Encoding something with that method works well enough for very short, well-defined animations.
Sorry, what are you imagining "that method", to be? What I'm suggesting is very high level descriptions of content - the level of a movie shooting script, along with information on cast/etc, but in a form designed by an AI to communicate efficiently with itself. The AI would use this to reconstruct the movie from scratch. To make it match the original as much as possible, it could imagine make ordered lists of each parameter, ordered by some measure of likelihood, and then store only these indexes into that huge multidimensional universe.
Anyway, I don't know on what basis you've decided that would only work for things that are short or well defined. To be clear, none of this is my crazy theory or something; this is the normal sort of thing people talk about when they talk about the frontiers of compression. Like, to compress Wikipedia beyond X, you need a powerful AI that starts doing work in this sort of way, and I'm sure you can find lots of speculative articles about how this might work. None of it is realistic now, but becomes very realistic in a world with AI that greatly surpasses human intelligence.
Even with the ~2000 exabytes of data we currently have "stored" in the world, the pointer to any point in it would be immensely huge
What are you talking about? A pointer to distinguish any bit within that 2000 exabytes of data is still going to be quite small - pointer size is logarithmic to the size of the set. Heck, I could reference any atom in the universe with ~100 bytes of data. But our compressed movie file wouldn't need to be anywhere near that extravagant. There are only so many white tops the actress would likely be wearing, etc...
I'm not sure I understand your complaint here. Are you saying it work well enough (1), that it wouldn't be exactly the same (2), or that it wouldn't operate on all possible inputs (3)?
On 1, there's no reason the output couldn't be much higher quality than the original. 2 seems likely to be true, but also a shallow complaint; the result could be very, very similar (I mean, who knows how much data it would take to encapsulate Michael Bay's directing preferences, but there's no hard reason to believe it's a big number) - or, again, better.
On 3, again you're trivially correct but it's not a fair complaint. If you're working against the set of "all possible arrays of pixels", then any significant compression is impossible. And yet video compression is useful technology, because the set of videos we're interested in compressing is an infinitesimal subset of possible videos.
My point was that for a realistic movie and realistic constraints on performance - ie. the normal parameters we'd expect such software to operate in - it's difficult to set hard limits on possible performance. Information theory doesn't provide us any kind of permanent objection (though, obviously, again, I don't expect that this software actually worked, or that any such software could feasibly work via "fractals", or any kind of pixel/image centric processing... it would need to be amazing, very content aware AI to actually get this performance).
It's not physically impossible - we can imagine a super AI (or huge team of humans) that could reconstruct a reasonable facsimile of a movie based on grainy thumbnails of the cast and a quick synopsis of the plot and shot framing (and with work you could fit that in 8k).
This makes sense as a function of RT getting information from more users and more critics.
When the site was smaller, user reviews came from people who were super into movies, and thus thought more like movie critics. Now it's a broader group of people who are more casual consumers, who think less like movie critics. On the other side, I think they're also casting a broader net on critics. This means a lower percentage of "big market populist" reviews that are trying to meet the needs of broad audiences, and more small market, niche audience "movie buff" reviewers.
Yeah - it's not a surprise that their tastes have diverged, but I don't know how you solve for that without changing their whole concept. Which maybe isn't a bad idea - limiting data to fresh/rotten for each review is going to produce very swingy results. I prefer something like metacritic; even if they can't perfectly translate every review score, I think it still ends up being a more useful score.
But of course it's fine to bundle a free browser with your OS...
Jebus yeah, I hear you man - I lived through that era! Kids today don't understand why we have to stay watchful. Stay vigilant.
But you watch - one day, one of these unscrupulous banana companies is going to try giving us a free browser again (that's the pattern through history: banana, bears, beats, browser) - but they won't get me. Je me souviens. I'll be there to say NON!
You free food hippies need to just get out of here: the fact that Amazon has a proper financial relationship with the distributor makes it worse, not better.
Do you really think the distributor wouldn't call it quits if they knew who Amazon was giving the banana to? They're the ones I feel for. Well, and the children who have to live in this neighborhood - what are they going to grow up thinking?
They just have to stop doing this, then. I mean, it's tragic when any business is harmed in any way - but this is just too far. The banana eatery business is what this country was built on, and I can't imagine the hardships faced by grocers selling less bananas than normal.
God only approves of food consumption if it's part of a legal financial transaction.
I've read a lot of Slashdot comments on this subject, so I'm sure there's no reason to worry. I'll summarize:
1. The Cotton Gin. Once there was a "cotton gin" and blacksmiths but we still have jobs, so no problem! 2. Humans scheduled to get big buff next patch 3. People have been wrong about this in the past, ipso facto QED they're wrong about it now: humans win forever. 4. Who wants some cheap crap? I want quality and craftmanship in my Cheetos, and only humans have feelings and I want personal touch and... my waitress was cute that one time? 5. We'll still need poets and robot repairs guys. Probably everyone will do that.
He's either overstating the case (Vive does have better tracking which helps) or maybe he's comparing different experiences. The real differentiator is what you're doing. On a proper setup, people don't get sick playing roomscale Vive games. That is true, and true for probably a hundred people I've had try my Vive. However, you can get people sick immediately with artificial movement, and especially artificial rotation. Quite often at the end of my demos, I'll see if people want to try feeling sick, and we'll pull out Fancy Skiing or something - and almost everyone gets queasy the first time.
So yeah, if he's showing people Job Simulator on Vive, and some flying game on Oculus, he probably isn't lying, but he's coming to the wrong conclusion.
You need 3 cameras to do much with Oculus Touch (bringing the cost over that of a Vive) and the tracking is nowhere near as good. I have both headsets: the Oculus is more comfortable and makes for a generally a better visual experience, but Vive tracking and general roomscale experience is better. The Oculus cameras are a dead end - either they'll move to some kind of inside-out tracking, or something like Lighthouses will win.
Yeah, Slashdot has become a reasonably depressing site, where people with no idea or insight come to be mad about technology. Sometimes it's grinding some pointless axe about a programming language, bit of software, or company - sometimes it's sour graping about expensive tech.
You, me, and everyone else who has played with modern VR understands current limitations. Meanwhile, the critics are stuck in some weird time loop from 2007 where it makes everyone sick and doesn't track right and looks terrible. I also like the complaint that it's not social - I game alone reasonably often, but pretty much never use the Vive alone, and I use it really quite often. Something about the experience makes it much better with other people.
My area (southern Alberta, Canada) and family have their own weird cultural stuff going on, so I'm sure there's lots of stuff like this that's common elsewhere, but that I haven't run into. Anywho, interesting stuff.
Lol - I didn't know that. Certainly makes sense that it would work. Probably would work even better mixed in with some booze. Or, hell, I don't know, uh, coca leaves and mercury.
..but Belladonna is intended to be one of the "homeopathic ingredients" here (along with some other nonsense). From their website:
Calcarea Phosphorica 12X HPUS: teething, dentition Chamomilla 6X HPUS (Chamomile): for peevishness, restlesness and irritability Coffea Cruda 6X HPUS: sleeplessness Belladonna 6X HPUS (0.0000003% Alkaloids, calculated): redness and teething discomfort
To be clear, belladonna seems like a possible legitimate treatment (it does deaden nerves) if you got a real dose - but probably isn't effective at the intended concentration (nor is it probably a great idea in any case... I mean, opium would probably deal with teething pain too, but your pharmacist is hardly going to give it to your for teething). So if the gel is ever doing anything (via the actual effects of belladonna, rather than the backwards magic effect of removing all of the belladonna), that itself is evidence that they've screwed up their process and are getting more in than they meant to. Which isn't what you want to hear when the ingredient in question is straight-up-good-way-to-murder-someone poison.
Having hands (well, controllers at least) and being able to walk around naturally/poke stuff helps a lot, as does better head tracking.
I mean, I don't know if that's "overwhelming benefits" (I'd recommend waiting for next generation to buy at this point) - but it's pretty cool, and worth trying out a Vive (or Rift with Touch) if you get a chance.
If all you've tried is GearVR, you haven't tried modern VR in any meaningful way - it's garbage. Really. Find someone with a proper computer and a Vive.
Well, I guess the difference is that while it would be a stretch to call this AI, it is very reasonable to call it machine learning - and that's what they called it. And yeah, I get it, you have some pointless axe to grind about people generally overusing the term AI. Nobody cares.
But I will say that this exact comment, which is a waste of space every time it crops up on this kind of story, at least provided me a laugh this time. Your shitty, pouty reply is just icing on the entertainment cake. I would have felt maybe a bit mean if you'd actually copped to shitposting without reading the summary. But I don't imagine you're that kind of person (ie. someone with a shred of dignity or self awareness).
I don't see anyone calling this AI except you. Whatever pattern matching you use to determine when to deploy this garbage stock comment, I think it has misfired here...or, wait, do you also have some pointless objection to calling this machine learning?
I'm not saying I necessarily agree with dude, but the situations are different.
When picking between Hillary and Bernie, voters aren't just deciding which of those two they want to win, they're also trying to help "a Democrat" winning the general election. The idea, commonly presented at the time, was that a protracted race would hurt Hillary as the candidates hit each other, and that in a long fight with Bernie Hillary would be forced further to the left. Thus if you, as a Democratic primary voter, believe Hillary is going to win regardless, you might as well avoid the long fight and just vote for her to help win the general.
You could expect sort-of-related polling effects in a general if there was a third party (ie. I want Hillary to win over Trump, but she's so far ahead I feel safe to use my vote to support the Green Party or whatever) - but I don't think that was much of a factor this time. You may have seen some depressed turnout for Hillary, though, as some may have felt it was a done deal.
Overall, whatever polling may have done, I think the much more explanatory factors are "Many people don't like Hillary" and "People are tired of doing what the media says" (which we saw previously with Brexit).
Make the site request permission, like it would for the camera or GPS location.
I don't see this is as some big thing, but just because I can't think of an important use case doesn't mean there isn't a good one somewhere. Surely somebody wanted this for some reason? Anyway, it seems weird to introduce it then take it back over concerns that seem pretty mild, and also pretty easy to address the same way other concerns have been addressed in the past.
I don't think you're being fair here. As before, I believe (based on experience) that Macs overall will require less support given time, and I'm aware of lots of problems with Windows PCs. I just don't believe it's going to stay at an 8 times ratio. Like you say, there's going to be lots of users that will require support no matter what they're given (ie. the kind of people who can't find the start menu). If nothing else, they'll be the leveler here. There's no way you'll keep the "Mac users who need support" at 1/8th as you bring more of "them" into the mix.
I think you're imagining that I'm fighting for some Windows side or something - I'm not, I'm just suggesting that we temper unrealistic expectations.
Moving on, it's "hand waving", not "hand waiving", and I was right to ignore his sad little anecdote. Lots of people who are into computers (of all types) manage to keep their computers going without support; the fact that he was able to keep his home Mac going brings nothing to the conversation. Would you find it interesting to know that my non-techie dad manages to keep a Windows PC going, year-after-year without help? Does that change your view of the whole thing? Of course not; that's the kind of dumb crap you bring up when you have no facts or knowledge, but you feel like you need to write -something- in your condescending garbage post.
Anyway, yeah, my post was a little sarcastic in that previous post, but when buddy is like "facepalm, don't you know Macs don't need any support?", my inclination to be polite kind of evaporates.
...at a medium sized company that supports Windows, Mac, and Linux desktops. I'm more on the programming side, but I stay on top of the support issues for various departments. Macs need tech support largely for the same reason Windows users do: because most users aren't terribly computer savvy, aren't confident enough to just try plugging things in, make dumb mistakes, and generally don't know where to find easy answers.
From my experience, Macs need very little tech support when we give them to, say, the publications department - but become much more problematic for field staff and managers (especially to start) because things aren't where they've grown to expect them to be, because of limited software availability, and because of more limited "local guru helpers" (ie. that guy in cubicle 4 who's into computers).
So when I say that I wouldn't think IBM will see this sort of support benefit ratio as they move to wider roll out, I'm doing so based on experience, and also on a suspicion that IBM has motivation to present this information in an exaggerated way (a suspicion confirmed by insider perspectives in other comments).
But now that I know that you, personally, haven't had problems with your Macs... well that changes everything. Thanks so much.
I mean, I'm sure our Linux users overall require the least tech support. But that's a function of who they are more than what they're using.
I don't doubt that Macs require less support, but 40% vs 5% says that something else is going on - and I doubt that sort of ratio will hold once people are converted in bulk.
..and Amazon will likely retain its lead in supporting large and/or public facing websites.
But there's a lot of businesses, usually non-IT focused ones, who will remain Microsoft shops. For them, hosting their internal, B2B, and smaller public applications on Azure makes a lot of sense, and it's going to be a growth area for MS for a while. Yes you can host your.NET applications lots of places, but all the Azure nonsense is baked into Visual Studio and so, to the extent it basically works, it's going to be the path of least resistance for a lot of people.
In a lot of ways, MS is in a different market than other cloud vendors.
Slashdot Mirror
Sorry, what are you imagining "that method", to be? What I'm suggesting is very high level descriptions of content - the level of a movie shooting script, along with information on cast/etc, but in a form designed by an AI to communicate efficiently with itself. The AI would use this to reconstruct the movie from scratch. To make it match the original as much as possible, it could imagine make ordered lists of each parameter, ordered by some measure of likelihood, and then store only these indexes into that huge multidimensional universe.
Anyway, I don't know on what basis you've decided that would only work for things that are short or well defined. To be clear, none of this is my crazy theory or something; this is the normal sort of thing people talk about when they talk about the frontiers of compression. Like, to compress Wikipedia beyond X, you need a powerful AI that starts doing work in this sort of way, and I'm sure you can find lots of speculative articles about how this might work. None of it is realistic now, but becomes very realistic in a world with AI that greatly surpasses human intelligence.
What are you talking about? A pointer to distinguish any bit within that 2000 exabytes of data is still going to be quite small - pointer size is logarithmic to the size of the set. Heck, I could reference any atom in the universe with ~100 bytes of data. But our compressed movie file wouldn't need to be anywhere near that extravagant. There are only so many white tops the actress would likely be wearing, etc...
I'm not sure I understand your complaint here. Are you saying it work well enough (1), that it wouldn't be exactly the same (2), or that it wouldn't operate on all possible inputs (3)?
On 1, there's no reason the output couldn't be much higher quality than the original. 2 seems likely to be true, but also a shallow complaint; the result could be very, very similar (I mean, who knows how much data it would take to encapsulate Michael Bay's directing preferences, but there's no hard reason to believe it's a big number) - or, again, better.
On 3, again you're trivially correct but it's not a fair complaint. If you're working against the set of "all possible arrays of pixels", then any significant compression is impossible. And yet video compression is useful technology, because the set of videos we're interested in compressing is an infinitesimal subset of possible videos.
My point was that for a realistic movie and realistic constraints on performance - ie. the normal parameters we'd expect such software to operate in - it's difficult to set hard limits on possible performance. Information theory doesn't provide us any kind of permanent objection (though, obviously, again, I don't expect that this software actually worked, or that any such software could feasibly work via "fractals", or any kind of pixel/image centric processing... it would need to be amazing, very content aware AI to actually get this performance).
It's not physically impossible - we can imagine a super AI (or huge team of humans) that could reconstruct a reasonable facsimile of a movie based on grainy thumbnails of the cast and a quick synopsis of the plot and shot framing (and with work you could fit that in 8k).
But yeah.. with "fractals", not so much.
This makes sense as a function of RT getting information from more users and more critics.
When the site was smaller, user reviews came from people who were super into movies, and thus thought more like movie critics. Now it's a broader group of people who are more casual consumers, who think less like movie critics. On the other side, I think they're also casting a broader net on critics. This means a lower percentage of "big market populist" reviews that are trying to meet the needs of broad audiences, and more small market, niche audience "movie buff" reviewers.
Yeah - it's not a surprise that their tastes have diverged, but I don't know how you solve for that without changing their whole concept. Which maybe isn't a bad idea - limiting data to fresh/rotten for each review is going to produce very swingy results. I prefer something like metacritic; even if they can't perfectly translate every review score, I think it still ends up being a more useful score.
Jebus yeah, I hear you man - I lived through that era! Kids today don't understand why we have to stay watchful. Stay vigilant.
But you watch - one day, one of these unscrupulous banana companies is going to try giving us a free browser again (that's the pattern through history: banana, bears, beats, browser) - but they won't get me. Je me souviens. I'll be there to say NON!
You free food hippies need to just get out of here: the fact that Amazon has a proper financial relationship with the distributor makes it worse, not better.
Do you really think the distributor wouldn't call it quits if they knew who Amazon was giving the banana to? They're the ones I feel for. Well, and the children who have to live in this neighborhood - what are they going to grow up thinking?
They just have to stop doing this, then. I mean, it's tragic when any business is harmed in any way - but this is just too far. The banana eatery business is what this country was built on, and I can't imagine the hardships faced by grocers selling less bananas than normal.
God only approves of food consumption if it's part of a legal financial transaction.
I've read a lot of Slashdot comments on this subject, so I'm sure there's no reason to worry. I'll summarize:
1. The Cotton Gin. Once there was a "cotton gin" and blacksmiths but we still have jobs, so no problem!
2. Humans scheduled to get big buff next patch
3. People have been wrong about this in the past, ipso facto QED they're wrong about it now: humans win forever.
4. Who wants some cheap crap? I want quality and craftmanship in my Cheetos, and only humans have feelings and I want personal touch and... my waitress was cute that one time?
5. We'll still need poets and robot repairs guys. Probably everyone will do that.
He's either overstating the case (Vive does have better tracking which helps) or maybe he's comparing different experiences. The real differentiator is what you're doing. On a proper setup, people don't get sick playing roomscale Vive games. That is true, and true for probably a hundred people I've had try my Vive. However, you can get people sick immediately with artificial movement, and especially artificial rotation. Quite often at the end of my demos, I'll see if people want to try feeling sick, and we'll pull out Fancy Skiing or something - and almost everyone gets queasy the first time.
So yeah, if he's showing people Job Simulator on Vive, and some flying game on Oculus, he probably isn't lying, but he's coming to the wrong conclusion.
You need 3 cameras to do much with Oculus Touch (bringing the cost over that of a Vive) and the tracking is nowhere near as good. I have both headsets: the Oculus is more comfortable and makes for a generally a better visual experience, but Vive tracking and general roomscale experience is better. The Oculus cameras are a dead end - either they'll move to some kind of inside-out tracking, or something like Lighthouses will win.
Yeah, Slashdot has become a reasonably depressing site, where people with no idea or insight come to be mad about technology. Sometimes it's grinding some pointless axe about a programming language, bit of software, or company - sometimes it's sour graping about expensive tech.
You, me, and everyone else who has played with modern VR understands current limitations. Meanwhile, the critics are stuck in some weird time loop from 2007 where it makes everyone sick and doesn't track right and looks terrible. I also like the complaint that it's not social - I game alone reasonably often, but pretty much never use the Vive alone, and I use it really quite often. Something about the experience makes it much better with other people.
My area (southern Alberta, Canada) and family have their own weird cultural stuff going on, so I'm sure there's lots of stuff like this that's common elsewhere, but that I haven't run into. Anywho, interesting stuff.
Lol - I didn't know that. Certainly makes sense that it would work. Probably would work even better mixed in with some booze. Or, hell, I don't know, uh, coca leaves and mercury.
..but Belladonna is intended to be one of the "homeopathic ingredients" here (along with some other nonsense). From their website:
Calcarea Phosphorica 12X HPUS: teething, dentition
Chamomilla 6X HPUS (Chamomile): for peevishness, restlesness and irritability
Coffea Cruda 6X HPUS: sleeplessness
Belladonna 6X HPUS (0.0000003% Alkaloids, calculated): redness and teething discomfort
To be clear, belladonna seems like a possible legitimate treatment (it does deaden nerves) if you got a real dose - but probably isn't effective at the intended concentration (nor is it probably a great idea in any case... I mean, opium would probably deal with teething pain too, but your pharmacist is hardly going to give it to your for teething). So if the gel is ever doing anything (via the actual effects of belladonna, rather than the backwards magic effect of removing all of the belladonna), that itself is evidence that they've screwed up their process and are getting more in than they meant to. Which isn't what you want to hear when the ingredient in question is straight-up-good-way-to-murder-someone poison.
Having hands (well, controllers at least) and being able to walk around naturally/poke stuff helps a lot, as does better head tracking.
I mean, I don't know if that's "overwhelming benefits" (I'd recommend waiting for next generation to buy at this point) - but it's pretty cool, and worth trying out a Vive (or Rift with Touch) if you get a chance.
If all you've tried is GearVR, you haven't tried modern VR in any meaningful way - it's garbage. Really. Find someone with a proper computer and a Vive.
Well, I guess the difference is that while it would be a stretch to call this AI, it is very reasonable to call it machine learning - and that's what they called it. And yeah, I get it, you have some pointless axe to grind about people generally overusing the term AI. Nobody cares.
But I will say that this exact comment, which is a waste of space every time it crops up on this kind of story, at least provided me a laugh this time. Your shitty, pouty reply is just icing on the entertainment cake. I would have felt maybe a bit mean if you'd actually copped to shitposting without reading the summary. But I don't imagine you're that kind of person (ie. someone with a shred of dignity or self awareness).
I don't see anyone calling this AI except you. Whatever pattern matching you use to determine when to deploy this garbage stock comment, I think it has misfired here. ..or, wait, do you also have some pointless objection to calling this machine learning?
I'm not saying I necessarily agree with dude, but the situations are different.
When picking between Hillary and Bernie, voters aren't just deciding which of those two they want to win, they're also trying to help "a Democrat" winning the general election. The idea, commonly presented at the time, was that a protracted race would hurt Hillary as the candidates hit each other, and that in a long fight with Bernie Hillary would be forced further to the left. Thus if you, as a Democratic primary voter, believe Hillary is going to win regardless, you might as well avoid the long fight and just vote for her to help win the general.
You could expect sort-of-related polling effects in a general if there was a third party (ie. I want Hillary to win over Trump, but she's so far ahead I feel safe to use my vote to support the Green Party or whatever) - but I don't think that was much of a factor this time. You may have seen some depressed turnout for Hillary, though, as some may have felt it was a done deal.
Overall, whatever polling may have done, I think the much more explanatory factors are "Many people don't like Hillary" and "People are tired of doing what the media says" (which we saw previously with Brexit).
Make the site request permission, like it would for the camera or GPS location.
I don't see this is as some big thing, but just because I can't think of an important use case doesn't mean there isn't a good one somewhere. Surely somebody wanted this for some reason? Anyway, it seems weird to introduce it then take it back over concerns that seem pretty mild, and also pretty easy to address the same way other concerns have been addressed in the past.
I don't think you're being fair here. As before, I believe (based on experience) that Macs overall will require less support given time, and I'm aware of lots of problems with Windows PCs. I just don't believe it's going to stay at an 8 times ratio. Like you say, there's going to be lots of users that will require support no matter what they're given (ie. the kind of people who can't find the start menu). If nothing else, they'll be the leveler here. There's no way you'll keep the "Mac users who need support" at 1/8th as you bring more of "them" into the mix.
I think you're imagining that I'm fighting for some Windows side or something - I'm not, I'm just suggesting that we temper unrealistic expectations.
Moving on, it's "hand waving", not "hand waiving", and I was right to ignore his sad little anecdote. Lots of people who are into computers (of all types) manage to keep their computers going without support; the fact that he was able to keep his home Mac going brings nothing to the conversation. Would you find it interesting to know that my non-techie dad manages to keep a Windows PC going, year-after-year without help? Does that change your view of the whole thing? Of course not; that's the kind of dumb crap you bring up when you have no facts or knowledge, but you feel like you need to write -something- in your condescending garbage post.
Anyway, yeah, my post was a little sarcastic in that previous post, but when buddy is like "facepalm, don't you know Macs don't need any support?", my inclination to be polite kind of evaporates.
...at a medium sized company that supports Windows, Mac, and Linux desktops. I'm more on the programming side, but I stay on top of the support issues for various departments. Macs need tech support largely for the same reason Windows users do: because most users aren't terribly computer savvy, aren't confident enough to just try plugging things in, make dumb mistakes, and generally don't know where to find easy answers.
From my experience, Macs need very little tech support when we give them to, say, the publications department - but become much more problematic for field staff and managers (especially to start) because things aren't where they've grown to expect them to be, because of limited software availability, and because of more limited "local guru helpers" (ie. that guy in cubicle 4 who's into computers).
So when I say that I wouldn't think IBM will see this sort of support benefit ratio as they move to wider roll out, I'm doing so based on experience, and also on a suspicion that IBM has motivation to present this information in an exaggerated way (a suspicion confirmed by insider perspectives in other comments).
But now that I know that you, personally, haven't had problems with your Macs... well that changes everything. Thanks so much.
I mean, I'm sure our Linux users overall require the least tech support. But that's a function of who they are more than what they're using.
I don't doubt that Macs require less support, but 40% vs 5% says that something else is going on - and I doubt that sort of ratio will hold once people are converted in bulk.
..and Amazon will likely retain its lead in supporting large and/or public facing websites.
But there's a lot of businesses, usually non-IT focused ones, who will remain Microsoft shops. For them, hosting their internal, B2B, and smaller public applications on Azure makes a lot of sense, and it's going to be a growth area for MS for a while. Yes you can host your .NET applications lots of places, but all the Azure nonsense is baked into Visual Studio and so, to the extent it basically works, it's going to be the path of least resistance for a lot of people.
In a lot of ways, MS is in a different market than other cloud vendors.
Transcription is obviously a lot more straightforward, and the goalposts should be pretty easy to set.