _continue_ improving until it hit the fundamental limitations of the materials used.
Now look at the remaining ways transistor doubling happens without shrinking the process size: increasing die area? that only increases transistor count, it does nothing for switching speed, propagation delay, power efficiency and introduces logistic problems, thermal problems and most importantly has little future (how sustainably can you keep increasing die area before we go back to room sized computers - except with out current transistor density those room sized computers would be consuming trillions more watts than their predecessors).
In short: your arguments seems to be that moore's observation is alive because transistor count still increasing in spite of it happening in an entirely different way at an slower rate, with almost none of the benefits that the original phenomenon endowed on processors with no sustainable way of continuing to increase transistor count in the foreseeable future.
Moore's law is informal observation, not a well defined law, if you interpret it as a pedant you will miss the point - look at it from the authors perspective.
The Chinese manufacturer had replace glycerine with propylene glycol to save money. Lots of children died.
I'm no toxicologist but I think you must mean "Diethylene glycol" not "Propylene Glycol"... if you look up the later on wikipedia in the human safety section [1] it states:
The acute oral toxicity of propylene glycol (E1520) is very low, and large quantities are required to cause perceptible health damage in humans
Where as Diethylene glycol (which is in the paper you reference at the very start of the toxological analysis section) and the wikipedia article [2] suggests it has high toxicity (albeit only empirically due to involvement in mass poisonings.):
Despite the discovery of DEG’s toxicity in 1937 and its involvement in mass poisonings around the world, the information available regarding human toxicity is limited. Some authors suggest the minimum toxic dose is estimated at 0.14 mg/kg of body weight and the lethal dose is between 1.0 and 1.63 g/kg of body weight...
[1] https://en.wikipedia.org/wiki/...
[2] https://en.wikipedia.org/wiki/...
Anyway it's nasty stuff... however it should be noted that most of these types of events on the Asian continent are more due to lack of strict regulation on food and medicine than malice. Fake medicine is a real problem over there due to the distribution channels, people but stuff in shops with no way to know how authentic it is... and we all know how good the Chinese are at making rip-offs, unfortunately when you swap out expensive components of a medicine without really knowing what you are doing the difference is death rather than a short lived knock-off.
...the only part of Moore's law is counting transistors.
Moores law was as observation stated in the 60s... what was implicitly meant is dennard scaling, which was not explicitly defined until the 70s. Doubling _was_ a side effect of halfing the transistor size and being able to fit double the transistors on the same die, but you get two things from this:
1. More transistor logic possibilties.
2. faster clock
Number 2 is obviously what everyone used to fixate on because it's an easy to quantify, single dimmensional number that you can slap on the front of boxes and don't have to explain to anyone.
I mean it's not like you confused it with single core performance right? RIGHT?
It pretty much used to mean that, and largely still does when interpretted as intended. We are reduced to improving single core speed through moving logic around only - the only architectural agnostic things to quantify here are tricks of computational reduction and speculation, and not everything _can_ be computationally reduced - many operations are just too fundamental that out of a larger context they are always going to be limmited to propagation delay which has stayed still since 2006.
but a whole lot of the problem is state-funded trolls and people's own hatefulness, idiocy, and credulity being amplified by the echo-chamber effect from self-grouping by disconnecting from conflicting views.
Those are problems, centralised or not, but they are not new or significantly more problematic than the new phenomena that you casually brushed aside:
Sure, there's some targeted ads in there that might not exist in a decentralized system
The problem is those ads have been re-purposed, they are now used to sell political views in the most underhanded subliminal way possible... no echo chamber required, most users don't invest enough time for that sort of engagement anyway, most users consume, which is why it's so much more effective to buy up one peoples major personal portals to the web and distort it into an alternate reality filled with bias appealing to the values of each group - it's not even that accurate, but it doesn't matter, they can reach everyone, so even a 20% success rate is really good, unlike "echo chambers" which only reach a small fraction of people.
As far as I can see, those problems would only be worse if social media were decentralised, as it's very unlikely you could shut someone down even if you knew for certain they were part of a deceitful manipulation campaign - decentralisation protects against censorship, for better *and* worse.
Yes, anonymity and freedom of speech has both negatives an positives, however the difference in the potential negatives of exploiting them in a reasonably designed decentralised network are that they occur on a level playing field, user data is siloed by design. It's not a silver bullet, nothing is, but the difference is it would take hard work and real networking effort to gather and exploit information, not merely a big wallet or a certain CEO under your thumb.
A network of networks cannot be centrally controlled because it has no centre. This is not some accidental design choice we could alter
Irony is the "negative uses" in recent past are all due to _centralisation_, which is entirely the product of the commercialisation of the internet and not it's _design_... assuming "negative uses" is referring to targeted manipulation on facebook, but also widespread censorship particularly China, Egypt etc which is due to ISPs + government being able to pressure or outright control such entities.
Your analogy doesn't really make any sense unless a surface grinder performs a redundant job among a collection of tools and is more trouble than it's worth. This is not about engineering it's closer to the psychology of UX in language design.
It encourages inheritance, which fundamentally breaks the box-model (a.k.a. the black-box-model) of programming.
Rather than breaking things into discrete, understandable chunks, it encourages massive classes that must be understood in their entirety.
I don't know any C++ (and I don't like the sound of it), but I see this same problem in JavaScript albeit in the form of prototypes (Prototypes can be used for pretty much the same purpose, and usually are, although since ES6 there are not "proper" classes).
I haven't travelled wide and far in the programming space, but I have a suspicion that people give too much credit to the "class" pattern because many languages have adopted it into it's syntax, I myself was like this for a while when younger, using classes for everything is easy - until it makes a mess that even the author can't track, now I actively try to avoid it, and find that it's completely inappropriate 95% of the time, for the other 5% things like vectors etc it's a perfect fit, and that's fine. So I suppose in summary it's people using the wrong tool for the wrong job too often, C++ sounds like it provides too many tools, and so it's inevitably going to breed bad code as people are more likely to missus them.
Slashdot Mirror
The $1000 item vending machine of the future: Apple.
In 10,000 years, aliens visit a strangely composite rock. Tiny intricate compartments of hydrocarbons and rare earth metals cover the surface almost entirely, the rectangular compartments are in such a formation as to make it infeasibly difficult to usefully recover any of the materials short of throwing them into a star - the alien archaeologists will find that the extinct inhabitants converted 99% of the useful earth material of the planet into these strange little formations - they drew the only possible conclusion: the inhabitants must have had blindingly strong religious connections to these small rectangles to have dedicated all resources and life on their planet into irreversibly manufacturing these small objects before quickly burying them in the ground.
Let me guess: the fucking page was trying to push 1MB of HTML, 2MB of CSS, 5MB of javascript and 10MB of images for each page hit?
That's a pretty fucking terrible guess... anyone who bothered to follow the link would see one of the most spartan sites on the web - then it's just a button (12) and a mouse click (network tab) away to find out that it's a whopping 67 KiB... so no, your completely and utterly wrong. Also serving static content is almost never the bottleneck in any website, pipes are enormous these days and static content is the easiest think in the world to distribute among a cluster.
...Analysis has already shown a large number of signatures from outside the UK, including North Korea and Russia.
Last time I checked there were shown to be 25 from Russia... The petition site is far from infallible but this is one of the most signed petitions since the sites inception, and clearly the one with the highest signing rate - Aren't you missing the point.
the AC is on Max cold, the radio is as loud as it goes, and then BAM you are wide awake because you've turned the trunk of a Corolla into an accordion
You were aware enough to try to use loud noises and cold air and you didn't think to pull over... do you think external stimulus replace sleep? was it not obvious to you that your decision to do those things instead of stopping was going to increase your chances of uncontrollably falling asleep?
Better that than dozing off without autopilot.
Do you really think she "dozed off"?
I think it's far more likely she consciously made a decision to rest her eyes while yielding to the autopilot. Tesla's autopilot is not fully autonomous... in which case anything less than pulling over to get some sleep or letting someone else drive is irresponsible.
Boghossian's behavior "raises ethical issues of concern."
That was the point you fucking morons.
How is this trolling? They are advertising a supposed feature of their phone, and supposing their competitors lack that feature then it's automatically trolling? I really hate Apple and think they are evil, but my view is not so warped as to think advertising advantages is trolling.
Able to resist being covered in increasingly caustic pesticides = bad modification.
If they don't harm humans or the environment, is that a bad modification?
Really?.. Yes... because it encourages the use of caustic pesticides...
Now I need to come up with some kind of quip orthogonal to "captain fucking obvious"
We actually don't call them stores any more, we call them town squares.
Heh, fuck, I could not better this in trying make a cringe-worthy capitalistic statement - It's so perfect I almost suspect it's astroturfing - Yes Apple! please do come and run our town, I for one welcome our new technological overlords.
Because Apple like to obsolete hardware quickly both officially and by making old hardware run slow on their latest OS version while EOLing older versions of their OS... I have a 10 year old MBP that is decent Linux machine, it doesn't matter that it's Linux specifically, it matters that you have the freedom to continue to boot other OS on your hardware... Apple have recently come to the conclusion that it is not your hardware, but it's theirs, even if you pay them.
This is officially goodbye Apple, you truly disrespect your users.
because the claims that people suffered damage were not supported by the facts advanced by the campaign group
Well of course... damage caused by violating certain human rights such as privacy tend to be intangible and not even possible to directly draw causal relationships between, just because you can't easily quantify it in monetary terms doesn't make it invalid or worthless.
Moore's law doesn't even take into account the physical size of the IC. So even if transistors stayed the same, simply throwing double the cores at the problem and making the chip twice as large still is very much Moore's original observation which had everything to do with manufacturing cost when he defined his law.
You are completely ignoring the context it was observed in... in an era where shrinking process size was the driving force in doubling transistor count - a phenomenon that improved a processor threefold: increases transistor count, switching speed, reduces local propagation delay, increases power efficiency and most importantly had a reasonably sustainable future to _continue_ improving until it hit the fundamental limitations of the materials used.
Now look at the remaining ways transistor doubling happens without shrinking the process size: increasing die area? that only increases transistor count, it does nothing for switching speed, propagation delay, power efficiency and introduces logistic problems, thermal problems and most importantly has little future (how sustainably can you keep increasing die area before we go back to room sized computers - except with out current transistor density those room sized computers would be consuming trillions more watts than their predecessors).
In short: your arguments seems to be that moore's observation is alive because transistor count still increasing in spite of it happening in an entirely different way at an slower rate, with almost none of the benefits that the original phenomenon endowed on processors with no sustainable way of continuing to increase transistor count in the foreseeable future.
Moore's law is informal observation, not a well defined law, if you interpret it as a pedant you will miss the point - look at it from the authors perspective.
The Chinese manufacturer had replace glycerine with propylene glycol to save money. Lots of children died.
I'm no toxicologist but I think you must mean "Diethylene glycol" not "Propylene Glycol"... if you look up the later on wikipedia in the human safety section [1] it states:
The acute oral toxicity of propylene glycol (E1520) is very low, and large quantities are required to cause perceptible health damage in humans
Where as Diethylene glycol (which is in the paper you reference at the very start of the toxological analysis section) and the wikipedia article [2] suggests it has high toxicity (albeit only empirically due to involvement in mass poisonings.):
Despite the discovery of DEG’s toxicity in 1937 and its involvement in mass poisonings around the world, the information available regarding human toxicity is limited. Some authors suggest the minimum toxic dose is estimated at 0.14 mg/kg of body weight and the lethal dose is between 1.0 and 1.63 g/kg of body weight...
[1] https://en.wikipedia.org/wiki/... [2] https://en.wikipedia.org/wiki/... Anyway it's nasty stuff... however it should be noted that most of these types of events on the Asian continent are more due to lack of strict regulation on food and medicine than malice. Fake medicine is a real problem over there due to the distribution channels, people but stuff in shops with no way to know how authentic it is... and we all know how good the Chinese are at making rip-offs, unfortunately when you swap out expensive components of a medicine without really knowing what you are doing the difference is death rather than a short lived knock-off.
...the only part of Moore's law is counting transistors.
Moores law was as observation stated in the 60s... what was implicitly meant is dennard scaling, which was not explicitly defined until the 70s. Doubling _was_ a side effect of halfing the transistor size and being able to fit double the transistors on the same die, but you get two things from this:
1. More transistor logic possibilties.
2. faster clock
Number 2 is obviously what everyone used to fixate on because it's an easy to quantify, single dimmensional number that you can slap on the front of boxes and don't have to explain to anyone.
I mean it's not like you confused it with single core performance right? RIGHT?
It pretty much used to mean that, and largely still does when interpretted as intended. We are reduced to improving single core speed through moving logic around only - the only architectural agnostic things to quantify here are tricks of computational reduction and speculation, and not everything _can_ be computationally reduced - many operations are just too fundamental that out of a larger context they are always going to be limmited to propagation delay which has stayed still since 2006.
... When doused in petrol and lit on fire, all computers BURN. Thanks captain fucking obvious.
https://www.youtube.com/watch?...
but a whole lot of the problem is state-funded trolls and people's own hatefulness, idiocy, and credulity being amplified by the echo-chamber effect from self-grouping by disconnecting from conflicting views.
Those are problems, centralised or not, but they are not new or significantly more problematic than the new phenomena that you casually brushed aside:
Sure, there's some targeted ads in there that might not exist in a decentralized system
The problem is those ads have been re-purposed, they are now used to sell political views in the most underhanded subliminal way possible... no echo chamber required, most users don't invest enough time for that sort of engagement anyway, most users consume, which is why it's so much more effective to buy up one peoples major personal portals to the web and distort it into an alternate reality filled with bias appealing to the values of each group - it's not even that accurate, but it doesn't matter, they can reach everyone, so even a 20% success rate is really good, unlike "echo chambers" which only reach a small fraction of people.
As far as I can see, those problems would only be worse if social media were decentralised, as it's very unlikely you could shut someone down even if you knew for certain they were part of a deceitful manipulation campaign - decentralisation protects against censorship, for better *and* worse.
Yes, anonymity and freedom of speech has both negatives an positives, however the difference in the potential negatives of exploiting them in a reasonably designed decentralised network are that they occur on a level playing field, user data is siloed by design. It's not a silver bullet, nothing is, but the difference is it would take hard work and real networking effort to gather and exploit information, not merely a big wallet or a certain CEO under your thumb.
A network of networks cannot be centrally controlled because it has no centre. This is not some accidental design choice we could alter
Irony is the "negative uses" in recent past are all due to _centralisation_, which is entirely the product of the commercialisation of the internet and not it's _design_ ... assuming "negative uses" is referring to targeted manipulation on facebook, but also widespread censorship particularly China, Egypt etc which is due to ISPs + government being able to pressure or outright control such entities.
Making it worthless unless you like super expensive and shit servers that only run one OS.
Your analogy doesn't really make any sense unless a surface grinder performs a redundant job among a collection of tools and is more trouble than it's worth. This is not about engineering it's closer to the psychology of UX in language design.
that was supposed to be "there are now proper classes" not "not"
It encourages inheritance, which fundamentally breaks the box-model (a.k.a. the black-box-model) of programming. Rather than breaking things into discrete, understandable chunks, it encourages massive classes that must be understood in their entirety.
I don't know any C++ (and I don't like the sound of it), but I see this same problem in JavaScript albeit in the form of prototypes (Prototypes can be used for pretty much the same purpose, and usually are, although since ES6 there are not "proper" classes).
I haven't travelled wide and far in the programming space, but I have a suspicion that people give too much credit to the "class" pattern because many languages have adopted it into it's syntax, I myself was like this for a while when younger, using classes for everything is easy - until it makes a mess that even the author can't track, now I actively try to avoid it, and find that it's completely inappropriate 95% of the time, for the other 5% things like vectors etc it's a perfect fit, and that's fine. So I suppose in summary it's people using the wrong tool for the wrong job too often, C++ sounds like it provides too many tools, and so it's inevitably going to breed bad code as people are more likely to missus them.
As the U.K. begins a two-week heat wave
What are you talking about ? it's been two weeks already! and it looks to keep on going.
Delivery is key ;P
or "Low-Flight Flight Demonstrator"
So... it's a Low Flight Flight Demonstrator Demonstration?