Design costs are fixed costs and therefore fungible, whereas manufacturing costs are not.
If you cannot segment the market, and cannot ever change your price (after all, the design effort was identical 10 years from now as it is today...) then your price has to be:
Breakeven U Unit Price = Manufacturing cost + Design cost / Total number of units sold
This will not earn a profit so it's an absolute minimum. I've chosen U for unified.
However, if you can segment into A and B by downgrading some A to B, you can profitably sell B at a lower price than "breakeven" and still profit.
Breakeven A Unit Price = Manufacturing cost + Design cost / Total number of A units sold Breakeven B Unit Price = Manufacturing cost
The high-end market got more expensive, according to the proportion of the market willing to pay for A quality. The low-end market got cheaper. In many cases this is what the market requires to happen.
It's easier to see how this works if you consider a scenario where the performance of unit B is exactly equal to the old unit A, call it A', which has the same manufacturing cost as A and B units, but had a lower design cost.
Breakeven A' Unit Price = Manufacturing cost + Design cost' / Total number of A' units sold Breakeven B' Unit Price = Manufacturing cost
As you can see, the new product B is strictly better than the old products: higher performing than unit B', and the same performance as A' but at a lower price. Product A isn't important except in the fact that it makes product B possible. Therefore it's absolutely clear that the A / B model must replace the A' / B' models in the free market.
Now consider that instead of A / B, we want to use U. Now it is NOT clear that this must replace the A' / B' models, because although U is higher performing than both of these models, the breakeven price of U is higher than the B' price, and depending on the difference in design cost, possibly higher than the A' price too.
Given:
1. A / B > A' / B' 2. ~(U > A' / B') C. ~(U > A / B)
The single-pricing model is not preferred by a rational free market over the two-price model.
The free market perfect competition models that say that a manufacturer should not downgrade performance for no downgrade in price apply in the limit, as sales go to infinity, which does not ever happen. At that point, the marginal design cost goes to zero, and THEN the market sells only unit A. Until then, the market calls for segmentation, and that's only a failure in the sense that capitalism is itself a failure.
This is what the market does. The marginal cost of producing a higher-quality good to a lower one is of only minor relevance to the optimal pricing strategy.
When that marginal cost is 0, near-0, or even very slightly negative, it gets people's knickers in a twist here on slashdot. But focussing on this is like micro-optimising a piece of code without profiling the entire system.
There was no false pretense. They said what it did, and it does that. This announcement doesn't make your purchase any worse. At all. Whatsoever. The guy who bought the chip yesterday has literally lost nothing. In fact, he gained an option to do something, which he can decline without any consequences and be exactly where he was the day before. You're worse off if you buy an iPod the day before Apple announces a price reduction (even one that isn't effective immediately).
No, it's like saying some consumers demand automobiles that will only go 150mph when they could go 300mph. Some people actually care about top speed. Some don't, at all. Some do, but care more about money. Welcome to the world of market segmentation.
The consumers aren't demanding a low-end CPU per se, they are demanding a cheap CPU. Others are demanding a faster CPU and are willing to pay. It's not just two groups either, it's a gradient.
It absolutely would cost Intel to enable the missing features. If their cheaper chips had the same capabilities as their high end chips, they have to cost the same amount of money, which eliminates market segmentation. Market segmentation generally benefits the company and the lowest-end consumers, while extracting more money from the highest-end consumers.
You can choose one price, equilibrium of all supply and demand, and then the people who want a high-end CPU and are willing to pay are winners because they get the same thing they wanted before, for cheaper. The people demanding the cheapest CPUs get no CPU, so they lose. There's a turning point somewhere in the middle between losers and winners.
I don't even understand why they would gain a lot in public relations. The number of people who care about this are few and this would be a one-shot. The manpower to disable the features seems pretty minimal too. Binning yields do mean that there will be a few lower-spec simply because the chip can't do better, which means that process already needs to exist.
Because Yahoo search is Bing search, so Yahoo taking top billing is Bing taking top billing. What surprises me is that they listed them separately, though it is interesting in that it shows a similar but non-identical success rate.
If you will die if you do not use the gun, then shoot to kill.
If you don't want to kill them, kill them anyway. Or accept that you will die.
You seem to think your scenario makes sense. What is this scenario anyway? Your daughter is being mind-controlled by alien forces to attack you with a broadsword? If that ever happens to you, reverse the polarity instead of using a gun. Prepare for *reality*.
Imagining that a gun will help you without killing your target and actually being able to do it when the situation arises are two different things.
How does that nuke his supposition? Google doesn't sell Chrome, it distributes Chrome without charge -- sometimes it even pays organisations to distribute Chrome. I don't know whether it profits from Chromebook sales, so I'll let that point slide for now.
But Google isn't obviously getting anything out of a person using Chrome rather than Firefox. Dig a bit deeper and maybe they can pick up user habits and data for advertising from Chrome, but that's exactly an argument to do what they can to remain Firefox's default search engine.
Google sells ads. As in, they get their money from advertisements, and therefore they care about competition for advertisements. The vast majority of their other activities are in service to their advertising empire. So if it's between their piece of the ad market and their piece of the browser market, their ad marketshare will win, unless there's something else I'm missing or some other card they have yet to play.
I totally admit I had the same initial reaction when I first heard about the "Christian Science Monitor", but a quick examination showed me that actually it's quite reasonable.
Still, your skipping 2 and 3 betrays a different viewpoint. I'm guessing you're American? I don't mean this as an insult, just an observation: there's a strong stereotype of Americans being insular about world events and choosing "bubbles" in which to consume information to a much greater degree than people of many other countries with a free press and the average person has enough wealth to pay attention to the news. I freely admit this is my own biased perspective of it and maybe you have a different spin on the same phenomenon.
As for 1, I didn't know what 3M stood for until just now. I'm sure they changed their name, in part, to avoid these problems.
Yes, that's what positively means. Although also ideally it would change the direction of their life in the way they like -- changing your mind doesn't mean permanent unhappiness.
I don't mean to be glib, but is there some problem with viewing changing kids' lives from neo-nazism as a good thing? It's not like this is a government fiat where they hold a gun to your head and tell you to believe; it's a charity advertising campaign.
I'm unclear what your position is. My guesses:
- A philosophical stance of non-interventionism under any circumstances? That seems to contradict you posting any argument. - Neo-nazism sympathy? - An irrelevant philosophical point that boils down to how morality cannot be derived from provable facts? - Trolling?
They're testing the contrapositive proposition. Just as:
non-human signals that are extremely unlikely to be natural -> extraterrestrial intelligence no extraterrestrial intelligence -> no non-human signals that are extremely unlikely to be natural
The contrapositive is absolutely falsifiable, therefore the experiment is scientific.
Falsifiable prediction: There is no intelligent life on other planets.
Experiment: Listen for radio signals that are patterned in a way that is implausible to occur from natural phenomena, and that therefore most likely came from intelligent actions. Determine if they might be a false positive (human satellites, etc.).
One megabyte for 3 hours is 10800 megabyte-seconds. Half a gigabyte for 8 seconds is 4096 megabyte-seconds, which is less than half of the memory usage time. Assuming those numbers are even close to accurate, he wasn't even really optimising for memory usage unless you discount concurrency, and if you discount concurrency then optimising for memory usage over execution time is absurd, even given old constraints.
You get $10k per target, which substantially exceeds the machine price, so while it's not perfectly objective it's not that far out of whack.
I do find this argument funny because it's essentially identical to the argument "Windows Exploits are more common because so many more people have Windows and therefore it's more rewarding to exploit Windows".
That's not necessarily a fair example. Strcpy isn't safe unless you can guarantee that the size of the destination buffer is greater than the number of characters up to and including the null terminator. Typically that means you have to call strlen anyway (or carry the string length with you as though it were a counted string).
The argument is valid right now because it's arguing against the idea that Apple saying something means that Apple will continue to believe the thing they said today, for all time.
They really did make the one-button mouse long past the time when everybody else had decided that two was a minimum. They really did change their mind on that (and even afterward some people praised the one-button design for some time). Pretty clear-cut: sometimes Apple's decisions change.
Apple didn't release a netbook, but it released a tablet. There's a real difference there and it isn't splitting hairs, but the point is that the statement lead people to conclude that Apple would be sticking with their $1000 notebooks as the low end portable hardware (and that this would be fine for the foreseeable future, and that iPods would not meaningfully converge with their traditional businesses). Likewise, arguing that touch devices are fundamentally different from desktops/laptops doesn't mean we won't find convergence eventually. The mouse is fundamentally different from the (hardware) keyboard, too.
As for PowerPC, the point is they said that, and then they changed their minds. That's partly because of changing circumstances, but let's not pretend that circumstances are never going to change again.
I'm of the opinion that eventually, they'll converge. That, or one will simply die when the other veers into a vastly superior or inferior direction (from the point of view of the marketplace). I'm not counting on it being anytime soon, though, like Wall Street seems to be doing.
One performance advantage of NULL-terminated strings is you can trivially maintain two independent representations of the same string, one of which has a static prefix.
char *str2 = str1 + prefix_length;
Any non-reallocating modification of one string instantly affects the other.
Is that kind of academic? You betcha'. Most of this is. If your performance is seriously gated by strlen, then you should use counted strings for that operation even in C (nobody will stop you).
The point of the article was what the ideal string encoding would be, with the argument that the magic terminator wasn't it. We can think of an infinite number of bad encodings, but I'm pretty sure the article was trying to come up with a good one.
But anyway, the article argued that it would cost one more byte to do address + length. The entire premise of the article, then, is of a 2-byte length. 65k characters ought to be enough for anyone... or in any case, it's not unreasonable to require a library for ultra-massive buffers like that (such a library could also allow it to be non-contiguous, which the rare cases of ultra-long strings can benefit from, memory-wise).
If you see somebody in a bikini and can't tell what sex organs they have, you might have a problem.
Besides, if you saw somebody you thought was attractive in a bikini, what would finding out she had dude parts do to change things? Do you fear that you have a secret bisexual alternate personality? I'm pretty sure the things that make me straight come way before that point.
One, you assume that the program operates only on memory. But it can have external input, which is an effectively infinite space. Polling the network card is an obvious case.
Two, you assume that you can define and measure a number of "steps", which are wholly unique. You could argue that any cause for this is just a special case of the other problem, but it's a real problem for real programs. And this all started by debating that "real computers" don't meet Turing criteria.
Counterexample, in pseudocode, using the internal timer:
int main() {
for(;;)
{
if (GetSecondsElapsedSinceBootAsEightBitUnsigned() != 100)
{
if (GetSecondsElapsedSinceBootAsEightBitUnsigned() != 100) break;
}
}
return 0; }
This is a program that may or may not terminate every 256 seconds, starting at 100 seconds in. It will terminate if and only if the timer "tick" occurs between the two if statements.
You may say that part of the N states the machine might be in includes every possible timing for the second "tick" to occur in relation to the code execution, or perhaps every possible subdivision of the time between "tick"s down to the planck time. In which case I'd argue you're stretching the definition of "state", but even so the point remains that it's possible to repeat the exact same state multiple times before hitting the end state, unless you make the unrealistic assumption that CPU execution time is pristinely perfect with no variance whatsoever for all time. So executing N+1 times has no guarantee of terminating this function.
I don't get why you assume that this system triggers on your oh shit reflex rather than you actually deciding to stop.
Clenching your leg muscles, mentally imagining stopping -- that's what the GP was referring to when he said "oh shit".
I mean, maybe it's hard to detect the right response but in principle it should be perfectly possible to get no false positives (other than the case where you totally wanted to hit the brakes but it turns out that your feet were obstructed, eg. a child had grabbed your shoelace).
I doubt it. I'd imagine that they pick up the signal once it was too late to change your mind. The article mentions the point of new return several times. At that point, you're pressing the brake; there's no stopping that and changing your mind.
He's not castigating people for taking these jobs, nor is he even saying that those jobs should just vanish into thin air. He's pointing out a societal problem that leads people to making these choices. Basically, he's agreeing with everything you said about researching the cure for cancer paying less than HFT, and calling that the problem. Surely you can imagine a society where curing cancer is paid like HFT and HFT is paid like curing cancer, even if you can't imagine how to get there from here. The only thing you said that countered the GP rather than reinforcing it was the argument that HFT applies upward salary pressure.
The narrative where people should not be allowed to make their own choices in favour of your choices, is yours alone. At best, it's mistaken, irrational hyperbole. At worst, it is a battle-cry against change and progress and any analysis whatsoever about the status quo.
Slashdot Mirror
Design costs are fixed costs and therefore fungible, whereas manufacturing costs are not.
If you cannot segment the market, and cannot ever change your price (after all, the design effort was identical 10 years from now as it is today...) then your price has to be:
Breakeven U Unit Price = Manufacturing cost + Design cost / Total number of units sold
This will not earn a profit so it's an absolute minimum. I've chosen U for unified.
However, if you can segment into A and B by downgrading some A to B, you can profitably sell B at a lower price than "breakeven" and still profit.
Breakeven A Unit Price = Manufacturing cost + Design cost / Total number of A units sold
Breakeven B Unit Price = Manufacturing cost
The high-end market got more expensive, according to the proportion of the market willing to pay for A quality. The low-end market got cheaper. In many cases this is what the market requires to happen.
It's easier to see how this works if you consider a scenario where the performance of unit B is exactly equal to the old unit A, call it A', which has the same manufacturing cost as A and B units, but had a lower design cost.
Breakeven A' Unit Price = Manufacturing cost + Design cost' / Total number of A' units sold
Breakeven B' Unit Price = Manufacturing cost
As you can see, the new product B is strictly better than the old products: higher performing than unit B', and the same performance as A' but at a lower price. Product A isn't important except in the fact that it makes product B possible. Therefore it's absolutely clear that the A / B model must replace the A' / B' models in the free market.
Now consider that instead of A / B, we want to use U. Now it is NOT clear that this must replace the A' / B' models, because although U is higher performing than both of these models, the breakeven price of U is higher than the B' price, and depending on the difference in design cost, possibly higher than the A' price too.
Given:
1. A / B > A' / B'
2. ~(U > A' / B')
C. ~(U > A / B)
The single-pricing model is not preferred by a rational free market over the two-price model.
The free market perfect competition models that say that a manufacturer should not downgrade performance for no downgrade in price apply in the limit, as sales go to infinity, which does not ever happen. At that point, the marginal design cost goes to zero, and THEN the market sells only unit A. Until then, the market calls for segmentation, and that's only a failure in the sense that capitalism is itself a failure.
This is what the market does. The marginal cost of producing a higher-quality good to a lower one is of only minor relevance to the optimal pricing strategy.
When that marginal cost is 0, near-0, or even very slightly negative, it gets people's knickers in a twist here on slashdot. But focussing on this is like micro-optimising a piece of code without profiling the entire system.
There was no false pretense. They said what it did, and it does that. This announcement doesn't make your purchase any worse. At all. Whatsoever. The guy who bought the chip yesterday has literally lost nothing. In fact, he gained an option to do something, which he can decline without any consequences and be exactly where he was the day before. You're worse off if you buy an iPod the day before Apple announces a price reduction (even one that isn't effective immediately).
No, it's like saying some consumers demand automobiles that will only go 150mph when they could go 300mph. Some people actually care about top speed. Some don't, at all. Some do, but care more about money. Welcome to the world of market segmentation.
The consumers aren't demanding a low-end CPU per se, they are demanding a cheap CPU. Others are demanding a faster CPU and are willing to pay. It's not just two groups either, it's a gradient.
It absolutely would cost Intel to enable the missing features. If their cheaper chips had the same capabilities as their high end chips, they have to cost the same amount of money, which eliminates market segmentation. Market segmentation generally benefits the company and the lowest-end consumers, while extracting more money from the highest-end consumers.
You can choose one price, equilibrium of all supply and demand, and then the people who want a high-end CPU and are willing to pay are winners because they get the same thing they wanted before, for cheaper. The people demanding the cheapest CPUs get no CPU, so they lose. There's a turning point somewhere in the middle between losers and winners.
I don't even understand why they would gain a lot in public relations. The number of people who care about this are few and this would be a one-shot. The manpower to disable the features seems pretty minimal too. Binning yields do mean that there will be a few lower-spec simply because the chip can't do better, which means that process already needs to exist.
Because Yahoo search is Bing search, so Yahoo taking top billing is Bing taking top billing. What surprises me is that they listed them separately, though it is interesting in that it shows a similar but non-identical success rate.
If you will die if you do not use the gun, then shoot to kill.
If you don't want to kill them, kill them anyway. Or accept that you will die.
You seem to think your scenario makes sense. What is this scenario anyway? Your daughter is being mind-controlled by alien forces to attack you with a broadsword? If that ever happens to you, reverse the polarity instead of using a gun. Prepare for *reality*.
Imagining that a gun will help you without killing your target and actually being able to do it when the situation arises are two different things.
How does that nuke his supposition? Google doesn't sell Chrome, it distributes Chrome without charge -- sometimes it even pays organisations to distribute Chrome. I don't know whether it profits from Chromebook sales, so I'll let that point slide for now.
But Google isn't obviously getting anything out of a person using Chrome rather than Firefox. Dig a bit deeper and maybe they can pick up user habits and data for advertising from Chrome, but that's exactly an argument to do what they can to remain Firefox's default search engine.
Google sells ads. As in, they get their money from advertisements, and therefore they care about competition for advertisements. The vast majority of their other activities are in service to their advertising empire. So if it's between their piece of the ad market and their piece of the browser market, their ad marketshare will win, unless there's something else I'm missing or some other card they have yet to play.
"is about to" is not an "if" word, it's a "when" word, with "when" = "very soon".
I totally admit I had the same initial reaction when I first heard about the "Christian Science Monitor", but a quick examination showed me that actually it's quite reasonable.
Still, your skipping 2 and 3 betrays a different viewpoint. I'm guessing you're American? I don't mean this as an insult, just an observation: there's a strong stereotype of Americans being insular about world events and choosing "bubbles" in which to consume information to a much greater degree than people of many other countries with a free press and the average person has enough wealth to pay attention to the news. I freely admit this is my own biased perspective of it and maybe you have a different spin on the same phenomenon.
As for 1, I didn't know what 3M stood for until just now. I'm sure they changed their name, in part, to avoid these problems.
Yes, that's what positively means. Although also ideally it would change the direction of their life in the way they like -- changing your mind doesn't mean permanent unhappiness.
I don't mean to be glib, but is there some problem with viewing changing kids' lives from neo-nazism as a good thing? It's not like this is a government fiat where they hold a gun to your head and tell you to believe; it's a charity advertising campaign.
I'm unclear what your position is. My guesses:
- A philosophical stance of non-interventionism under any circumstances? That seems to contradict you posting any argument.
- Neo-nazism sympathy?
- An irrelevant philosophical point that boils down to how morality cannot be derived from provable facts?
- Trolling?
They're testing the contrapositive proposition. Just as:
non-human signals that are extremely unlikely to be natural -> extraterrestrial intelligence
no extraterrestrial intelligence -> no non-human signals that are extremely unlikely to be natural
The contrapositive is absolutely falsifiable, therefore the experiment is scientific.
Falsifiable prediction: There is no intelligent life on other planets.
Experiment: Listen for radio signals that are patterned in a way that is implausible to occur from natural phenomena, and that therefore most likely came from intelligent actions. Determine if they might be a false positive (human satellites, etc.).
One megabyte for 3 hours is 10800 megabyte-seconds. Half a gigabyte for 8 seconds is 4096 megabyte-seconds, which is less than half of the memory usage time. Assuming those numbers are even close to accurate, he wasn't even really optimising for memory usage unless you discount concurrency, and if you discount concurrency then optimising for memory usage over execution time is absurd, even given old constraints.
You get $10k per target, which substantially exceeds the machine price, so while it's not perfectly objective it's not that far out of whack.
I do find this argument funny because it's essentially identical to the argument "Windows Exploits are more common because so many more people have Windows and therefore it's more rewarding to exploit Windows".
That's not necessarily a fair example. Strcpy isn't safe unless you can guarantee that the size of the destination buffer is greater than the number of characters up to and including the null terminator. Typically that means you have to call strlen anyway (or carry the string length with you as though it were a counted string).
The argument is valid right now because it's arguing against the idea that Apple saying something means that Apple will continue to believe the thing they said today, for all time.
They really did make the one-button mouse long past the time when everybody else had decided that two was a minimum. They really did change their mind on that (and even afterward some people praised the one-button design for some time). Pretty clear-cut: sometimes Apple's decisions change.
Apple didn't release a netbook, but it released a tablet. There's a real difference there and it isn't splitting hairs, but the point is that the statement lead people to conclude that Apple would be sticking with their $1000 notebooks as the low end portable hardware (and that this would be fine for the foreseeable future, and that iPods would not meaningfully converge with their traditional businesses). Likewise, arguing that touch devices are fundamentally different from desktops/laptops doesn't mean we won't find convergence eventually. The mouse is fundamentally different from the (hardware) keyboard, too.
As for PowerPC, the point is they said that, and then they changed their minds. That's partly because of changing circumstances, but let's not pretend that circumstances are never going to change again.
I'm of the opinion that eventually, they'll converge. That, or one will simply die when the other veers into a vastly superior or inferior direction (from the point of view of the marketplace). I'm not counting on it being anytime soon, though, like Wall Street seems to be doing.
Could be that he was familiar with both Gnome and Windows in the first place.
Every time? No.
One performance advantage of NULL-terminated strings is you can trivially maintain two independent representations of the same string, one of which has a static prefix.
char *str2 = str1 + prefix_length;
Any non-reallocating modification of one string instantly affects the other.
Is that kind of academic? You betcha'. Most of this is. If your performance is seriously gated by strlen, then you should use counted strings for that operation even in C (nobody will stop you).
The point of the article was what the ideal string encoding would be, with the argument that the magic terminator wasn't it. We can think of an infinite number of bad encodings, but I'm pretty sure the article was trying to come up with a good one.
But anyway, the article argued that it would cost one more byte to do address + length. The entire premise of the article, then, is of a 2-byte length. 65k characters ought to be enough for anyone... or in any case, it's not unreasonable to require a library for ultra-massive buffers like that (such a library could also allow it to be non-contiguous, which the rare cases of ultra-long strings can benefit from, memory-wise).
If you see somebody in a bikini and can't tell what sex organs they have, you might have a problem.
Besides, if you saw somebody you thought was attractive in a bikini, what would finding out she had dude parts do to change things? Do you fear that you have a secret bisexual alternate personality? I'm pretty sure the things that make me straight come way before that point.
Crap, sorry, the first if was supposed to be == and the second if !=. I think that should be clear from context.
You have two problems:
One, you assume that the program operates only on memory. But it can have external input, which is an effectively infinite space. Polling the network card is an obvious case.
Two, you assume that you can define and measure a number of "steps", which are wholly unique. You could argue that any cause for this is just a special case of the other problem, but it's a real problem for real programs. And this all started by debating that "real computers" don't meet Turing criteria.
Counterexample, in pseudocode, using the internal timer:
int main()
{
for(;;)
{
if (GetSecondsElapsedSinceBootAsEightBitUnsigned() != 100)
{
if (GetSecondsElapsedSinceBootAsEightBitUnsigned() != 100) break;
}
}
return 0;
}
This is a program that may or may not terminate every 256 seconds, starting at 100 seconds in. It will terminate if and only if the timer "tick" occurs between the two if statements.
You may say that part of the N states the machine might be in includes every possible timing for the second "tick" to occur in relation to the code execution, or perhaps every possible subdivision of the time between "tick"s down to the planck time. In which case I'd argue you're stretching the definition of "state", but even so the point remains that it's possible to repeat the exact same state multiple times before hitting the end state, unless you make the unrealistic assumption that CPU execution time is pristinely perfect with no variance whatsoever for all time. So executing N+1 times has no guarantee of terminating this function.
I don't get why you assume that this system triggers on your oh shit reflex rather than you actually deciding to stop.
Clenching your leg muscles, mentally imagining stopping -- that's what the GP was referring to when he said "oh shit".
I mean, maybe it's hard to detect the right response but in principle it should be perfectly possible to get no false positives (other than the case where you totally wanted to hit the brakes but it turns out that your feet were obstructed, eg. a child had grabbed your shoelace).
I doubt it. I'd imagine that they pick up the signal once it was too late to change your mind. The article mentions the point of new return several times. At that point, you're pressing the brake; there's no stopping that and changing your mind.
There's also the physical time involved in moving your foot left and then pressing the brake. I expect that's dominant.
Additionally, your reference gives a range of thought signal speeds.
You're not paying any attention to what he said.
He's not castigating people for taking these jobs, nor is he even saying that those jobs should just vanish into thin air. He's pointing out a societal problem that leads people to making these choices. Basically, he's agreeing with everything you said about researching the cure for cancer paying less than HFT, and calling that the problem. Surely you can imagine a society where curing cancer is paid like HFT and HFT is paid like curing cancer, even if you can't imagine how to get there from here. The only thing you said that countered the GP rather than reinforcing it was the argument that HFT applies upward salary pressure.
The narrative where people should not be allowed to make their own choices in favour of your choices, is yours alone. At best, it's mistaken, irrational hyperbole. At worst, it is a battle-cry against change and progress and any analysis whatsoever about the status quo.