I personally will not upgrade from 3.6 (on my Windows machines at least) just because of plain stubbornness. I strongly feel that Mozilla as a whole and Asa Dotzler in particular need to be somehow punished for ignoring "LTS" (including corporate and academics) users altogether. I am all for "spread the Firefox hate" campaign.
Call me a troll, but I was a loyal Firefox user since late 2003 (it was called "Firebird" then)... until they started to push versions upon me, destroying binary plugins and losing their identity as a stable browser in the process. Now I'm a Firefox hater.
In Russian every word has gender, primarily, but not always, determined from its ending. We don't use "it" (which is reserved for neutral gender only and would sound rude if applied to non-neutral-gender word), but "he/she" instead (note that this applies to words, not objects, e.g. "browser" is "he" because it ends with male-sounding -er, while "Mozilla" is "she", because it ends with female "-a").
Thus, "Russia" which ends with "-ia" (-ya) is female, but "Soviet Union" was male, because it ended with consonant (in Russian language, too). "United States" are plural, so we say "they" (rare in actual usage though), while America is female (again, "-a"). China is male though (ands with consonant -y in Russian, Kitay).
Don't you think this is a bubble that is going to pop sooner or later? Using mobile devices always feels constrained - even if I *can* do something on my phone, I often postpone the task until I am at the computer, so I can do it more comfortably (without any "please ignore the typos" messages).
I know that's stupid and tinfoil hattish, but one might think that some country found a way to control quakes and is testing it. First Japan, then East Coast of the USA, then Turkey, now geographical center of USA.
I'd advise Israeli and Californians to intensify training people how to survive major quakes, just in case.
To be fair, screens are also order of magnitude larger than they used to be (even more if you count memory and not pixels, since 99% of us have 32 bits per pixel these days and back then 8 bit was common), not to mention that having 2+ monitors is now standard for desktops. This, plus proliferation of antialased rendering offsets advancements in CPU power - to the point that navigating source code in QtCreator on my Linux box is not that smooth as I'd like it to be. Generally speaking, concerns about font rendering performance are still relevant.
Censorship is never a solution. Soviet Union made owning a photo of an erected penis a criminal offense. It is perhaps not a coincidence that Soviet society was plagued with sexual maniacs like Chikatilo.
Firefox lost its identity, trying hard to copy Chrome, despite Chrome's lesser popularity. If this is not changed, Firefox will fall behind Chrome when Firefox 26 is released (which is next year).
Canonical embraced it to the point that *buntu has python scripts running in background (printer-applet and such). This is perhaps a misuse of an interpreted language.
I think that such abstraction may be fine on certain high (application) level (i.e. basically like it's now), but not on API level. If at all, this should be organized like OSI model, where each layer would see data with less semantic "metadata" involved.
And even on application level, this approach ignores users like me. I am more or less fine with not knowing how and where my e-mails are stored (although I do backups to my disk, because I don't 100% trust Google), but I certainly would not like not knowing which disk my files are stored on. E.g. if police comes and takes all my computer equipment for checkup (because someone reported that I am using pirated software - things like that happen in my part of the world, at least to people who run their own businesses) I would like to be sure that my sensitive data is safe on a TrueCrypt'ed pendrive that I store elsewhere.
To sum up, I think that there's no need for such a change. It's not like people cannot grasp the concept of "traditional" file system - DropBox shows that people are perfectly able to store them in a "cloud". Making file storage opaque for the user is a short-term thinking, which helps rapid expansion due to being "newbie"-friendly, but limits everyone's ability to control the data (and also introduces a single point of failure that would be that all-encompassing "database"). And 'control freaks' are a sizeable population.
Well, we can have multiple hierarchies. Just as Google started to allow hierarchies of labels in GMail.
But frankly speaking, I wouldn't like to have my files in one big flat "All files" root directory with different labels/filters attached. Sure, they are like that on the disk at some level. Sure, most people treat their files like that (see desktop folder of an average user). Yet I, for one, like rigid hierarchies even if that means that sometimes I need to pick one of possible categorizations of a file (or duplicate it, or create a link), because they bring order to otherwise chaotic and messy natural way of thinking, just like strict logical reasoning brings order to our understanding of the world.
Your algorithm is susceptible to falling into local extrema. What if it's cheaper to pay for certain kind of energy, but it can be reasonably expected that its price will sharply go up during next 50 years? Isn't it better to pay more for another kind of energy which will stay roughly on the same level for next few hundred years?
This is not to say that I believe in "we're running out of oil/gas" predictions. I am just pointing out an obvious flaw of a "greedy algorithm" (choose whatever is cheaper now and hope that the price won't ever go up or costs of transition to next cheap thing will be affordable) - it's like buying real estate near dormant volcano just because it's cheap.
Well, programming practically is the computer-world equivalent of construction worker or cleaners. Sure, it's useful so people actually can get things done, but it isn't practically challenging or something lots of people can't do if given teaching. Developers have to make the important decisions regarding a product. If you wanted to work in the gaming industry, would you rather want to be a coder or actually the game designer?
Being game programmer is much more fun than being a game designer. Most "important decisions" boil down to balancing and scenario =) Stuff like gameplay mechanics (weapon design and AI behavior) is proposed and basically decided by programmers who permanently experiment with code, designers do influence that, but they have to account for what's technically possible and it's programmers who say "this cannot be done in reasonable time":)
To sum up, the more you understand, the more power you have. Programmers can and often do understand game designer's job, but not vice versa - so programmers can both design and implement stuff, but game designers can only design =) They are valued for having more consumer-oriented mindset than programmers, if not for that, they would be just glorified testers.
Re:Evolution = only good news for not-yet-born peo
on
The Real Job Threat
·
· Score: 1
Thanks for insightful post, I learned a lot from it.
The halting problem has nothing to do with this. By your reasoning humans are impossible (having improved on themselves in many domains). Guess what... they're not. Which must obviously mean that humans are fundamentally limited in the problems they can understand, not that they are somehow magically above the problem.
Actually, my - lame - theory only means that humans won't create anything that will "understand" (i.e. explain, predict, optimize, possibly describe properties) specific thing (process) better than the authors themselves. I.e. you will never get a tool that replaces a scientist or a programmer.
And humans haven't understand themselves in their full complexity yet. This is a bit separate question, but I think we will never be able to do that. We can possibly create a neural network accurately reflecting human brain - but it may turn out that you can't "teach" this network without providing it with exactly the same inputs a human baby is provided with, starting from physical feelings and not ending with interaction with other people. And suddenly modelling a single person requires you to accurately model a group of people, creating a chicken-and-egg problem.
However even having a neural network is not understanding it, I'd argue. At least - in my engineerish, non-scientific opinion - you can't, by just looking at it, tell what will it output for any given input and how to transform it into other forms while preserving its properties. I think that - in general - we only really understand formal rules that describe how this network "works" (i.e. how to calculate its outputs), but we don't really grasp it.
Sorry for not having read your wikipedia links yet - maybe they already answer my doubts. I will.
Once we figure out how to accurately simulate neuron interaction, we should be able to create a brain much more intelligent than ours.
I think that is going beyond creation/creator relationship and becomes usual evolution. I.e. there won't be "them" and "us", people will just evolve into beings with more intelligent brain and powerful bodies, but still sharing identity with us (inasmuch as we consider Neanderthal men our ancestors).
But I don't believe we will be able to accurately reproduce the way our brain works, at least not in purely hand-made form. We may get close enough models, attempting to copy it as a "black box", but I believe that an entity cannot fully understand its own operation...
This is similar to a halting problem, you want to have a program that could analyze and "understand" itself to the point that it could rewrite itself in a better way... If it were possible to create such a program then it would probably not exist at all, because whoever would write it would also knew better ways to write it - and it's turtles all the way down to understanding the "ultimate optimal form" of such a program and writing it in that form so it is no longer capable of self-imrpovement.
We can design and make machines to do things better than we can. That there are fewer jobs for us to do, given the above, should not come as a surprise.
This only holds true if level of education among humans stays the same. We may eliminate repetitive jobs, but humans will learn new techniques which won't be easily automated.
I think it's a reasonable assumption that one cannot design anything that has more chances to pass natural selection than its creator, so our creations will always be bounded by our own intellect.
of course everyone calls the Ubuntu version by its alliterative animal name
I, for one, prefer numbers and often cannot map them to animal names (especially given that names are different for English and localized Ubuntu versions), at least for past releases. Also, I don't remember the last time when someone else around me used animal names in spoken language (well, this can be partially explained by the fact that those names aren't easy to remember for non-English speakers).
Hard to estimate, given that a lot of second-hand iPhones are sold jail-broken. Certainly less users would buy it if they couldn't run pirate apps or use it with their favourite telco.
Well... I see your point, but I think that we need to teach people to program in the same way as hardware works (and unstructured BASIC happens to be closer to that). Ideally, I'd want every programmer to start with learning assembly and understanding what limitations we put on it in higher level languages and what trade-offs are involved - but in a less ideal world, they may begin with BASIC.
Also, there's this thing called "Data-Oriented Design/Programming" which despises OOP for not thinking about data flow, but about relations between objects. I personally think that DOD is refreshing, especially seeing as OOP lead us into world of slow bloatware. Again, BASIC does not really get into the way of learning DOD.
In Moscow subway, if your train is getting closer to city center, you hear male voice ("boss hurries you to work"), and if you are travelling in opposite direction, there's a female announcer ("wife calls you home") - they switch half-way for most trains. Male-oriented hint, but still.
Slashdot Mirror
The did release a cross-platform (running on FreeBSD) implementation of .NET 1.0... "from the start".
I personally will not upgrade from 3.6 (on my Windows machines at least) just because of plain stubbornness. I strongly feel that Mozilla as a whole and Asa Dotzler in particular need to be somehow punished for ignoring "LTS" (including corporate and academics) users altogether. I am all for "spread the Firefox hate" campaign.
Call me a troll, but I was a loyal Firefox user since late 2003 (it was called "Firebird" then)... until they started to push versions upon me, destroying binary plugins and losing their identity as a stable browser in the process. Now I'm a Firefox hater.
In Russian every word has gender, primarily, but not always, determined from its ending. We don't use "it" (which is reserved for neutral gender only and would sound rude if applied to non-neutral-gender word), but "he/she" instead (note that this applies to words, not objects, e.g. "browser" is "he" because it ends with male-sounding -er, while "Mozilla" is "she", because it ends with female "-a").
Thus, "Russia" which ends with "-ia" (-ya) is female, but "Soviet Union" was male, because it ended with consonant (in Russian language, too). "United States" are plural, so we say "they" (rare in actual usage though), while America is female (again, "-a"). China is male though (ands with consonant -y in Russian, Kitay).
Don't you think this is a bubble that is going to pop sooner or later? Using mobile devices always feels constrained - even if I *can* do something on my phone, I often postpone the task until I am at the computer, so I can do it more comfortably (without any "please ignore the typos" messages).
I certainly don't want this feeling on the PC.
Half of the complains about GNOME 3 is just from a bunch of Old Farts who do not want to see anything change.
We are using an OS with 40 years of tradition for a reason.
Well, actually there has already been unusual quake in Israel this year. Perhaps Mother Earth is shaking people off?
I know that's stupid and tinfoil hattish, but one might think that some country found a way to control quakes and is testing it. First Japan, then East Coast of the USA, then Turkey, now geographical center of USA.
I'd advise Israeli and Californians to intensify training people how to survive major quakes, just in case.
To be fair, screens are also order of magnitude larger than they used to be (even more if you count memory and not pixels, since 99% of us have 32 bits per pixel these days and back then 8 bit was common), not to mention that having 2+ monitors is now standard for desktops. This, plus proliferation of antialased rendering offsets advancements in CPU power - to the point that navigating source code in QtCreator on my Linux box is not that smooth as I'd like it to be. Generally speaking, concerns about font rendering performance are still relevant.
Censorship is never a solution. Soviet Union made owning a photo of an erected penis a criminal offense. It is perhaps not a coincidence that Soviet society was plagued with sexual maniacs like Chikatilo.
Firefox lost its identity, trying hard to copy Chrome, despite Chrome's lesser popularity. If this is not changed, Firefox will fall behind Chrome when Firefox 26 is released (which is next year).
Canonical embraced it to the point that *buntu has python scripts running in background (printer-applet and such). This is perhaps a misuse of an interpreted language.
This should remind people that posting their opinions on controversial topics under their "real names" decreases their chances to get a job.
Zuckerberg's idea of "single personality" makes it hard to separate professional and personal attitude.
Normally you pick one of possible categories, and I would grouped it with desk phone because of similar function.
I think that such abstraction may be fine on certain high (application) level (i.e. basically like it's now), but not on API level. If at all, this should be organized like OSI model, where each layer would see data with less semantic "metadata" involved.
And even on application level, this approach ignores users like me. I am more or less fine with not knowing how and where my e-mails are stored (although I do backups to my disk, because I don't 100% trust Google), but I certainly would not like not knowing which disk my files are stored on. E.g. if police comes and takes all my computer equipment for checkup (because someone reported that I am using pirated software - things like that happen in my part of the world, at least to people who run their own businesses) I would like to be sure that my sensitive data is safe on a TrueCrypt'ed pendrive that I store elsewhere.
To sum up, I think that there's no need for such a change. It's not like people cannot grasp the concept of "traditional" file system - DropBox shows that people are perfectly able to store them in a "cloud". Making file storage opaque for the user is a short-term thinking, which helps rapid expansion due to being "newbie"-friendly, but limits everyone's ability to control the data (and also introduces a single point of failure that would be that all-encompassing "database"). And 'control freaks' are a sizeable population.
Well, we can have multiple hierarchies. Just as Google started to allow hierarchies of labels in GMail.
But frankly speaking, I wouldn't like to have my files in one big flat "All files" root directory with different labels/filters attached. Sure, they are like that on the disk at some level. Sure, most people treat their files like that (see desktop folder of an average user). Yet I, for one, like rigid hierarchies even if that means that sometimes I need to pick one of possible categorizations of a file (or duplicate it, or create a link), because they bring order to otherwise chaotic and messy natural way of thinking, just like strict logical reasoning brings order to our understanding of the world.
People who say that hierarchical filesystems suck probably have a big mess on their table in real life.
Your algorithm is susceptible to falling into local extrema. What if it's cheaper to pay for certain kind of energy, but it can be reasonably expected that its price will sharply go up during next 50 years? Isn't it better to pay more for another kind of energy which will stay roughly on the same level for next few hundred years?
This is not to say that I believe in "we're running out of oil/gas" predictions. I am just pointing out an obvious flaw of a "greedy algorithm" (choose whatever is cheaper now and hope that the price won't ever go up or costs of transition to next cheap thing will be affordable) - it's like buying real estate near dormant volcano just because it's cheap.
Well, programming practically is the computer-world equivalent of construction worker or cleaners. Sure, it's useful so people actually can get things done, but it isn't practically challenging or something lots of people can't do if given teaching. Developers have to make the important decisions regarding a product. If you wanted to work in the gaming industry, would you rather want to be a coder or actually the game designer?
Being game programmer is much more fun than being a game designer. Most "important decisions" boil down to balancing and scenario =) Stuff like gameplay mechanics (weapon design and AI behavior) is proposed and basically decided by programmers who permanently experiment with code, designers do influence that, but they have to account for what's technically possible and it's programmers who say "this cannot be done in reasonable time" :)
To sum up, the more you understand, the more power you have. Programmers can and often do understand game designer's job, but not vice versa - so programmers can both design and implement stuff, but game designers can only design =) They are valued for having more consumer-oriented mindset than programmers, if not for that, they would be just glorified testers.
The halting problem has nothing to do with this. By your reasoning humans are impossible (having improved on themselves in many domains). Guess what ... they're not. Which must obviously mean that humans are fundamentally limited in the problems they can understand, not that they are somehow magically above the problem.
Actually, my - lame - theory only means that humans won't create anything that will "understand" (i.e. explain, predict, optimize, possibly describe properties) specific thing (process) better than the authors themselves. I.e. you will never get a tool that replaces a scientist or a programmer.
And humans haven't understand themselves in their full complexity yet. This is a bit separate question, but I think we will never be able to do that. We can possibly create a neural network accurately reflecting human brain - but it may turn out that you can't "teach" this network without providing it with exactly the same inputs a human baby is provided with, starting from physical feelings and not ending with interaction with other people. And suddenly modelling a single person requires you to accurately model a group of people, creating a chicken-and-egg problem.
However even having a neural network is not understanding it, I'd argue. At least - in my engineerish, non-scientific opinion - you can't, by just looking at it, tell what will it output for any given input and how to transform it into other forms while preserving its properties. I think that - in general - we only really understand formal rules that describe how this network "works" (i.e. how to calculate its outputs), but we don't really grasp it.
Sorry for not having read your wikipedia links yet - maybe they already answer my doubts. I will.
Once we figure out how to accurately simulate neuron interaction, we should be able to create a brain much more intelligent than ours.
I think that is going beyond creation/creator relationship and becomes usual evolution. I.e. there won't be "them" and "us", people will just evolve into beings with more intelligent brain and powerful bodies, but still sharing identity with us (inasmuch as we consider Neanderthal men our ancestors).
But I don't believe we will be able to accurately reproduce the way our brain works, at least not in purely hand-made form. We may get close enough models, attempting to copy it as a "black box", but I believe that an entity cannot fully understand its own operation...
This is similar to a halting problem, you want to have a program that could analyze and "understand" itself to the point that it could rewrite itself in a better way... If it were possible to create such a program then it would probably not exist at all, because whoever would write it would also knew better ways to write it - and it's turtles all the way down to understanding the "ultimate optimal form" of such a program and writing it in that form so it is no longer capable of self-imrpovement.
We can design and make machines to do things better than we can. That there are fewer jobs for us to do, given the above, should not come as a surprise.
This only holds true if level of education among humans stays the same. We may eliminate repetitive jobs, but humans will learn new techniques which won't be easily automated.
I think it's a reasonable assumption that one cannot design anything that has more chances to pass natural selection than its creator, so our creations will always be bounded by our own intellect.
of course everyone calls the Ubuntu version by its alliterative animal name
I, for one, prefer numbers and often cannot map them to animal names (especially given that names are different for English and localized Ubuntu versions), at least for past releases. Also, I don't remember the last time when someone else around me used animal names in spoken language (well, this can be partially explained by the fact that those names aren't easy to remember for non-English speakers).
Hard to estimate, given that a lot of second-hand iPhones are sold jail-broken. Certainly less users would buy it if they couldn't run pirate apps or use it with their favourite telco.
Well... I see your point, but I think that we need to teach people to program in the same way as hardware works (and unstructured BASIC happens to be closer to that). Ideally, I'd want every programmer to start with learning assembly and understanding what limitations we put on it in higher level languages and what trade-offs are involved - but in a less ideal world, they may begin with BASIC.
Also, there's this thing called "Data-Oriented Design/Programming" which despises OOP for not thinking about data flow, but about relations between objects. I personally think that DOD is refreshing, especially seeing as OOP lead us into world of slow bloatware. Again, BASIC does not really get into the way of learning DOD.
In Moscow subway, if your train is getting closer to city center, you hear male voice ("boss hurries you to work"), and if you are travelling in opposite direction, there's a female announcer ("wife calls you home") - they switch half-way for most trains. Male-oriented hint, but still.