I think the larger trend is that lectures are moving to video, and class time is thereby freed up for Q&A (both directions, so some "Socratic" from the professor), discussion, and various other interactive things. The best part about video is that if you can follow it at 1.0x speed you can probably follow it at 1.6x speed, and when you can't follow it there is the rewind button.
Unless you mean the publishers, they don't even today. You work for free for the journal and give them exclusive access, in return for a shot at publicity/"impact".
Note quite, a shot at tenure. "Publish or Perish". Everyone in academia knew this going in.
Even if Bitcoin does raise the block size one day, it's still not clear to me how it's ever going to scale high enough to become a general purpose currency. Going from 7 transactions/second to 56 sounds nice and all, but Visa and Mastercard handle about 2000 transactions/second. Each.
Are they going to eventually go to a 600+ MB block size? Or are a sizeable number of transactions going to have to go through bank-like intermediaries?
As I wrote before: "may adopt a larger block size **or some other remedy** at any time in the future".
My point is that Classic Bitcoin is not going to sit idly by and say: well, the old 1.0 version of our design can't handle our upcoming performance requirements. Lets just pack things up and call it a failure. No, they can implement whatever changes are necessary, block size or otherwise.
Actually, it does not matter. Bitcoin Cash is redundant. They may be correct as to the technical merits of why a change is needed regarding maximum block size but their fork is unnecessary. Ordinary Bitcoin may adopt a larger block size or some other remedy at any time in the future. The argument is only over whether something needs to be done today. Well that and internal developer politics.
Regular Bitcoin is not locked into some doomed course. They can make a block size change on the timeframe they think appropriate; or if Bitcoin Cash does enjoy increasing popularity they can make the change to put an end to the defections to the other side. Either way Bitcoin Cash seems doomed. Although I'm sure some speculators will find a way to make money off the hype.
Fanboys have a habit of twisting a disadvantage into a full advantage. Much like how back in the Power PC days. Apple use to show how the Power PC processor had handled a few Photoshop filters better than Intel chips, while the Intel Chips were in general faster overall.
As someone who did a lot of Intel and PowerPC programming back then, including optimized PowerPC and x86 assembly. PowerPC was technically faster even for general code. About 20% for CPUs of the same clock rate. Although I had many years more experience with x86 assembly, PowerPC also provided unique optimization opportunities that helped specialized code. Where Intel won overall was not having the better performing hardware (again, comparing CPUs with the same clock rate), rather they had a higher clock rate to overcome the inefficiencies of their architecture. "In theory" PowerPC had the advantage of a clean RISC design versus the legacy CISC of x86. CISC is much harder to design with and improve. However Intel overcame this by throwing tons of money into the project and frankly pulled off miracles. Looking back it was not that PowerPC failed to improve over time, its that no one (outside of Intel) expected x86 could reach the performance levels it has. Intel's engineers are miracle workers and deserve and enormous amount of credit. PowerPC expected Intel x86 to hit the wall many years earlier that it did, and PowerPC failed as a result.
That said, Apple's move to Intel was long overdue. Perhaps they should have never bothered with PowerPC and went from 68K to x86. Apple doubling their marketshare after the switch to Intel was not about slightly faster CPUs running native code. It was about being able to run Windows apps at reasonable speed even under emulation. Once on x86 CPUs they no longer had to emulate the underlying instruction set. Once they had this users no longer had to pick MacOS or Windows, they could have both. Even better, Apple added a dual boot option to run Windows natively when emulation might not be OK.
It's ridiculous to begin with. Comparing a specific phone to a platform? Has there ever been a requirement for Android hardware vendors to only manufacture phones with good cameras?
Apparently you missed most of the summary:
"I would never use an Android phone for photos! Here is the problem: It's Android. Android is an open source (mostly) operating system that has to be neutral to all parties. This sounds good until you get into the details. Ever wonder why a Samsung phone has a confused and bewildering array of photo options? Should I use the Samsung Camera? Or the Android Camera? Samsung gallery or Google Photos? It's because when Samsung innovates with the underlying hardware (like a better camera) they have to convince Google to allow that innovation to be surfaced to other applications via the appropriate API. That can take YEARS. Also the greatest innovation isn't even happening at the hardware level -- it's happening at the computational photography level. (Google was crushing this 5 years ago -- they had had 'auto awesome' that used AI techniques to automatically remove wrinkles, whiten teeth, add vignetting, etc... but recently Google has fallen back). Apple doesn't have all these constraints. They innovate in the underlying hardware, and just simply update the software with their latest innovations (like portrait mode) and ship it."
I can tell you from experience degreed CS individuals also have gaps in their knowledge - so don't go there.
Its not the existence of gaps, its what those specific gaps are. CS individuals are less likely to have gaps in common important areas. For example, many self-taught programmers aspiring to game development crash and burn when they apply at a game studio due to a limited understanding of data structure and algorithms.
Today there are plenty of on-line training opportunities to supplement any self study. So, opportunities to 'self teach' isn't just limited to reading computer magazines.
No one said it is. Anything you are doing on your own outside of coursework or the job is self study. And University students have a lot of time for self study. Their University coursework adds to what they are doing on their own, it doesn't take anything away.
As for dropping into populations - that also applies to people being dropped into a population in a company.
Populations are usually far more diverse and far more accessible at a University than at a company. Experience at a company tends to be more focused, spare time to help with something not work related more limited due to family pressures, etc.
Experience beats training over the long haul.
Absolutely, the point is you will have more experience adding the formal studies than if you forgo them.
Finally - there are autodidacts that have the ability to not only learn material on their own, but also have acquired or developed on their own models that allow that information to be properly integrated and used in a work context.
And the broader your base of knowledge the more likely that is to happen due to cross-pollination. Those uninteresting classes that were forced upon you in the University sometimes turn out to be quite useful on the job years later. Again, formal training is additive, it adds to what you are doing on your own.
The fact that you believe code lives in a vacuum... res ipso loquitor.
Sorry, no one claimed that code lives in a vacuum. All that is claimed is that (1) some things are left as an exercise for the student, OS configuration, integrated development environments, the fashionable programming language of the day, etc. All these are things students are expected to learn as needed on their own time. University time is better spent on things more persistent such as data structures and algorithms. The students can apply the concepts of this topic to whatever language they use, whatever operating system they target, etc. And it is also claimed that (2) gaps in core topics like data structures and algorithms is far more detrimental than gaps in OS configuration for someone in a software development career. For an IT career that would be reversed.
As I said, "more likely". There is always the unlikely exception. But what you offer are poor examples. There is no shortage of self-taught that are stumped by such things. And frankly, those are minor things in a University environment that students would be expected to figure out on their own. Recall my mention of learning as much from peers and independent study as from professors.
The point you are missing is that going the University route is additive, it adds to what you would have done on your own. It doesn't plug all possible gaps, it just avoids the more common and important ones. Gaps with respect to data structures and algorithms is a far greater hazard than configuring a python environment.
Yes, I rember when I first learned what pointers were: "Oh... so that's what they call what I've been doing with indirect indexesd addressing. Pointer arrays." Then I found out about Yourdain & Constantine. Sorry, all that information is available much more so now than then, and it didn't even hold me back then.
And I read Knuth for fun in my drop-out days. Availability is not the issue. The issue is that many self-taught pass on available info that seems uninteresting to them, or in their erroneous belief not relevant to them. A curated list of topics to study will most likely provide a person with a better base than what *appears* to relevant to their personal interests.
I certainly wouldn't comment without having experienced the system.
Experiencing the system 25 years into a career can be quite different than experiencing the system 5 years into a career. At that later stage in life did you have the free time to take advantage of all the outside-the-classroom opportunities to learn? Even though I was working 25-30 hours a week I was in my 20s so I had plenty of time to work with peers (fellow students) on personal projects, plenty of time to get access to some lab/workstation for no reason other than telling a professor I am curious about something and no its unrelated to any coursework, sitting in on classes I was not enrolled in because I was curious about something, etc.
I too have a 30+ year career. However I earned my CS degree 5 years into that career, not 25 years into it. Who might have the better perspective on how a CS degree affects the early portion of one's career?
Each will have gaps. The difference is that the autodidactic fills them on demand while the one requiring hand-holding is simply standing there with his limp dock in his hand.
You failed to note that the university trained also do a lot of independent study, they too do a lot of learning without any hand holding, at least for CS and related programs. Plus the gaps are less likely to be in more critical areas, data structures and algorithms come to mind. There are many self-taught aspiring game programmers who apply for a job at a game studio and are surprised that the screening test/questions have a lot to do with data structures and algorithms and not so much with Direct3D or OpenGL.
I hope that made you feel better, buying all that bullshit. Now, the truth.
Yes, please try to explain the "truth" to someone who has seen things from both sides. Who dropped out of a computer science program the middle of their sophomore year to pursue a unique startup opportunity, worked in such an environment for a couple years, moved to a more traditional software development job, went back to school and finished their degree.
Colleges do not teach people how to be good programmers.
The classroom is not the sole source of knowledge for computer science and related fields. You learn as much from your peers, fellow students, and from self-study as you do from professors. A University is a unique environment. Not only is it dense with peers from your field of study but also with peers from other areas. Other areas that may provide insights that are beneficial. A University also can provide access to equipment and software unavailable by your own means. Been there, done that. Access to people, info, equipment, software, etc, is amazing at a University. Add to that the University making me take classes I expected to be of little value, and much to my surprise turned out quite useful on the jobs years later.
In summary, a formal degree program can make you a better programmer than you would have otherwise been. It is additive to what you are doing on your own.
Very few jobs actually require a graduate degree...
Graduate degrees include Masters, not Just PhDs.
... teaching and research and maybe, some very very specialized field where it's just researchers who are doing the work; like getting a project form a uni.
Few, yes. Its rare that a computer vision related job pops up locally, computer vision being my area of research in my Masters program. Where the Masters is more commonly useful is in a position involving leading a development team in some way, of course it has to be combined with experience.
I personally find that a degree is about a sense of entitlement that has absolutely nothing to do with job skills.
As someone who has done the self-study route, the drop-out route when a commercial opportunity presented itself, the working in the industry route... I found that going back and finishing the degree (while working) was beneficial with respect to job skills.
You are made to think the way that your betters believe that you should think
You are severely misinformed, at least for courses of study like computer science. In such a program you learn as much from your peers and from self study as you do from your professors. I also had little problem arguing with my professors about something. They in fact seemed to enjoy a student do so rather than just repeat back to them the book or the lecture.
A degree has nothing to do with intelligence and not even much to do with persistence.
No one claimed a degree is some exclusive evidence of intelligence. What it is evidence of is a broad comprehensive body of relevant knowledge, again speaking from a computer science type of perspective. People going the self-taught route exclusively often have gaps. Topics they did not study since they were not interested and/or mistakenly thought unimportant. For example aspiring video game programmers who spent little to no time studying data structures.
Regarding persistence, it absolutely demonstrates the ability to complete a multi-year project filled with things you have no inner passion for.
You'll make it through as long as you can pay the bill and show up occasionally
And get weeded out in job interviews. Its why even those with degrees are subject to the various programming "tests" as part of the interview process.
Might there be a third category besides those who went to college and dropped out and those who completed? Hint: Begins with "auto" and ends with "didactic"
Self taught individuals often have gaps in their knowledge. They passed on learning about something they have little to no inherent interest in. Over several decades I've worked with many self taught individuals doing software development that I'd be happy to work with again, but only two had the personal motivation and discipline to read and learn the equivalent of a traditional CS/CE/etc coarse of study. And of course I've seen graduates that are pretty damn clueless and useless. However, overall, formal training seems to have an advantage over self taught.
Also, what makes you think formally trained and self taught are mutually exclusive? In courses of study like computer science you will likely learn as much from your colleagues and your own independent study as from the professors. That is part of the university experience too. Being dropped into a population of skilled and motivated people that one can also learn from; being dropped into an environment where you will have access to resources you would have never had otherwise.
Again, the main difference is that the university trained are more likely to have studied things they had little interest in, but things that make them more capable, things that often had a value they did not anticipate.
The ultimate killer feature is that bitcoin can't be counterfeited or manipulated by a central authority.
Yes, it can. 51% Attack, Majority Hash Rate Attack.
"To date, Ghash.io has twice come dangerously close to obtaining 51% of the bitcoin network's hashing power. The popular pool was able to gain 42% of the network in January, and, just last week, the pool reached a worrying 50%.
In theory, obtaining a majority of network power could potentially enable massive double spending and the ability to prevent transaction confirmations, among other potentially nefarious acts.
However, despite widespread concern about the vulnerability of the bitcoin network to large mining pools, there remains no easy solution to the issue." Jun 20, 2014 https://www.coindesk.com/51-at...
Slashdot Mirror
Note that a bank safe deposit box is one of the offsite options
Great news for law enforcement, this should help them get through that backlog of iPhones they want to examine. :-(
I think the larger trend is that lectures are moving to video, and class time is thereby freed up for Q&A (both directions, so some "Socratic" from the professor), discussion, and various other interactive things. The best part about video is that if you can follow it at 1.0x speed you can probably follow it at 1.6x speed, and when you can't follow it there is the rewind button.
There is also the need to publish in order to become tenured. Which has various economic and social benefits.
Unless you mean the publishers, they don't even today. You work for free for the journal and give them exclusive access, in return for a shot at publicity/"impact".
Note quite, a shot at tenure. "Publish or Perish". Everyone in academia knew this going in.
Even if Bitcoin does raise the block size one day, it's still not clear to me how it's ever going to scale high enough to become a general purpose currency. Going from 7 transactions/second to 56 sounds nice and all, but Visa and Mastercard handle about 2000 transactions/second. Each.
Are they going to eventually go to a 600+ MB block size? Or are a sizeable number of transactions going to have to go through bank-like intermediaries?
As I wrote before: "may adopt a larger block size **or some other remedy** at any time in the future".
My point is that Classic Bitcoin is not going to sit idly by and say: well, the old 1.0 version of our design can't handle our upcoming performance requirements. Lets just pack things up and call it a failure. No, they can implement whatever changes are necessary, block size or otherwise.
Actually, it does not matter. Bitcoin Cash is redundant. They may be correct as to the technical merits of why a change is needed regarding maximum block size but their fork is unnecessary. Ordinary Bitcoin may adopt a larger block size or some other remedy at any time in the future. The argument is only over whether something needs to be done today. Well that and internal developer politics.
Regular Bitcoin is not locked into some doomed course. They can make a block size change on the timeframe they think appropriate; or if Bitcoin Cash does enjoy increasing popularity they can make the change to put an end to the defections to the other side. Either way Bitcoin Cash seems doomed. Although I'm sure some speculators will find a way to make money off the hype.
Fanboys have a habit of twisting a disadvantage into a full advantage. Much like how back in the Power PC days. Apple use to show how the Power PC processor had handled a few Photoshop filters better than Intel chips, while the Intel Chips were in general faster overall.
As someone who did a lot of Intel and PowerPC programming back then, including optimized PowerPC and x86 assembly. PowerPC was technically faster even for general code. About 20% for CPUs of the same clock rate. Although I had many years more experience with x86 assembly, PowerPC also provided unique optimization opportunities that helped specialized code. Where Intel won overall was not having the better performing hardware (again, comparing CPUs with the same clock rate), rather they had a higher clock rate to overcome the inefficiencies of their architecture. "In theory" PowerPC had the advantage of a clean RISC design versus the legacy CISC of x86. CISC is much harder to design with and improve. However Intel overcame this by throwing tons of money into the project and frankly pulled off miracles. Looking back it was not that PowerPC failed to improve over time, its that no one (outside of Intel) expected x86 could reach the performance levels it has. Intel's engineers are miracle workers and deserve and enormous amount of credit. PowerPC expected Intel x86 to hit the wall many years earlier that it did, and PowerPC failed as a result.
That said, Apple's move to Intel was long overdue. Perhaps they should have never bothered with PowerPC and went from 68K to x86. Apple doubling their marketshare after the switch to Intel was not about slightly faster CPUs running native code. It was about being able to run Windows apps at reasonable speed even under emulation. Once on x86 CPUs they no longer had to emulate the underlying instruction set. Once they had this users no longer had to pick MacOS or Windows, they could have both. Even better, Apple added a dual boot option to run Windows natively when emulation might not be OK.
It's ridiculous to begin with. Comparing a specific phone to a platform? Has there ever been a requirement for Android hardware vendors to only manufacture phones with good cameras?
Apparently you missed most of the summary:
"I would never use an Android phone for photos! Here is the problem: It's Android. Android is an open source (mostly) operating system that has to be neutral to all parties. This sounds good until you get into the details. Ever wonder why a Samsung phone has a confused and bewildering array of photo options? Should I use the Samsung Camera? Or the Android Camera? Samsung gallery or Google Photos? It's because when Samsung innovates with the underlying hardware (like a better camera) they have to convince Google to allow that innovation to be surfaced to other applications via the appropriate API. That can take YEARS. Also the greatest innovation isn't even happening at the hardware level -- it's happening at the computational photography level. (Google was crushing this 5 years ago -- they had had 'auto awesome' that used AI techniques to automatically remove wrinkles, whiten teeth, add vignetting, etc... but recently Google has fallen back). Apple doesn't have all these constraints. They innovate in the underlying hardware, and just simply update the software with their latest innovations (like portrait mode) and ship it."
I challenge you to come up with a less farcical argument. :-)
I can tell you from experience degreed CS individuals also have gaps in their knowledge - so don't go there.
Its not the existence of gaps, its what those specific gaps are. CS individuals are less likely to have gaps in common important areas. For example, many self-taught programmers aspiring to game development crash and burn when they apply at a game studio due to a limited understanding of data structure and algorithms.
Today there are plenty of on-line training opportunities to supplement any self study. So, opportunities to 'self teach' isn't just limited to reading computer magazines.
No one said it is. Anything you are doing on your own outside of coursework or the job is self study. And University students have a lot of time for self study. Their University coursework adds to what they are doing on their own, it doesn't take anything away.
As for dropping into populations - that also applies to people being dropped into a population in a company.
Populations are usually far more diverse and far more accessible at a University than at a company. Experience at a company tends to be more focused, spare time to help with something not work related more limited due to family pressures, etc.
Experience beats training over the long haul.
Absolutely, the point is you will have more experience adding the formal studies than if you forgo them.
Finally - there are autodidacts that have the ability to not only learn material on their own, but also have acquired or developed on their own models that allow that information to be properly integrated and used in a work context.
And the broader your base of knowledge the more likely that is to happen due to cross-pollination. Those uninteresting classes that were forced upon you in the University sometimes turn out to be quite useful on the job years later. Again, formal training is additive, it adds to what you are doing on your own.
The fact that you believe code lives in a vacuum ... res ipso loquitor.
Sorry, no one claimed that code lives in a vacuum. All that is claimed is that (1) some things are left as an exercise for the student, OS configuration, integrated development environments, the fashionable programming language of the day, etc. All these are things students are expected to learn as needed on their own time. University time is better spent on things more persistent such as data structures and algorithms. The students can apply the concepts of this topic to whatever language they use, whatever operating system they target, etc. And it is also claimed that (2) gaps in core topics like data structures and algorithms is far more detrimental than gaps in OS configuration for someone in a software development career. For an IT career that would be reversed.
As I said, "more likely". There is always the unlikely exception. But what you offer are poor examples. There is no shortage of self-taught that are stumped by such things. And frankly, those are minor things in a University environment that students would be expected to figure out on their own. Recall my mention of learning as much from peers and independent study as from professors.
The point you are missing is that going the University route is additive, it adds to what you would have done on your own. It doesn't plug all possible gaps, it just avoids the more common and important ones. Gaps with respect to data structures and algorithms is a far greater hazard than configuring a python environment.
Yes, I rember when I first learned what pointers were: "Oh ... so that's what they call what I've been doing with indirect indexesd addressing. Pointer arrays." Then I found out about Yourdain & Constantine. Sorry, all that information is available much more so now than then, and it didn't even hold me back then.
And I read Knuth for fun in my drop-out days. Availability is not the issue. The issue is that many self-taught pass on available info that seems uninteresting to them, or in their erroneous belief not relevant to them. A curated list of topics to study will most likely provide a person with a better base than what *appears* to relevant to their personal interests.
The military is an even better way to demonstrate that you can finish what you start, despite having to deal with the boring and the unpleasant.
I certainly wouldn't comment without having experienced the system.
Experiencing the system 25 years into a career can be quite different than experiencing the system 5 years into a career. At that later stage in life did you have the free time to take advantage of all the outside-the-classroom opportunities to learn? Even though I was working 25-30 hours a week I was in my 20s so I had plenty of time to work with peers (fellow students) on personal projects, plenty of time to get access to some lab/workstation for no reason other than telling a professor I am curious about something and no its unrelated to any coursework, sitting in on classes I was not enrolled in because I was curious about something, etc.
I too have a 30+ year career. However I earned my CS degree 5 years into that career, not 25 years into it. Who might have the better perspective on how a CS degree affects the early portion of one's career?
Each will have gaps. The difference is that the autodidactic fills them on demand while the one requiring hand-holding is simply standing there with his limp dock in his hand.
You failed to note that the university trained also do a lot of independent study, they too do a lot of learning without any hand holding, at least for CS and related programs. Plus the gaps are less likely to be in more critical areas, data structures and algorithms come to mind. There are many self-taught aspiring game programmers who apply for a job at a game studio and are surprised that the screening test/questions have a lot to do with data structures and algorithms and not so much with Direct3D or OpenGL.
I hope that made you feel better, buying all that bullshit. Now, the truth.
Yes, please try to explain the "truth" to someone who has seen things from both sides. Who dropped out of a computer science program the middle of their sophomore year to pursue a unique startup opportunity, worked in such an environment for a couple years, moved to a more traditional software development job, went back to school and finished their degree.
Colleges do not teach people how to be good programmers.
The classroom is not the sole source of knowledge for computer science and related fields. You learn as much from your peers, fellow students, and from self-study as you do from professors. A University is a unique environment. Not only is it dense with peers from your field of study but also with peers from other areas. Other areas that may provide insights that are beneficial. A University also can provide access to equipment and software unavailable by your own means. Been there, done that. Access to people, info, equipment, software, etc, is amazing at a University. Add to that the University making me take classes I expected to be of little value, and much to my surprise turned out quite useful on the jobs years later.
In summary, a formal degree program can make you a better programmer than you would have otherwise been. It is additive to what you are doing on your own.
Very few jobs actually require a graduate degree ...
Graduate degrees include Masters, not Just PhDs.
Few, yes. Its rare that a computer vision related job pops up locally, computer vision being my area of research in my Masters program. Where the Masters is more commonly useful is in a position involving leading a development team in some way, of course it has to be combined with experience.
I personally find that a degree is about a sense of entitlement that has absolutely nothing to do with job skills.
As someone who has done the self-study route, the drop-out route when a commercial opportunity presented itself, the working in the industry route ... I found that going back and finishing the degree (while working) was beneficial with respect to job skills.
You are made to think the way that your betters believe that you should think
You are severely misinformed, at least for courses of study like computer science. In such a program you learn as much from your peers and from self study as you do from your professors. I also had little problem arguing with my professors about something. They in fact seemed to enjoy a student do so rather than just repeat back to them the book or the lecture.
A degree has nothing to do with intelligence and not even much to do with persistence.
No one claimed a degree is some exclusive evidence of intelligence. What it is evidence of is a broad comprehensive body of relevant knowledge, again speaking from a computer science type of perspective. People going the self-taught route exclusively often have gaps. Topics they did not study since they were not interested and/or mistakenly thought unimportant. For example aspiring video game programmers who spent little to no time studying data structures.
Regarding persistence, it absolutely demonstrates the ability to complete a multi-year project filled with things you have no inner passion for.
You'll make it through as long as you can pay the bill and show up occasionally
And get weeded out in job interviews. Its why even those with degrees are subject to the various programming "tests" as part of the interview process.
Might there be a third category besides those who went to college and dropped out and those who completed? Hint: Begins with "auto" and ends with "didactic"
Self taught individuals often have gaps in their knowledge. They passed on learning about something they have little to no inherent interest in. Over several decades I've worked with many self taught individuals doing software development that I'd be happy to work with again, but only two had the personal motivation and discipline to read and learn the equivalent of a traditional CS/CE/etc coarse of study. And of course I've seen graduates that are pretty damn clueless and useless. However, overall, formal training seems to have an advantage over self taught.
Also, what makes you think formally trained and self taught are mutually exclusive? In courses of study like computer science you will likely learn as much from your colleagues and your own independent study as from the professors. That is part of the university experience too. Being dropped into a population of skilled and motivated people that one can also learn from; being dropped into an environment where you will have access to resources you would have never had otherwise.
Again, the main difference is that the university trained are more likely to have studied things they had little interest in, but things that make them more capable, things that often had a value they did not anticipate.
The ultimate killer feature is that bitcoin can't be counterfeited or manipulated by a central authority.
Yes, it can. 51% Attack, Majority Hash Rate Attack.
"To date, Ghash.io has twice come dangerously close to obtaining 51% of the bitcoin network's hashing power. The popular pool was able to gain 42% of the network in January, and, just last week, the pool reached a worrying 50%. In theory, obtaining a majority of network power could potentially enable massive double spending and the ability to prevent transaction confirmations, among other potentially nefarious acts. However, despite widespread concern about the vulnerability of the bitcoin network to large mining pools, there remains no easy solution to the issue." Jun 20, 2014
https://www.coindesk.com/51-at...
If its an online purchase where the goods are to be delivered does waiting for the desired number of verifications matter?