If they binge on solar, so much the better. It may not generate as much energy as other items per area of land covered, but upkeep on solar generation is very low, and the ecological impact is relatively minor compared to burning fossil fuels. Their use of the technology will get them to make it better, which benefits everyone.
I would be happy if they would return to their pre-Retina MBP engineering, allowing RAM to be upgraded, and a generic M.2 slot instead of Apple's locked down connector. A replaceable battery would be useful as well.
Same with their desktops. Having the ability to replace the HDD or SSD without having to deal with spudgers and hair dryers would be quite useful, especially in environments with a number of Macs.
1TB max internal storage is inexcusable for that machine. Yes, in theory, people are supposed to have FC HBAs and such, but the Mac Pro was designed to be a workstation, to handle high end tasks, so desktop RAID would be useful.
The current cylinder Mac Pro needs to be moved to a place as a midrange workstation, and the full tower brought back as the flagship Mac workhorse. I'd pay $4000 for a full tower with upgradable components. I'd pay $2000-$3000 for a non-upgradable canister Mac that has decent video (2016-era GPUs, not 2012/2013 era), a decent Xeon or i7, 32 GB of RAM, and at least one standard M.2 slot, preferably two, so I can have RAID 1.
I won't pay that much for 3+ year tech that cannot be upgraded in any meaningful manner.
You hit the nail on the head. "Good enough" has knocked Moore's Law off the rails. Since there isn't that much demand, other than adding cores for virtualization [1], it isn't surprising that Intel is backing off the gas pedal with CPU development.
There are other things as well to add to a CPU. Disk I/O hasn't kept up with capacity gains, and there is always working on better power management which is something I'm sure Intel's enterprise customers are heavily damanding for PR reasons.
[1]: The ideal would be faster cores, since Microsoft has hopped on the Oracle and Sybase bandwagon and started licensing by core, and not CPU socket, but more cores is better than nothing.
I know I am a late to this... but Windows Server 2012 and W2012R2 install Server Core by default. W2016 doesn't even give you the option for a GUI until you have the machine installed and are at a PowerShell prompt. Exchange has been using PowerShell for a lot of its configuration for almost a decade now.
Symantec Encryption Desktop is pretty close. You can hit a key, type your key's passphrase, have the mail decrypted in the window. Kpgp is similar on Linux, although you do have to cut and paste with it. I can't see how it can get easier than that.
If it does this, this is useful. I know with Exchange, I ended up setting up TLS connectors manually between sites that were in constant communication with each other. This way, anything going from foo.com to bar.com and back would be encrypted. Having this new standard will make life easier, because making connectors between sites would not be as important.
I just wish someone can address endpoint encryption. TLS has been constantly updated, while S/MIME and OpenPGP are pretty much untouched since their millennial introductions, and endpoint encryption is something that should be considered as a core security tool. Even having S/MIME sign all documents seems like something only I seem to do, but it has saved me in the past come audit time.
Sometimes having the books on Kindle (or in storage) is good. For example, when I was fixing a generator, and the starter decides to just stop working. Pull out smartphone, pull up manual, find a fuse that popped, replace it, good to go. I wouldn't be carrying a physical generator service manual everywhere I go, so being able to tap on a phone, find the part and pull it, was quite nice.
Regular books have their place as well. Best thing is to buy both.
What does this give over the existing protocols, other than using TLS? It looks like once the E-mail is received by the client side, it is stored decrypted, so it only solved a part of the problem.
What is so wrong with getting people to use a standard like S/MIME or OpenPGP, which truly secure messages, regardless if it is in-flight, sitting on a hard disk, or sitting on a spool file on a relay? The advantage of OpenPGP is that it functions independently of the messaging protocol, so security is assured, even if there is no other encryption in any part of the chain, other than the endpoints.
You would be surprised at how inexpensive 3G cards and antennas are. I wouldn't be surprised to find more devices just using that for a constant, unstoppable Internet connection if they can't find a link out.
Or, they can do what modern consoles do. No Internet connection, no worky. You agreed to this, and that all info the device finds, can be given or sold freely by the device maker, in the EULA, when you opened the box.
You can make IoT secure. Devices can be put on separate network segments that can't see each other, are firewalled, with an IDS/IPS in place to minimize damage if compromised. Logs can be exported one way via syslog to a secure server, which can be searched by Splunk or an elk stack machine. Warnings can be handled by an application running locally that can do email or SMS. Hub/spoke architectures can be used with low bandwidth devices using Bluetooth. Heck, most IoT devices could be hardwired. The deadbolt? Many, many buildings have used electric strikes and locks, and that technology is reliable enough for home use. Alarm systems are better hard wired anyway.
However, there is no money to be made by making IoT secure. As mentioned in other/. posts, the mantra, "security has no ROI" thrums loudly among most businesses. The IoT problems are solvable. It is a matter of won't, not can't.
I'm in the same boat. Due to numerous other Wi-Fi links around where I live, at best, I get reliable signal in one room, but that pretty much it. Because there are just so many devices yakking on Wi-Fi, even the 5Ghz band, where devices are supposedly to find the channel that is used the least, are saturated.
As for IoT devices, I do watch occasionally the Fiver channel on YT, which always has some new IoT item. Some are cool, others... why bother? If I were to spend the price premium for a "smart" fridge, I'd buy a refrigerator which runs on CNG or LP gas, as well as electric. Smart deadbolt? I'd like one that can tell me the status, and lock the deadbolt... but mechanically cannot unlock it from remote.
I've never understood why IoT devices don't move to a hub/spoke model. A hardened, central hub that does the Internet communicating, and the devices use Bluetooth and are paired with the hub (or hubs). This way, physical proximity is needed to the devices to had endpoints, and the hub can have IDS/IPS rules to handle compromised endpoint devices. This would go a long ways in solving the IoT security disaster.
I would say that HTC deserves props because they allow one to unlock the bootloader, and even though HTC may not do OS updates, there is always CM, which decently supports devices, and is kept up to date reasonably well. Add GApps and NovaLauncher, and the UI is decent.
I do agree that Nexus is top dog, but at least one can keep HTC devices current with a custom OS without much effort.
Similar with my MBP at home... it has a Thunderbolt to DVI adapter, Thunderbolt to GigE adapter, and the USB ports wind up going to decent powered hubs so I can plug in a keyboard, mouse, external HDD stuff, etc.
How hard would it be for Apple to redesign the MagSafe connector to handle DVI, USB-c, and FireWire? That way, all that is connected is one cord, no formal docking station needed.
Of course, there is the good ol' Dell docking station. Plop the laptop onto that... and it just works without issue. As an added bonus, I can slide a lever, add a Kensington lock, and both the docking station and the laptop are somewhat resistant to walking off. With my MBP, I'd have to have a metal shop fab me a cage for the device, since Apple in their infinite wisdom has decreed Macs immune to theft, so no Kensington lock slot is present on any of their offerings anymore.
One ideal would be a MagSafe-like connector that doesn't just handle power, but has USB-C and Thunderbolt type functionality. This way, it is easy to connect/disconnect, the device has a lot of insertion cycles, and only one connection has to be made for everything, even the external GPU unit.
My question is how a new cable would be handled... would it just be the same voltages as a RJ-45 cable, except terminating in a smaller connector? Will that connector be as easy to splice on, in the field, at 2:00 in the morning, when having to do an emergency upgrade, under the raised tile, while hung over?
We already have technology to deal with putting a RJ connector in a smaller factor. Anyone remember the old PCMCIA modems which had a push-in, push-out connector, called the xJack? I've used those in the field ages ago, and they were rugged enough. If a device doesn't have enough space for the RJ-45's thickness, then why not use that style of connector?
Relatively small numbers though. Each location would need to have at least one tech on staff during business hours. If not, and a machine had some type of fault, that location would lose far more money in lost sales than what would pay the cost of the tech. As per Google, Carl's Jr. has 1385 locations. This means having at least 1385 techs minimum working at each location every hour it was open, who know everything about the mechanism, from what temperatures the burgers are cooked at to making sure the nozzles that are being used for mayo are clean, to dealing with mice and rats that might be in a sack of buns. This can be done, but it will take lots of iterations and work. Heck, basic automobiles didn't get to a decently reliable state for over 100 years since the IC engine was invented.
Those techs won't be cheap either. They may also wind up having to be licensed as electricians and plumbers depending on local regulations (and there will be some cities who will step and require that), further driving up costs. The cost of a master electrician on staff 24/7 can pay for a lot of minimum wage burger flippers.
I have done that in the past, namely using Aladdin eTokens which worked well... but they required hunting down the PKCS drivers, and those were extremely hard to find.
For security, these very nice, especially for TrueCrypt/VeraCrypt volumes, PGP Desktop keys, and such. Because they don't have a USB flash drive partition, they don't work well for NetBackup backup key storage, or keyfile storage for BorgBackup, attic, zbackup, and other utilities. The biggest advantage is the brute force resistance. Someone tries more than several times on the password on the token, the token erases the key and everything on it.
However, my concern about those is longevity. I might have a number of tokens strewn about, but after 10-20 years, will the flash memory on those be readable? Once the electrons escape out the gates, there is no recovery.
Another alternative I've looked at are IronKeys, which mount as USB volumes once one enters the password. Again, useful against brute forcing, but I do worry about needing a recovery file for an archive made 10-15 years ago, not having it.
Then, there is the legal aspect. I am sure, there would be plenty of law firms slavering at the thought of bringing action against a deep-pocketed organization that explicitly said that something needs to be more dangerous.
There is also the fact that if one stands up and tells people that they can't contribute enough to justify a basic level of existence, those people don't have to keep following the rules. It doesn't take much to take a miserable and depressed populace who is told over and over again that they don't matter, and turn them into a violent insurgency.
Personally, if push comes to shove, having a citizen's dividend is a lot cheaper than having to pay for troops, weapons, jails, courts, refugee camps, and other stuff to crack down on crime and domestic terrorism should people realize they have no way of earning any type of wage and feeding their families, and turn to rioting.
What about the worker on site who has to maintain it? Machines and moving parts wear out, sanitation laws must be met, inspections must be done, and so on. There is no way a fast food business can run completely unmanned.
Even if the machines worked perfectly, there will be people trying their best to screw around with them in hopes of getting them to fuck up, then suing the fast food joint for millions because the automated coffee maker didn't realize the cup was yanked out of the slot, or the sliding door was jammed with a stick.
Don't forget sanitation laws. There is a reason why the vending machines that dispensed tea, coffee, hot chocolate, and Jack Daniel's by just pressing a button to have it fling into a cup are all gone. Is the machine going to be able to clean the nozzles and lines, or will there have to be a set of tech people changing out "belts and hoses" so the place meets health standards? To boot, those maintenance workers are not going to work for minimum wage, so payroll costs likely will increase, even though there are fewer people per fast food joint.
A good example of this is the self checkouts at Lowe's and Wal-Mart. I see a lot of people not bothering with them and just waiting for a human. Even then, those checkouts still have issues (nothing like it locking up, ever repeating, "place item in bagging area" until someone can punch in a PIN.)
Robo chefs will be just the same. They will have issues, and there will be need to be people on site to fix them. So, instead of having 3-4 checkers at minimum wage, there is one maintenance tech, who has to pass a licensing exam (as the robotics likely will need an electrician's license or professional certification to work with), who costs more to pay than all the people that were tossed.
Yes, it is easy to toss food service workers... more highly skilled people, not so much, even with the H-1B floodgates tossed wide open. You are not going to find and train someone that skilled for all those locations, and make any money in the long run.
tl;dr... it always sounds cool that the stuff shirts can always threaten to fire the food service workers, but it just means they have to hire a lot more skilled people to maintain the equipment... and a mistake by a tech can cost a fast food joint a lot more, especially if the entire production line is stopped.
What happens in California will just ripple over to Seattle, Portland, and Austin. VCs are VCs, and if people are worried about their jobs in SV, they better be worried in Austin, especially because if companies start laying off, they will be keeping their core offices in California, and axing the satellite offices elsewhere. Austin has a lot of companies, but not enough to stand alone if the economy tanked in California. You can't sell your new ads your new dot.com is making, if most potential buyers are in survival mode or at the courthouse filing bankruptcy paperwork.
In some ways, Austin is just a relatively cheap suburb of SF, or Santa Cruz... except located about 1750 miles away.
Slashdot Mirror
If they binge on solar, so much the better. It may not generate as much energy as other items per area of land covered, but upkeep on solar generation is very low, and the ecological impact is relatively minor compared to burning fossil fuels. Their use of the technology will get them to make it better, which benefits everyone.
I would be happy if they would return to their pre-Retina MBP engineering, allowing RAM to be upgraded, and a generic M.2 slot instead of Apple's locked down connector. A replaceable battery would be useful as well.
Same with their desktops. Having the ability to replace the HDD or SSD without having to deal with spudgers and hair dryers would be quite useful, especially in environments with a number of Macs.
1TB max internal storage is inexcusable for that machine. Yes, in theory, people are supposed to have FC HBAs and such, but the Mac Pro was designed to be a workstation, to handle high end tasks, so desktop RAID would be useful.
The current cylinder Mac Pro needs to be moved to a place as a midrange workstation, and the full tower brought back as the flagship Mac workhorse. I'd pay $4000 for a full tower with upgradable components. I'd pay $2000-$3000 for a non-upgradable canister Mac that has decent video (2016-era GPUs, not 2012/2013 era), a decent Xeon or i7, 32 GB of RAM, and at least one standard M.2 slot, preferably two, so I can have RAID 1.
I won't pay that much for 3+ year tech that cannot be upgraded in any meaningful manner.
You hit the nail on the head. "Good enough" has knocked Moore's Law off the rails. Since there isn't that much demand, other than adding cores for virtualization [1], it isn't surprising that Intel is backing off the gas pedal with CPU development.
There are other things as well to add to a CPU. Disk I/O hasn't kept up with capacity gains, and there is always working on better power management which is something I'm sure Intel's enterprise customers are heavily damanding for PR reasons.
[1]: The ideal would be faster cores, since Microsoft has hopped on the Oracle and Sybase bandwagon and started licensing by core, and not CPU socket, but more cores is better than nothing.
I know I am a late to this... but Windows Server 2012 and W2012R2 install Server Core by default. W2016 doesn't even give you the option for a GUI until you have the machine installed and are at a PowerShell prompt. Exchange has been using PowerShell for a lot of its configuration for almost a decade now.
Symantec Encryption Desktop is pretty close. You can hit a key, type your key's passphrase, have the mail decrypted in the window. Kpgp is similar on Linux, although you do have to cut and paste with it. I can't see how it can get easier than that.
If it does this, this is useful. I know with Exchange, I ended up setting up TLS connectors manually between sites that were in constant communication with each other. This way, anything going from foo.com to bar.com and back would be encrypted. Having this new standard will make life easier, because making connectors between sites would not be as important.
I just wish someone can address endpoint encryption. TLS has been constantly updated, while S/MIME and OpenPGP are pretty much untouched since their millennial introductions, and endpoint encryption is something that should be considered as a core security tool. Even having S/MIME sign all documents seems like something only I seem to do, but it has saved me in the past come audit time.
Sometimes having the books on Kindle (or in storage) is good. For example, when I was fixing a generator, and the starter decides to just stop working. Pull out smartphone, pull up manual, find a fuse that popped, replace it, good to go. I wouldn't be carrying a physical generator service manual everywhere I go, so being able to tap on a phone, find the part and pull it, was quite nice.
Regular books have their place as well. Best thing is to buy both.
Same thing happened to me. I have a Kindle Keyboard that sits plugged in, and apparently, it just autoupdated. No postcard for me.
What does this give over the existing protocols, other than using TLS? It looks like once the E-mail is received by the client side, it is stored decrypted, so it only solved a part of the problem.
What is so wrong with getting people to use a standard like S/MIME or OpenPGP, which truly secure messages, regardless if it is in-flight, sitting on a hard disk, or sitting on a spool file on a relay? The advantage of OpenPGP is that it functions independently of the messaging protocol, so security is assured, even if there is no other encryption in any part of the chain, other than the endpoints.
You would be surprised at how inexpensive 3G cards and antennas are. I wouldn't be surprised to find more devices just using that for a constant, unstoppable Internet connection if they can't find a link out.
Or, they can do what modern consoles do. No Internet connection, no worky. You agreed to this, and that all info the device finds, can be given or sold freely by the device maker, in the EULA, when you opened the box.
You can make IoT secure. Devices can be put on separate network segments that can't see each other, are firewalled, with an IDS/IPS in place to minimize damage if compromised. Logs can be exported one way via syslog to a secure server, which can be searched by Splunk or an elk stack machine. Warnings can be handled by an application running locally that can do email or SMS. Hub/spoke architectures can be used with low bandwidth devices using Bluetooth. Heck, most IoT devices could be hardwired. The deadbolt? Many, many buildings have used electric strikes and locks, and that technology is reliable enough for home use. Alarm systems are better hard wired anyway.
However, there is no money to be made by making IoT secure. As mentioned in other /. posts, the mantra, "security has no ROI" thrums loudly among most businesses. The IoT problems are solvable. It is a matter of won't, not can't.
I'm in the same boat. Due to numerous other Wi-Fi links around where I live, at best, I get reliable signal in one room, but that pretty much it. Because there are just so many devices yakking on Wi-Fi, even the 5Ghz band, where devices are supposedly to find the channel that is used the least, are saturated.
As for IoT devices, I do watch occasionally the Fiver channel on YT, which always has some new IoT item. Some are cool, others... why bother? If I were to spend the price premium for a "smart" fridge, I'd buy a refrigerator which runs on CNG or LP gas, as well as electric. Smart deadbolt? I'd like one that can tell me the status, and lock the deadbolt... but mechanically cannot unlock it from remote.
I've never understood why IoT devices don't move to a hub/spoke model. A hardened, central hub that does the Internet communicating, and the devices use Bluetooth and are paired with the hub (or hubs). This way, physical proximity is needed to the devices to had endpoints, and the hub can have IDS/IPS rules to handle compromised endpoint devices. This would go a long ways in solving the IoT security disaster.
I would say that HTC deserves props because they allow one to unlock the bootloader, and even though HTC may not do OS updates, there is always CM, which decently supports devices, and is kept up to date reasonably well. Add GApps and NovaLauncher, and the UI is decent.
I do agree that Nexus is top dog, but at least one can keep HTC devices current with a custom OS without much effort.
Similar with my MBP at home... it has a Thunderbolt to DVI adapter, Thunderbolt to GigE adapter, and the USB ports wind up going to decent powered hubs so I can plug in a keyboard, mouse, external HDD stuff, etc.
How hard would it be for Apple to redesign the MagSafe connector to handle DVI, USB-c, and FireWire? That way, all that is connected is one cord, no formal docking station needed.
Of course, there is the good ol' Dell docking station. Plop the laptop onto that... and it just works without issue. As an added bonus, I can slide a lever, add a Kensington lock, and both the docking station and the laptop are somewhat resistant to walking off. With my MBP, I'd have to have a metal shop fab me a cage for the device, since Apple in their infinite wisdom has decreed Macs immune to theft, so no Kensington lock slot is present on any of their offerings anymore.
One ideal would be a MagSafe-like connector that doesn't just handle power, but has USB-C and Thunderbolt type functionality. This way, it is easy to connect/disconnect, the device has a lot of insertion cycles, and only one connection has to be made for everything, even the external GPU unit.
My question is how a new cable would be handled... would it just be the same voltages as a RJ-45 cable, except terminating in a smaller connector? Will that connector be as easy to splice on, in the field, at 2:00 in the morning, when having to do an emergency upgrade, under the raised tile, while hung over?
We already have technology to deal with putting a RJ connector in a smaller factor. Anyone remember the old PCMCIA modems which had a push-in, push-out connector, called the xJack? I've used those in the field ages ago, and they were rugged enough. If a device doesn't have enough space for the RJ-45's thickness, then why not use that style of connector?
Relatively small numbers though. Each location would need to have at least one tech on staff during business hours. If not, and a machine had some type of fault, that location would lose far more money in lost sales than what would pay the cost of the tech. As per Google, Carl's Jr. has 1385 locations. This means having at least 1385 techs minimum working at each location every hour it was open, who know everything about the mechanism, from what temperatures the burgers are cooked at to making sure the nozzles that are being used for mayo are clean, to dealing with mice and rats that might be in a sack of buns. This can be done, but it will take lots of iterations and work. Heck, basic automobiles didn't get to a decently reliable state for over 100 years since the IC engine was invented.
Those techs won't be cheap either. They may also wind up having to be licensed as electricians and plumbers depending on local regulations (and there will be some cities who will step and require that), further driving up costs. The cost of a master electrician on staff 24/7 can pay for a lot of minimum wage burger flippers.
I have done that in the past, namely using Aladdin eTokens which worked well... but they required hunting down the PKCS drivers, and those were extremely hard to find.
For security, these very nice, especially for TrueCrypt/VeraCrypt volumes, PGP Desktop keys, and such. Because they don't have a USB flash drive partition, they don't work well for NetBackup backup key storage, or keyfile storage for BorgBackup, attic, zbackup, and other utilities. The biggest advantage is the brute force resistance. Someone tries more than several times on the password on the token, the token erases the key and everything on it.
However, my concern about those is longevity. I might have a number of tokens strewn about, but after 10-20 years, will the flash memory on those be readable? Once the electrons escape out the gates, there is no recovery.
Another alternative I've looked at are IronKeys, which mount as USB volumes once one enters the password. Again, useful against brute forcing, but I do worry about needing a recovery file for an archive made 10-15 years ago, not having it.
Then, there is the legal aspect. I am sure, there would be plenty of law firms slavering at the thought of bringing action against a deep-pocketed organization that explicitly said that something needs to be more dangerous.
There is also the fact that if one stands up and tells people that they can't contribute enough to justify a basic level of existence, those people don't have to keep following the rules. It doesn't take much to take a miserable and depressed populace who is told over and over again that they don't matter, and turn them into a violent insurgency.
Personally, if push comes to shove, having a citizen's dividend is a lot cheaper than having to pay for troops, weapons, jails, courts, refugee camps, and other stuff to crack down on crime and domestic terrorism should people realize they have no way of earning any type of wage and feeding their families, and turn to rioting.
What about the worker on site who has to maintain it? Machines and moving parts wear out, sanitation laws must be met, inspections must be done, and so on. There is no way a fast food business can run completely unmanned.
Even if the machines worked perfectly, there will be people trying their best to screw around with them in hopes of getting them to fuck up, then suing the fast food joint for millions because the automated coffee maker didn't realize the cup was yanked out of the slot, or the sliding door was jammed with a stick.
Don't forget sanitation laws. There is a reason why the vending machines that dispensed tea, coffee, hot chocolate, and Jack Daniel's by just pressing a button to have it fling into a cup are all gone. Is the machine going to be able to clean the nozzles and lines, or will there have to be a set of tech people changing out "belts and hoses" so the place meets health standards? To boot, those maintenance workers are not going to work for minimum wage, so payroll costs likely will increase, even though there are fewer people per fast food joint.
A good example of this is the self checkouts at Lowe's and Wal-Mart. I see a lot of people not bothering with them and just waiting for a human. Even then, those checkouts still have issues (nothing like it locking up, ever repeating, "place item in bagging area" until someone can punch in a PIN.)
Robo chefs will be just the same. They will have issues, and there will be need to be people on site to fix them. So, instead of having 3-4 checkers at minimum wage, there is one maintenance tech, who has to pass a licensing exam (as the robotics likely will need an electrician's license or professional certification to work with), who costs more to pay than all the people that were tossed.
Yes, it is easy to toss food service workers... more highly skilled people, not so much, even with the H-1B floodgates tossed wide open. You are not going to find and train someone that skilled for all those locations, and make any money in the long run.
tl;dr... it always sounds cool that the stuff shirts can always threaten to fire the food service workers, but it just means they have to hire a lot more skilled people to maintain the equipment... and a mistake by a tech can cost a fast food joint a lot more, especially if the entire production line is stopped.
What happens in California will just ripple over to Seattle, Portland, and Austin. VCs are VCs, and if people are worried about their jobs in SV, they better be worried in Austin, especially because if companies start laying off, they will be keeping their core offices in California, and axing the satellite offices elsewhere. Austin has a lot of companies, but not enough to stand alone if the economy tanked in California. You can't sell your new ads your new dot.com is making, if most potential buyers are in survival mode or at the courthouse filing bankruptcy paperwork.
In some ways, Austin is just a relatively cheap suburb of SF, or Santa Cruz... except located about 1750 miles away.