Container is a generic term. Solaris provides a service called Zones, which are based on FreeBSD Jails (but with a few improvements over the older Jails, such as full network-stack virtualisation and support for SysV IPC, both of which I think are now supported in FreeBSD). Linux provides a bunch of services that can be cobbled together to provide more or less the same abstractions.
Yes, the Office is the same - you download the same software and can use it offline.
It's also worth noting that the O365 subscription gives you access to the Windows, Mac, Android and iOS versions. Unfortunately, my employer set it up in a stupid way. If I use the web version then I'm redirected to a web page run by our org that handles the authentication, but for the mobile apps I have to type in my credentials directly to the app (which is then sent to our servers for authentication). This would be fine, except that they don't let me set a different password for that and the systems that give access to a bunch of information that is considered highly confidential and where I'd probably be breaking the law (GDPR) if I provided the password to Microsoft.
The danger of the cloud is that it's a single point of failure and a high-value target. Your 12 guys may not be the most competent in the world, but that doesn't matter unless your data are the most valuable in the world. Someone attacking Azure might not be attacking you, but if they happen to get your data at the same time then selling it is just bonus money for them. Similarly, people are more likely to try to DDoS Azure or AWS than they are your in-house server.
Meltdown allowed you to dump the contents of memory of other VMs on the same nodes as you (including those in Azure Secure Cloud SGX enclaves, if you have access to the preview of that). Microsoft has now rolled out some mitigations, but the vulnerability existed from before the creation of Azure until recently. Anything in multi-tenant systems in that window may have been compromised. What's your plan for preventing that in the future and for dealing with it if it's happened?
The film adaptation was pretty close to the book, thought there were a number of things in the book that were either hard to with the technology of the time or just hard to translate to film entirely (the TV addiction in the book, for example).
I don't think the book aged well. Even at the time of release, microfiche existed and given that TVs were everywhere in their future it was surprising that no one realised that you could put one page of a book in each frame of a TV recording and store books like that, rather than having people memorise them. Or even sneak them in as individual frames in other shows and let everyone else distribute them, but not notice their existence without a special TV: one frame of book every 200 frames would give you one novel every half hour to hour of TV show.
If you don't consider learning from a body of knowledge no human had previously "discovering something new" than you'll have to accept that leaning from a body of knowledge built up by humans over centuries isn't discovering anything new either, and in that case most people never discovery anything new in their lives.
I don't understand your point. In the latter category, you learn from your predecessors but then you add to the body of knowledge that they've created. If In the former category, you learn from your predecessors and add nothing to the body of knowledge they've created. The second is a scientific endeavour, the first is the intellectual equivalent of binge-watching a TV show.
Where would their AI run? If they're sending a physical machine, or they're sending some kind of energy construct that is equivalent to a physical machine, then that physical presence would let them acquire samples. If they expect their AI to run on our computers, then they need to understand how our computers work. To give an analogy, we understand pretty much every possible way of building a steam engine that's possible with Victorian technology, including a bunch that the Victorians didn't come up with, but that doesn't mean that we could tell someone how to sabotage a Victorian-era steam engine site-unseen, because there are a lot of different designs with different failure modes.
That's what I thought, but the ones in my old phone and a tiny laptop that supported SD cards for extra storage both died after a couple of years with very few writes. I think the cheap ones just die over time.
I'm on a pre-pay contract where I typically spend well under £5/month, so I wondered if your numbers were actually correct. Here's what I discovered from my provider:
Galaxy S9: £99 + £40/month.
iPhone 7: £79 + £36/month.
iPhone SE: £19 + £18/month.
Galaxy J3: £19 + £14/month.
Alcatel Pixi 3 (never heard of it - really low-end Android phone that's worse than my 4-year-old one): £19 + £7/month.
So the difference between the latest phone and the cheapest currently-made iPhone is the difference between £40/month and £18/month. That's on a 24-month plan, so it works out at £960 versus £432.
The contracts are actually much better value than I thought for the iPhone SE. The cost from Apple is £349, so you're paying about £3.50/month for a plan that gives you 4GB of data and unlimited minutes. The iPhone 7 plan is a little over £20/month more than the cost of the phone direct from Apple, and that plan comes with only 1GB/month.
Their cheapest SIM-only plan is £5/month and gives unlimited texts, but only 200 minutes and 500MB, so if you can sell on the iPhone for close to its retail value then it's best to get the iPhone SE and just keep the SIM.
Not sure about a $1000 laptop. My partner is using my old one, which cost about double that in 2011, because the one that she bought about two year ago for a little under $1,000 is, even after a replacement battery, basically dead (and Dell support is crap). A $1,000 laptop might still work after 3-4 years, but increasingly the low-end laptops are built with cheap components making them effectively disposable devices.
My phone is from 2013 and is not a brick. The one it replaced, from 2010 is, because it isn't supported by LineageOS and so is stuck with a version of the TLS stack that doesn't have updated root certs and doesn't support modern cyphers, so it can't actually connect to most things that use TLS. My phone was not $1,000, it was under $200. It still gets regular third-party updates. To compare a more expensive phone, Samsung Galaxy S4 was released about 4 years ago (though it was $650, not $1,000, at launch). It has a big.LITTLE CPU arrangement, where the LITTLE cluster is the same as the only CPUs in my phone and where the big cluster is a quad-core 1.6GHz Cortex A15 (superscalar, out of order, reasonable performance) - some versions had better CPUs. LineageOS supports it with the latest version, so it can still get security updates. It had 2GB of RAM (twice as much as mine), 16-32GB of Flash (plus an SD card slot that can handle up to an additional 256GB), and a user-replaceable battery (so the thing that's most likely to die can be swapped out). Oh, and a 1080p screen. Oh, and it supports NFC, so it can work with Google Pay if you're willing to trust Google for that kind of thing.
About the only thing that a more modern smartphone has that this lacks is a 64-bit CPU, but Android shows no sign of dropping 32-bit support any time soon and I've yet to see any new 64-bit-only apps (unlike iOS, where 32-bit is effectively dead). I can't think of anything that you'd do with a smartphone that wouldn't work well on an S4 (though perhaps I'm lacking in imagination).
For me, the difference is that I use a laptop for as a primary work computer, so it's easy to justify spending money (ideally, my employer's money) on a decent one.
Google doesn't care if you buy a new phone, they just care if you use in-app payments through Google Play Services (or, ideally, buy new apps through the Play Store) or have their tracking stuff installed. They've actually made Android faster. When I upgraded to the a LineageOS based on the newer AOSP, most of my apps were noticeably faster due to improvements in the ART compiler (which, to be fair, started from a really slow place).
I suspect it is a matter of degree. If someone presents you with a new discovery that opens up new fields of research, that's exciting. If someone dumps a few thousand years of scientific research on you so that it will take a decade or more of relearning to get to the point where you even understand some of the state of the art, let alone are able to contribute to it, then that's probably disappointing. Would you really enjoy knowing that you would never discover anything new again?
During the Democratic Primary, polls were carried out on the general public asking them to rate their preferences on pairs of R vs D candidates. Clinton was the only one that was not seen as definitely better than Trump. Now, it's possible that another candidate might have slipped in the polls, but none of the others had such a poor starting performance. I suspect that if the ballots had said Trump, Not Clinton, Not Trump, Clinton, then the two most popular choices would have been Not Clinton and Not Trump.
Exactly. A lot of this kind of science fiction plot falls down on basic economics. Oh no, aliens have come to steal our water! Wait, they've invested vast amounts of energy to cross interstellar spaces to get two of the most common elements in the universe from the bottom of a gravity well? They've come to enslave us... so that we can work orders of magnitude less efficiently than their machines. They've come to steal or planet almost makes sense, though unless they evolved on a very similar one they could build space habitats with more landmass that would be more comfortable a lot more easily.
At a sufficiently high level of technology, most resources boil down to energy and that's the thing that you'd use in vast quantities reaching us. I suppose that they might want to come and build a Dyson sphere around the sun to use it as a power source, but then they wouldn't need us.
I guess the only thing that really makes sense is if they wanted our planet and needed some kind of receiver for their invasion ship to slow down. The threat then would be 'build a thing so we can slow down and invade you, or we'll just crash into you at high speed and destroy all life on your world'. Which would probably be phrased as 'please help us slow down, our brakes broke and we have no way of avoiding hitting your world!'
Here's the problem: the danger of ideas is very closely tied to the intelligence that is interpreting them. Look at religions, for an example of self-replicating ideas: even the most successful ones haven't managed to spread to half of the population, and they've evolved in situ. For a counter example, look at people trying to export 'Western' ideas to the middle or far east. They're usually not able to construct infectious ideas in such a way that they aren't cancelled out by existing ideas. This is among people who have the same wetware running their consciousness. Now imagine trying to design such an idea that would spread among a civilisation whose brain-equvialents work completely differently to yours. I don't claim that it's impossible, but I do claim that it would need to be targeted and would not be possible to design without some information about the targets.
My $200 Moto G has 32GB of onboard flash memory, and still has an SD slot. I never even get close to using all of my storage.
Mine is an earlier generation and has even less, but the current iPhones on sale have 64GB to 256GB. 32GB is a lot for me - the largest things I download to my phone are vector maps for OSMAnd - but it's pretty small for someone who watches a lot of videos on their phone. An SD card means that they can think 'well, a 64GB SD card is cheap, so I can always add a load more storage for videos'. I also suspect that a lot of people don't really have a good feel for how much they use and will just see that 32GB is a lot smaller than 256GB and expect that no one would sell a phone with 256GB if people didn't actually use 256GB of space and so that they will probably need that much too.
That's increasingly untrue. For example, if I get a train ticket I can either collect it physically, pay to have it physically posted to me, or use a smartphone app with the barriers. Most airlines now make it very easy to use their app, but much more annoying to get printed boarding passes. RyanAir is particularly bad for this - you can't print your own boarding pass more than 5 days in advance and they charge you to print one for you, so if you're going away for more than 5 days then you need to find a printer somewhere - or just use their smartphone app. My mother doesn't have a smartphone and keeps hitting cases where companies expect it and you have to go through a slow path if you don't have one. In contrast, it's easy to survive without the telephone functionality of a modern phone.
It's possible that we don't understand quantum mechanics properly, but so far it's not been shown that quantum communication violates causality: all messages are propagated at the speed of light (or slower).
USB-C was intended to be significantly more robust than microUSB and the group that designed it included some of the Apple folk behind the Lightning connector, so I doubt it has the same problems.
I'm not too bothered about the dongle - it can live with my headphones - the problem for me is that you need a different and annoying adaptor if you want to listen to music and charge your phone, which is pretty common while travelling.
I suspect it's more the 'isn't noticeably better than the one I have now'. I have a cheap first-gen Moto G. It's crappy, but it does everything I need it to. A newer phone does the same stuff, but marginally faster. My previous phone was an HTC Desire (bravo) and it had such a small amount of on-board flash that it was hard to upgrade. It was also very slow (one core that was about the same speed as the four in my Moto G). That upgrade was worth it, but the next one will probably only happen when the current phone dies. I could afford a new phone, but I don't see much point in getting one.
I suspect that now there are so many high-end smartphones that people can no longer tell them apart: it's not like the first-gen iPhone when there were a handful of large touchscreen phones and owning one was very obvious conspicuous consumption. People who want to buy a shiny gadget to show off have probably moved on to something less mainstream.
Headphone jacks are staying because waterproof ones are expensive, but no one expects a cheap phone to be waterproof. SD cards are staying because you can stick less on-board flash in and convince people that an SD card will let them add a useful amount later (spoiler: it doesn't, and unless you're using really expensive SD cards, expect them to die in normal phone operation, at which point they're likely to give up and buy a new phone). User-replaceable batteries are also going, but it varies a bit. Custom non-replaceable batteries are more expensive than cheap off-the-shelf ones for large-volume runs (fewer parts, but there's a large fixed cost in the battery), so they're going from the high-volume cheap phones but still there for the smaller niche cheap ones.
It's in the same space as flash, perhaps closer to XPoint, but doesn't really have the investment of either to bring it to market in quantity and then doesn't have the economies of scale to make it competitive. Last I heard, the biggest chips were 64MB. It's hard to compete with flash because of the incredibly high volumes of flash that can be used to amortise R&D costs. Last I heard, HP was hoping that flash would run into a wall in scalability and they'd get more investment for alternatives in the same space, but that doesn't seem to have happened yet.
That depends a little bit on the relative energy costs. Imagine that you have a machine that can send matter anywhere in the galaxy, but the energy costs of doing so are huge and proportional to the amount of mass that you send. If you wanted to destroy a planet then the simplest way of doing so would be to drop a few large rocks where they'd fall towards the start and intersect the planet's orbit. Those rocks would have to be pretty huge though, so the energy costs would be very large. Alternatively, you could send a small machine to the planet's orbit, where it would intercept satellite communications, learn how to reproduce them, then start reverse engineering attacks on the information infrastructure and psychology of the planet's inhabitants. The second approach would take a few orders of magnitude less energy. Depending on your mechanism, you might not send anything physical - if you can open a wormhole and send photons through it rather than sending any matter, then that may be even cheaper.
We got applied nuclear fission (and fusion soon after) at almost exactly the same time we got long-range rocketry working. Coincidence?
I enjoy this kind of conspiracy theory, but unfortunately it doesn't really stand up. Getting a working nuclear bomb requires high explosives (in shaped charges) to achieve the critical mass at a density that maintains a chain reaction for long enough. This is basically the same sort of chemistry that you need for rocket propellants. Nuclear fission reactors require materials science able to build the containment vessels, which are very similar to rocket exhaust jets in requirements.
Rockets are very old, it's only the advances in materials sciences that made large human-carrying ones possible. There's a long chain of discoveries going back to the 19th century the led to the discovery of fission, which is easy to achieve (though not to very useful degrees) once you can refine uranium. Refining uranium requires centrifuges that, again, depend on the materials technology to be able to build rapidly spinning things that don't fly apart.
Without the advances in alloys during the first world war, we probably wouldn't have had either rockets or fission in the second world war. As to fusion, once you discover fission is possible then fusion is pretty obvious and a Farnsworth Fusor is fairly easy to build (though building one that's energy positive is, so far, not possible).
Slashdot Mirror
Container is a generic term. Solaris provides a service called Zones, which are based on FreeBSD Jails (but with a few improvements over the older Jails, such as full network-stack virtualisation and support for SysV IPC, both of which I think are now supported in FreeBSD). Linux provides a bunch of services that can be cobbled together to provide more or less the same abstractions.
Yes, the Office is the same - you download the same software and can use it offline.
It's also worth noting that the O365 subscription gives you access to the Windows, Mac, Android and iOS versions. Unfortunately, my employer set it up in a stupid way. If I use the web version then I'm redirected to a web page run by our org that handles the authentication, but for the mobile apps I have to type in my credentials directly to the app (which is then sent to our servers for authentication). This would be fine, except that they don't let me set a different password for that and the systems that give access to a bunch of information that is considered highly confidential and where I'd probably be breaking the law (GDPR) if I provided the password to Microsoft.
The danger of the cloud is that it's a single point of failure and a high-value target. Your 12 guys may not be the most competent in the world, but that doesn't matter unless your data are the most valuable in the world. Someone attacking Azure might not be attacking you, but if they happen to get your data at the same time then selling it is just bonus money for them. Similarly, people are more likely to try to DDoS Azure or AWS than they are your in-house server.
Meltdown allowed you to dump the contents of memory of other VMs on the same nodes as you (including those in Azure Secure Cloud SGX enclaves, if you have access to the preview of that). Microsoft has now rolled out some mitigations, but the vulnerability existed from before the creation of Azure until recently. Anything in multi-tenant systems in that window may have been compromised. What's your plan for preventing that in the future and for dealing with it if it's happened?
The film adaptation was pretty close to the book, thought there were a number of things in the book that were either hard to with the technology of the time or just hard to translate to film entirely (the TV addiction in the book, for example).
I don't think the book aged well. Even at the time of release, microfiche existed and given that TVs were everywhere in their future it was surprising that no one realised that you could put one page of a book in each frame of a TV recording and store books like that, rather than having people memorise them. Or even sneak them in as individual frames in other shows and let everyone else distribute them, but not notice their existence without a special TV: one frame of book every 200 frames would give you one novel every half hour to hour of TV show.
If you don't consider learning from a body of knowledge no human had previously "discovering something new" than you'll have to accept that leaning from a body of knowledge built up by humans over centuries isn't discovering anything new either, and in that case most people never discovery anything new in their lives.
I don't understand your point. In the latter category, you learn from your predecessors but then you add to the body of knowledge that they've created. If In the former category, you learn from your predecessors and add nothing to the body of knowledge they've created. The second is a scientific endeavour, the first is the intellectual equivalent of binge-watching a TV show.
Where would their AI run? If they're sending a physical machine, or they're sending some kind of energy construct that is equivalent to a physical machine, then that physical presence would let them acquire samples. If they expect their AI to run on our computers, then they need to understand how our computers work. To give an analogy, we understand pretty much every possible way of building a steam engine that's possible with Victorian technology, including a bunch that the Victorians didn't come up with, but that doesn't mean that we could tell someone how to sabotage a Victorian-era steam engine site-unseen, because there are a lot of different designs with different failure modes.
That's what I thought, but the ones in my old phone and a tiny laptop that supported SD cards for extra storage both died after a couple of years with very few writes. I think the cheap ones just die over time.
Galaxy S9: £99 + £40/month.
iPhone 7: £79 + £36/month.
iPhone SE: £19 + £18/month.
Galaxy J3: £19 + £14/month.
Alcatel Pixi 3 (never heard of it - really low-end Android phone that's worse than my 4-year-old one): £19 + £7/month.
So the difference between the latest phone and the cheapest currently-made iPhone is the difference between £40/month and £18/month. That's on a 24-month plan, so it works out at £960 versus £432.
The contracts are actually much better value than I thought for the iPhone SE. The cost from Apple is £349, so you're paying about £3.50/month for a plan that gives you 4GB of data and unlimited minutes. The iPhone 7 plan is a little over £20/month more than the cost of the phone direct from Apple, and that plan comes with only 1GB/month.
Their cheapest SIM-only plan is £5/month and gives unlimited texts, but only 200 minutes and 500MB, so if you can sell on the iPhone for close to its retail value then it's best to get the iPhone SE and just keep the SIM.
Not sure about a $1000 laptop. My partner is using my old one, which cost about double that in 2011, because the one that she bought about two year ago for a little under $1,000 is, even after a replacement battery, basically dead (and Dell support is crap). A $1,000 laptop might still work after 3-4 years, but increasingly the low-end laptops are built with cheap components making them effectively disposable devices.
My phone is from 2013 and is not a brick. The one it replaced, from 2010 is, because it isn't supported by LineageOS and so is stuck with a version of the TLS stack that doesn't have updated root certs and doesn't support modern cyphers, so it can't actually connect to most things that use TLS. My phone was not $1,000, it was under $200. It still gets regular third-party updates. To compare a more expensive phone, Samsung Galaxy S4 was released about 4 years ago (though it was $650, not $1,000, at launch). It has a big.LITTLE CPU arrangement, where the LITTLE cluster is the same as the only CPUs in my phone and where the big cluster is a quad-core 1.6GHz Cortex A15 (superscalar, out of order, reasonable performance) - some versions had better CPUs. LineageOS supports it with the latest version, so it can still get security updates. It had 2GB of RAM (twice as much as mine), 16-32GB of Flash (plus an SD card slot that can handle up to an additional 256GB), and a user-replaceable battery (so the thing that's most likely to die can be swapped out). Oh, and a 1080p screen. Oh, and it supports NFC, so it can work with Google Pay if you're willing to trust Google for that kind of thing.
About the only thing that a more modern smartphone has that this lacks is a 64-bit CPU, but Android shows no sign of dropping 32-bit support any time soon and I've yet to see any new 64-bit-only apps (unlike iOS, where 32-bit is effectively dead). I can't think of anything that you'd do with a smartphone that wouldn't work well on an S4 (though perhaps I'm lacking in imagination).
For me, the difference is that I use a laptop for as a primary work computer, so it's easy to justify spending money (ideally, my employer's money) on a decent one.
Google doesn't care if you buy a new phone, they just care if you use in-app payments through Google Play Services (or, ideally, buy new apps through the Play Store) or have their tracking stuff installed. They've actually made Android faster. When I upgraded to the a LineageOS based on the newer AOSP, most of my apps were noticeably faster due to improvements in the ART compiler (which, to be fair, started from a really slow place).
If you can afford someone to carry your phone, why not have them answer it as well, so you don't need to talk to people on the phone.
I suspect it is a matter of degree. If someone presents you with a new discovery that opens up new fields of research, that's exciting. If someone dumps a few thousand years of scientific research on you so that it will take a decade or more of relearning to get to the point where you even understand some of the state of the art, let alone are able to contribute to it, then that's probably disappointing. Would you really enjoy knowing that you would never discover anything new again?
During the Democratic Primary, polls were carried out on the general public asking them to rate their preferences on pairs of R vs D candidates. Clinton was the only one that was not seen as definitely better than Trump. Now, it's possible that another candidate might have slipped in the polls, but none of the others had such a poor starting performance. I suspect that if the ballots had said Trump, Not Clinton, Not Trump, Clinton, then the two most popular choices would have been Not Clinton and Not Trump.
Exactly. A lot of this kind of science fiction plot falls down on basic economics. Oh no, aliens have come to steal our water! Wait, they've invested vast amounts of energy to cross interstellar spaces to get two of the most common elements in the universe from the bottom of a gravity well? They've come to enslave us... so that we can work orders of magnitude less efficiently than their machines. They've come to steal or planet almost makes sense, though unless they evolved on a very similar one they could build space habitats with more landmass that would be more comfortable a lot more easily.
At a sufficiently high level of technology, most resources boil down to energy and that's the thing that you'd use in vast quantities reaching us. I suppose that they might want to come and build a Dyson sphere around the sun to use it as a power source, but then they wouldn't need us.
I guess the only thing that really makes sense is if they wanted our planet and needed some kind of receiver for their invasion ship to slow down. The threat then would be 'build a thing so we can slow down and invade you, or we'll just crash into you at high speed and destroy all life on your world'. Which would probably be phrased as 'please help us slow down, our brakes broke and we have no way of avoiding hitting your world!'
Here's the problem: the danger of ideas is very closely tied to the intelligence that is interpreting them. Look at religions, for an example of self-replicating ideas: even the most successful ones haven't managed to spread to half of the population, and they've evolved in situ. For a counter example, look at people trying to export 'Western' ideas to the middle or far east. They're usually not able to construct infectious ideas in such a way that they aren't cancelled out by existing ideas. This is among people who have the same wetware running their consciousness. Now imagine trying to design such an idea that would spread among a civilisation whose brain-equvialents work completely differently to yours. I don't claim that it's impossible, but I do claim that it would need to be targeted and would not be possible to design without some information about the targets.
My $200 Moto G has 32GB of onboard flash memory, and still has an SD slot. I never even get close to using all of my storage.
Mine is an earlier generation and has even less, but the current iPhones on sale have 64GB to 256GB. 32GB is a lot for me - the largest things I download to my phone are vector maps for OSMAnd - but it's pretty small for someone who watches a lot of videos on their phone. An SD card means that they can think 'well, a 64GB SD card is cheap, so I can always add a load more storage for videos'. I also suspect that a lot of people don't really have a good feel for how much they use and will just see that 32GB is a lot smaller than 256GB and expect that no one would sell a phone with 256GB if people didn't actually use 256GB of space and so that they will probably need that much too.
That's increasingly untrue. For example, if I get a train ticket I can either collect it physically, pay to have it physically posted to me, or use a smartphone app with the barriers. Most airlines now make it very easy to use their app, but much more annoying to get printed boarding passes. RyanAir is particularly bad for this - you can't print your own boarding pass more than 5 days in advance and they charge you to print one for you, so if you're going away for more than 5 days then you need to find a printer somewhere - or just use their smartphone app. My mother doesn't have a smartphone and keeps hitting cases where companies expect it and you have to go through a slow path if you don't have one. In contrast, it's easy to survive without the telephone functionality of a modern phone.
It's possible that we don't understand quantum mechanics properly, but so far it's not been shown that quantum communication violates causality: all messages are propagated at the speed of light (or slower).
USB-C was intended to be significantly more robust than microUSB and the group that designed it included some of the Apple folk behind the Lightning connector, so I doubt it has the same problems.
I'm not too bothered about the dongle - it can live with my headphones - the problem for me is that you need a different and annoying adaptor if you want to listen to music and charge your phone, which is pretty common while travelling.
I suspect it's more the 'isn't noticeably better than the one I have now'. I have a cheap first-gen Moto G. It's crappy, but it does everything I need it to. A newer phone does the same stuff, but marginally faster. My previous phone was an HTC Desire (bravo) and it had such a small amount of on-board flash that it was hard to upgrade. It was also very slow (one core that was about the same speed as the four in my Moto G). That upgrade was worth it, but the next one will probably only happen when the current phone dies. I could afford a new phone, but I don't see much point in getting one.
I suspect that now there are so many high-end smartphones that people can no longer tell them apart: it's not like the first-gen iPhone when there were a handful of large touchscreen phones and owning one was very obvious conspicuous consumption. People who want to buy a shiny gadget to show off have probably moved on to something less mainstream.
Headphone jacks are staying because waterproof ones are expensive, but no one expects a cheap phone to be waterproof. SD cards are staying because you can stick less on-board flash in and convince people that an SD card will let them add a useful amount later (spoiler: it doesn't, and unless you're using really expensive SD cards, expect them to die in normal phone operation, at which point they're likely to give up and buy a new phone). User-replaceable batteries are also going, but it varies a bit. Custom non-replaceable batteries are more expensive than cheap off-the-shelf ones for large-volume runs (fewer parts, but there's a large fixed cost in the battery), so they're going from the high-volume cheap phones but still there for the smaller niche cheap ones.
It's in the same space as flash, perhaps closer to XPoint, but doesn't really have the investment of either to bring it to market in quantity and then doesn't have the economies of scale to make it competitive. Last I heard, the biggest chips were 64MB. It's hard to compete with flash because of the incredibly high volumes of flash that can be used to amortise R&D costs. Last I heard, HP was hoping that flash would run into a wall in scalability and they'd get more investment for alternatives in the same space, but that doesn't seem to have happened yet.
Yup. SourceForge has also been suffering, so I imagine that the 486 in the corner that's handling both sites is now completely full of dust.
That depends a little bit on the relative energy costs. Imagine that you have a machine that can send matter anywhere in the galaxy, but the energy costs of doing so are huge and proportional to the amount of mass that you send. If you wanted to destroy a planet then the simplest way of doing so would be to drop a few large rocks where they'd fall towards the start and intersect the planet's orbit. Those rocks would have to be pretty huge though, so the energy costs would be very large. Alternatively, you could send a small machine to the planet's orbit, where it would intercept satellite communications, learn how to reproduce them, then start reverse engineering attacks on the information infrastructure and psychology of the planet's inhabitants. The second approach would take a few orders of magnitude less energy. Depending on your mechanism, you might not send anything physical - if you can open a wormhole and send photons through it rather than sending any matter, then that may be even cheaper.
We got applied nuclear fission (and fusion soon after) at almost exactly the same time we got long-range rocketry working. Coincidence?
I enjoy this kind of conspiracy theory, but unfortunately it doesn't really stand up. Getting a working nuclear bomb requires high explosives (in shaped charges) to achieve the critical mass at a density that maintains a chain reaction for long enough. This is basically the same sort of chemistry that you need for rocket propellants. Nuclear fission reactors require materials science able to build the containment vessels, which are very similar to rocket exhaust jets in requirements.
Rockets are very old, it's only the advances in materials sciences that made large human-carrying ones possible. There's a long chain of discoveries going back to the 19th century the led to the discovery of fission, which is easy to achieve (though not to very useful degrees) once you can refine uranium. Refining uranium requires centrifuges that, again, depend on the materials technology to be able to build rapidly spinning things that don't fly apart.
Without the advances in alloys during the first world war, we probably wouldn't have had either rockets or fission in the second world war. As to fusion, once you discover fission is possible then fusion is pretty obvious and a Farnsworth Fusor is fairly easy to build (though building one that's energy positive is, so far, not possible).