im not saying it shouldn't be upgradeable, just that it not being upgradeable doesn't make it non-pro.
True. Even different generations of iMacs and iBooks could be upgraded. Apple even made it easy to flip-up the keyboard to insert RAM, disks, and wireless.
What should worry is that if we can't upgrade RAM and disk or buy at non-Apple pricing. The battery is glued. All together the expected lifetime of the MacBook Pro 2016 is probably reduced from 5 to 3 years. Combined with new pricing and the TCO is then ≈ +80% higher while the environmental impact is +60% compared to previous. All of that to save 2.5 mm!
Nope. The RAM and Hard drive should be the minimally replaceable items. The GPU (MXM board) and CPU are "nice to have" upgradable but for logical reasons involving cooling, they're not likely. Up until SSD's, you wouldn't even think of integrating the hard drive onto the mainboard, because mechanical drives barely last 2 years in a laptop. SSD's push this up to 3 years, but they are not going to last the life of the laptop. All the machines that I have access to that use SSD's haven't had a SSD last longer than 3 years, and when they die, they die terribly, forget recovering data, because the OS will typically be disabled in doing so. This IMO is what that hidden port is for, recovering the SSD data after it dies.
Laptops can be used for 10 years before they are truely useless. People generally hang on to a laptop or desktop for 7 years and get that life by replacing the hard drive after 2-3, and upgrading the ram from the installed base to the maximum.
That said, "Pro" these are not. These at best are Prosumer. They are a far cry from Professional use. They make for nice looking movie props. Functionality-wise even replacing the F-keys is something that is going to get a lot of flack for tone-deafness product designs.
Will people buy it? Of course. But let's not fool ourselves. The reason they sell at all is that Apple products have one of the best build qualities and resale value. If this were Google or Microsoft (neither whom possess any knack for industrial design, only cheap disposable products that they want you to replace every 18 months) it would cost half as much and be only a quarter as capable. I see all these people cooing over the Microsoft Surface Studio Desktop and I'm like .... It's just slightly worse than Wacom's offering, and less stylish.
Apple 's professional products are without capable professional computers.
Three years for the life of an SSD didn't sound very long to me (for any laptop, not just an Apple laptop) so I thought I’d try and see where this figure came from. Here's what I found:
The short version is that you're wrong, but I did find plenty of articles from knowledgeable tech journals stating that since 2012 (the early days of SSD use in Apple laptops), the lifetime of an SSD has far outstripped the useable lifetime of the actual machine.
Here's a report on real-world torture test carried out in 2014:
As a worst case scenario, the drive in the old Macbook Air would be good for 13 years; it's theoretical maximum was 50 years.
These days, consumer-grade SSDs can write in the order of petabytes of information before they fail, and this is from measurements taken years ago.
The problem is that they all fail differently; some go suddenly, and some degrade gracefully over time, depending on the sophistication of the SSD controller. Given that Apple has some expertise in preserving the life of SSDS through write compression and use of its own custom controller then I reckon that your 'three years' figure is woefully off-mark. Try fifteen years plus.
If you want to complain about something, you'd have a better chance claiming that Apple should allow the SSD to be removed so it can be used in your next laptop.
And if your SSD does fail then it shouldn't be a problem if you take backups.
No doubt someone will now chime in and claim that not all professionals can afford to take backups.
You cannot reach conclusions based on the tests of a minute sample of drives and drives can fail for reasons other than read/write exhaustion.
If they had taken six Apple based SSDs it would have been a little more relevant.
Time will tell how drives (and specifically Apple drives) perform.
As I have said, if Apple has conducted lifespan testing based on actual user hardware (and it surely has) and found drives to be an order of magnitude more reliable than anything before it, it would be great if they slapped a 10 year warranty on them. It would be great marketing material too.
And not making them user upgradeable is what people are complaining about. Not to move them up to a new Mac but to swap out any failed drive (for whatever the reason) and put in a higher capacity one.
Why should I have to buy all the internal capacity I might need for the lifetime of the machine and at Apple's prices at the time of purchase? A better approach is to let me upgrade down the line - and at cheaper prices. The same goes for RAM although in my opinion, no Pro machine shipping in 2017 should be shipping with 8GB RAM or just 256 SSD.
With the advent of the cloud, storing more and more off the device, suspect that HD size is as important as it used to be.
Utterly impractical in all but the least demanding workflows. Streaming multiple tracks of audio or video over the internet isn't going to work, so one is looking at wasting precious hours down- and uploading the many GBs of files associated with every project.
Storage space is still just as important as it's always been. Cloud storage is a useful step in the right direction, but it's nowhere near ready for prime time yet.
misa said: [...] Functionality-wise even replacing the F-keys is something that is going to get a lot of flack for tone-deafness product designs.
Speaking only to that specific comment and not the rest of your post, I don't understand what you mean. The Touch Bar can still be used as standard row of F-keys, just like before. It CAN be used for other things but it doesn't HAVE to be.
And "pros" (i.e., people who use their laptop professionally--to earn money) don't generally keep using a 5 or 10 year old laptop. If a "pro" can use a 5 year old laptop, then frankly they probably never needed a "pro" computer in the first place. Therefore pros are the least likely to care that the various internal components are difficult to repair/replace.
Here's something else that pros do: they back up their work so the failure of the 'hard drive' (some of the 'concerned pros' round here seem unable to tell the difference') won't be a complete catastrophe.
A rule of thumb: if the work you're doing doesn't need backing up, or if this machine is worth more than the work you're doing with it, then it isn't meant for you.
Losing data isn't the point. I have backups, but what do I do when the SSD in this new MacBook Pro starts failing in three or four years? I can't replace it. I'll have my files backed up, but will I have to buy a whole new computer to restore them? Or at least be faced with a costly-due-to-complexity third-party repair and the downtime associated with that?
I don't know yet, but I'm at least a little concerned about it. When they're cheap and thus easily replaced it's easy to tolerate the fact that consumer electronics can't be serviced. When the device costs five-thousand bucks like this thing did, it's harder to swallow losing an entire computer to a failed storage system.
With the advent of the cloud, storing more and more off the device, suspect that HD size is as important as it used to be.
Also, curious if it was only my machine, the touch bar got pretty hot, too hot to use really. Anyone else noticing this?
The cloud should be considered a convenience, not a pseudo replacement for local storage. Apple and all the others want us to be cloud dependent so they can sell us more services but most of the planet doesn't have users with the upload speeds that would make cloud computing as fast as local computing. Add to that that ISPs are doing everything they can to bring back data usage plans to replace unlimited plans AND want to charge the likes of Apple, Google etc for providing content that puts a strain on their infrastructure.
The cloud isn't just Amazon and Apple but your local cloud resources.
Pros with high computing requirements have moved to the cloud because it provides far more computing capability than even a high end workstation. That and large datasets that won't fit on anything short of a raid array are available anywhere you can VPN.
The need for high powered local workstions is less than before because the workflow has changed.
Video pros and creatives still need high powered workstations but even there render farms and large editing bays are more than a single box with shared computational load.
When you say 'high computing power' what are referring to exactly? There are many pros that don't even have high speed access to basic cloud services like storage much less computational access.
Like the FEM and FTDT example provided by another poster. You can run them on a local workstation or you can instantiate a VM on modern day "big iron" and do a run for a few bucks on EC2. There is a CAELinux AMI that gives you all the common FEM tools in a prebuilt package.
I haven't supported that sort of thing in over a decade but we used to do them on multi node ROCK clusters using OpenFOAM so still not a single uber workstation workflow. Today you can just do this: http://cfd.direct/cloud/
Do spot purchases of time when the rate is low. It's much more efficient than cobbling together 40 antiquated workstations to form your own cluster. Likewise, its cheaper to put your build farm on EC2 than local workstations and do on demand pricing.
With the advent of the cloud, storing more and more off the device, suspect that HD size is as important as it used to be.
Utterly impractical in all but the least demanding workflows. Streaming multiple tracks of audio or video over the internet isn't going to work, so one is looking at wasting precious hours down- and uploading the many GBs of files associated with every project.
Storage space is still just as important as it's always been. Cloud storage is a useful step in the right direction, but it's nowhere near ready for prime time yet.
Also, curious if it was only my machine, the touch bar got pretty hot, too hot to use really. Anyone else noticing this?
I've barely used it, but the few times I have it hasn't been warm. 15" model.
First, 512GB isn't expensive an upgrade and 2TB, while pricey is available.
Second, with USB-C you have access to external storage at very high speeds. You are better off sneakernetting large assets via external drives between your laptop and desktop than transferring over the wire.
Third, cloud storage include local cloud assets like SAN and NAS.
So really the only existing workflows impacted are folks with small/mid sized projects that normally fit on internal storage with moderate speed needs. That may be a lot of independent pros that are one person shops but not nearly the bulk of pros or pro sales.
And "pros" (i.e., people who use their laptop professionally--to earn money) don't generally keep using a 5 or 10 year old laptop. If a "pro" can use a 5 year old laptop, then frankly they probably never needed a "pro" computer in the first place. Therefore pros are the least likely to care that the various internal components are difficult to repair/replace.
Here's something else that pros do: they back up their work so the failure of the 'hard drive' (some of the 'concerned pros' round here seem unable to tell the difference') won't be a complete catastrophe.
A rule of thumb: if the work you're doing doesn't need backing up, or if this machine is worth more than the work you're doing with it, then it isn't meant for you.
Losing data isn't the point. I have backups, but what do I do when the SSD in this new MacBook Pro starts failing in three or four years? I can't replace it. I'll have my files backed up, but will I have to buy a whole new computer to restore them? Or at least be faced with a costly-due-to-complexity third-party repair and the downtime associated with that?
I don't know yet, but I'm at least a little concerned about it. When they're cheap and thus easily replaced it's easy to tolerate the fact that consumer electronics can't be serviced. When the device costs five-thousand bucks like this thing did, it's harder to swallow losing an entire computer to a failed storage system.
That's what your iMac is for. If your business is so small you only have a MBP then what a Mini or your old laptop is for.
Larger shops have extra equipment to FedEx to a site. Smaller guys better be smart enough not to have a single point of failure whether it's a body, lens or laptop. If LensRentals is your backup for smashing a $5000 lens then a place like meetingtomorrow.com is your same insurance for a $5000 laptop until you figure out a permanent solution.
Note that lenses are not user serviceable either. Or for that matter a D5, Scarlett or Ursa.
With the advent of the cloud, storing more and more off the device, suspect that HD size is as important as it used to be.
Also, curious if it was only my machine, the touch bar got pretty hot, too hot to use really. Anyone else noticing this?
The cloud should be considered a convenience, not a pseudo replacement for local storage. Apple and all the others want us to be cloud dependent so they can sell us more services but most of the planet doesn't have users with the upload speeds that would make cloud computing as fast as local computing. Add to that that ISPs are doing everything they can to bring back data usage plans to replace unlimited plans AND want to charge the likes of Apple, Google etc for providing content that puts a strain on their infrastructure.
The cloud isn't just Amazon and Apple but your local cloud resources.
Pros with high computing requirements have moved to the cloud because it provides far more computing capability than even a high end workstation. That and large datasets that won't fit on anything short of a raid array are available anywhere you can VPN.
The need for high powered local workstions is less than before because the workflow has changed.
Video pros and creatives still need high powered workstations but even there render farms and large editing bays are more than a single box with shared computational load.
When you say 'high computing power' what are referring to exactly? There are many pros that don't even have high speed access to basic cloud services like storage much less computational access.
Like the FEM and FTDT example provided by another poster. You can run them on a local workstation or you can instantiate a VM on modern day "big iron" and do a run for a few bucks on EC2. There is a CAELinux AMI that gives you all the common FEM tools in a prebuilt package.
I haven't supported that sort of thing in over a decade but we used to do them on multi node ROCK clusters using OpenFOAM so still not a single uber workstation workflow. Today you can just do this: http://cfd.direct/cloud/
Do spot purchases of time when the rate is low. It's much more efficient than cobbling together 40 antiquated workstations to form your own cluster. Likewise, its cheaper to put your build farm on EC2 than local workstations and do on demand pricing.
Now give me a rough percentage of pro users that these new MBPs are are aimed at, that have the need of cloud based clusters and whose needs can even be carried out through clusters? The last time I had any involvement with MPI, one of the limitations was that not all problems were suitable for breaking down to be resolved in parallel. I think you will find that only a very, very small percentage of MBP users would benefit from these solutions.
So, glued in battery, not replaceable? I have a mid-2010 MacBook pro which has been a really excellent computer. Have had to replace the battery once, pretty cheap/easy. Machine still works like a champ.
I've also upgraded the internal hard drive. Bummer that you can't upgrade the internal storage.
I think lots of apples design compromises make sense. Not being able to easily upgrade the battery or to be able to upgrade the internal storage at all seems pretty lame. Over the course of a computer's life, both of these will need to be done
All USBc doesn't bother me. Can't tinker doesn't bother me. Price of machine, that bothers me. I wish they were even at bit more reasonably priced.
Agreed. USBc will be fine, cough up a few $$ for dongles - no big deal. The fact that you can't upgrade the internal storage means I would definitely want the 2TB version and that is a crazy expensive option for an already very expensive computer.
Like the FEM and FTDT example provided by another poster. You can run them on a local workstation or you can instantiate a VM on modern day "big iron" and do a run for a few bucks on EC2. There is a CAELinux AMI that gives you all the common FEM tools in a prebuilt package.
I haven't supported that sort of thing in over a decade but we used to do them on multi node ROCK clusters using OpenFOAM so still not a single uber workstation workflow. Today you can just do this: http://cfd.direct/cloud/
Do spot purchases of time when the rate is low. It's much more efficient than cobbling together 40 antiquated workstations to form your own cluster. Likewise, its cheaper to put your build farm on EC2 than local workstations and do on demand pricing.
Now give me a rough percentage of pro users that these new MBPs are are aimed at, that have the need of cloud based clusters and whose needs can even be carried out through clusters? The last time I had any involvement with MPI, one of the limitations was that not all problems were suitable for breaking down to be resolved in parallel. I think you will find that only a very, very small percentage of MBP users would benefit from these solutions.
You can get a high compute mode with high single threaded performance if desired.
I would guess most Pros if the definition of pros is "folks who depend on their computers to make money" as opposed to "folks with the same needs as myself".
Given you can edit 4K on a MacBook (using proxies) there's very few pro tasks that the new MBPs can't handle. Those that it can't (like build farms and multiple VMs) are often low data transfer (upload new source code) far more easily done in a cloud environment if your devops support is circa 2015 vs circa 2005.
The remaining are all extreme and rare edge cases where folks want to natively edit and grade 5K r3d on the road with just a MBP.
4K without proxy is doable on a MBP in FCPX. 4.6K and 4GB video ram becomes the limiter in DaVinci.
So, glued in battery, not replaceable? I have a mid-2010 MacBook pro which has been a really excellent computer. Have had to replace the battery once, pretty cheap/easy. Machine still works like a champ.
I've also upgraded the internal hard drive. Bummer that you can't upgrade the internal storage.
I think lots of apples design compromises make sense. Not being able to easily upgrade the battery or to be able to upgrade the internal storage at all seems pretty lame. Over the course of a computer's life, both of these will need to be done
Replaceable, yes. User replaceable, no. The problem is that to replace the glued on battery you will also have to replace the casing it is glued to. Under warranty, that really isn't such a big deal. Out of warranty it will be anywhere from a costly headache to an utter nightmare.
nht said: I would guess most Pros if the definition of pros is "folks who depend on their computers to make money" as opposed to "folks with the same needs as myself".
So, glued in battery, not replaceable? I have a mid-2010 MacBook pro which has been a really excellent computer. Have had to replace the battery once, pretty cheap/easy. Machine still works like a champ. I've also upgraded the internal hard drive. Bummer that you can't upgrade the internal storage. I think lots of apples design compromises make sense. Not being able to easily upgrade the battery or to be able to upgrade the internal storage at all seems pretty lame. Over the course of a computer's life, both of these will need to be done
Replaceable, yes. User replaceable, no. The problem is that to replace the glued on battery you will also have to replace the casing it is glued to. Under warranty, that really isn't such a big deal. Out of warranty it will be anywhere from a costly headache to an utter nightmare.
You're complaining from total ignorance, I'm afraid.
Ever since Apple's batteries became non-user-replacable seven or eight years ago, they have offered battery replacement via authorised service providers for $129 (plus taxes) for all models regardless of age, and $199 for the glued-in retina MacBooks Pro (since 2012).
Like the FEM and FTDT example provided by another poster. You can run them on a local workstation or you can instantiate a VM on modern day "big iron" and do a run for a few bucks on EC2. There is a CAELinux AMI that gives you all the common FEM tools in a prebuilt package.
I haven't supported that sort of thing in over a decade but we used to do them on multi node ROCK clusters using OpenFOAM so still not a single uber workstation workflow. Today you can just do this: http://cfd.direct/cloud/
Do spot purchases of time when the rate is low. It's much more efficient than cobbling together 40 antiquated workstations to form your own cluster. Likewise, its cheaper to put your build farm on EC2 than local workstations and do on demand pricing.
Now give me a rough percentage of pro users that these new MBPs are are aimed at, that have the need of cloud based clusters and whose needs can even be carried out through clusters? The last time I had any involvement with MPI, one of the limitations was that not all problems were suitable for breaking down to be resolved in parallel. I think you will find that only a very, very small percentage of MBP users would benefit from these solutions.
You can get a high compute mode with high single threaded performance if desired.
I would guess most Pros if the definition of pros is "folks who depend on their computers to make money" as opposed to "folks with the same needs as myself".
Given you can edit 4K on a MacBook (using proxies) there's very few pro tasks that the new MBPs can't handle. Those that it can't (like build farms and multiple VMs) are often low data transfer (upload new source code) far more easily done in a cloud environment if your devops support is circa 2015 vs circa 2005.
The remaining are all extreme and rare edge cases where folks want to natively edit and grade 5K r3d on the road with just a MBP.
4K without proxy is doable on a MBP in FCPX. 4.6K and 4GB video ram becomes the limiter in DaVinci.
As I said, a very, very small amount of pros. As I also said, a very small amount of pros have access to fast up/downloads. Most have only fast downloads.
As for the definition of 'pros' I would say most try to make money through using their computer. My needs though are irrelevant. We are generalising but would you try to put a percentage on the amount of MBP users that need access to cluster based cloud computing?
Would less than one percent be a reasonable guess?
Like the FEM and FTDT example provided by another poster. You can run them on a local workstation or you can instantiate a VM on modern day "big iron" and do a run for a few bucks on EC2. There is a CAELinux AMI that gives you all the common FEM tools in a prebuilt package.
I haven't supported that sort of thing in over a decade but we used to do them on multi node ROCK clusters using OpenFOAM so still not a single uber workstation workflow. Today you can just do this: http://cfd.direct/cloud/
Do spot purchases of time when the rate is low. It's much more efficient than cobbling together 40 antiquated workstations to form your own cluster. Likewise, its cheaper to put your build farm on EC2 than local workstations and do on demand pricing.
Now give me a rough percentage of pro users that these new MBPs are are aimed at, that have the need of cloud based clusters and whose needs can even be carried out through clusters? The last time I had any involvement with MPI, one of the limitations was that not all problems were suitable for breaking down to be resolved in parallel. I think you will find that only a very, very small percentage of MBP users would benefit from these solutions.
You can get a high compute mode with high single threaded performance if desired.
I would guess most Pros if the definition of pros is "folks who depend on their computers to make money" as opposed to "folks with the same needs as myself".
Given you can edit 4K on a MacBook (using proxies) there's very few pro tasks that the new MBPs can't handle. Those that it can't (like build farms and multiple VMs) are often low data transfer (upload new source code) far more easily done in a cloud environment if your devops support is circa 2015 vs circa 2005.
The remaining are all extreme and rare edge cases where folks want to natively edit and grade 5K r3d on the road with just a MBP.
4K without proxy is doable on a MBP in FCPX. 4.6K and 4GB video ram becomes the limiter in DaVinci.
As I said, a very, very small amount of pros. As I also said, a very small amount of pros have access to fast up/downloads. Most have only fast downloads.
As for the definition of 'pros' I would say most try to make money through using their computer. My needs though are irrelevant. We are generalising but would you try to put a percentage on the amount of MBP users that need access to cluster based cloud computing?
Would less than one percent be a reasonable guess?
No. 1% is a very poor guess.
I would say that 90% of the developers complaining about the 16gb limit is better served with better devops that leverages a cloud infrastructure. Unit test locally then push the code to the test branch (or trunk...whatever floats your boat) and it gets deployed into a test environment located in the (private or public) cloud. Any differences between your local dev environment and the deployed environment is exposed and testing is done with a more representative data set. None of this needs fast uploads or downloads. Just sufficient bandwidth to SSH or RDP into your test environment.
Likewise the majority of folks with high computation needs but modest input values are better served with using either a private or public cloud infrastructure rather than single high performance laptop. Most modern tools of this kind are already built for distributed computing across nodes. Even those apps that cannot be used across multiple nodes are better served with a high performance instance than a 4 core local laptop. A single c4.xlarge instance $0.312 per hour using the current spot instance pricing. That's a Xeon E5-2666 v3 3.5 Ghz instance with 18 cores and 60GB of RAM.
Folks with both high computational needs and large local data sets (like 4K video editing) are better served by desktops and not laptops. They are not served well by either the cloud or the laptop.
So who's left? A small subset of use cases that have high computational needs, large local data sets and need to do so in the field where it is more efficient to do so on a small 15" screen rather than shipping a 27" iMac in a Pelican case to the site[1].
Your 1% estimate is more applicable to folks that need 32GB and more than 2TB internal storage than the general pro use of MBPs.
--
[1] Having helped a few local pro photographer out I've seen them bring the iMac so potential buyers (aka parents) can see the photos of the shoot (ballet, etc) on a bigger screen and buy them. Which is where they make the majority of their money.
Like the FEM and FTDT example provided by another poster. You can run them on a local workstation or you can instantiate a VM on modern day "big iron" and do a run for a few bucks on EC2. There is a CAELinux AMI that gives you all the common FEM tools in a prebuilt package.
I haven't supported that sort of thing in over a decade but we used to do them on multi node ROCK clusters using OpenFOAM so still not a single uber workstation workflow. Today you can just do this: http://cfd.direct/cloud/
Do spot purchases of time when the rate is low. It's much more efficient than cobbling together 40 antiquated workstations to form your own cluster. Likewise, its cheaper to put your build farm on EC2 than local workstations and do on demand pricing.
Now give me a rough percentage of pro users that these new MBPs are are aimed at, that have the need of cloud based clusters and whose needs can even be carried out through clusters? The last time I had any involvement with MPI, one of the limitations was that not all problems were suitable for breaking down to be resolved in parallel. I think you will find that only a very, very small percentage of MBP users would benefit from these solutions.
You can get a high compute mode with high single threaded performance if desired.
I would guess most Pros if the definition of pros is "folks who depend on their computers to make money" as opposed to "folks with the same needs as myself".
Given you can edit 4K on a MacBook (using proxies) there's very few pro tasks that the new MBPs can't handle. Those that it can't (like build farms and multiple VMs) are often low data transfer (upload new source code) far more easily done in a cloud environment if your devops support is circa 2015 vs circa 2005.
The remaining are all extreme and rare edge cases where folks want to natively edit and grade 5K r3d on the road with just a MBP.
4K without proxy is doable on a MBP in FCPX. 4.6K and 4GB video ram becomes the limiter in DaVinci.
As I said, a very, very small amount of pros. As I also said, a very small amount of pros have access to fast up/downloads. Most have only fast downloads.
As for the definition of 'pros' I would say most try to make money through using their computer. My needs though are irrelevant. We are generalising but would you try to put a percentage on the amount of MBP users that need access to cluster based cloud computing?
Would less than one percent be a reasonable guess?
No. 1% is a very poor guess.
I would say that 90% of the developers complaining about the 16gb limit is better served with better devops that leverages a cloud infrastructure. Unit test locally then push the code to the test branch (or trunk...whatever floats your boat) and it gets deployed into a test environment located in the (private or public) cloud. Any differences between your local dev environment and the deployed environment is exposed and testing is done with a more representative data set. None of this needs fast uploads or downloads. Just sufficient bandwidth to SSH or RDP into your test environment.
Likewise the majority of folks with high computation needs but modest input values are better served with using either a private or public cloud infrastructure rather than single high performance laptop. Most modern tools of this kind are already built for distributed computing across nodes. Even those apps that cannot be used across multiple nodes are better served with a high performance instance than a 4 core local laptop. A single c4.xlarge instance $0.312 per hour using the current spot instance pricing. That's a Xeon E5-2666 v3 3.5 Ghz instance with 18 cores and 60GB of RAM.
Folks with both high computational needs and large local data sets (like 4K video editing) are better served by desktops and not laptops. They are not served well by either the cloud or the laptop.
So who's left? A small subset of use cases that have high computational needs, large local data sets and need to do so in the field where it is more efficient to do so on a small 15" screen rather than shipping a 27" iMac in a Pelican case to the site[1].
Your 1% estimate is more applicable to folks that need 32GB and more than 2TB internal storage than the general pro use of MBPs.
--
[1] Having helped a few local pro photographer out I've seen them bring the iMac so potential buyers (aka parents) can see the photos of the shoot (ballet, etc) on a bigger screen and buy them. Which is where they make the majority of their money.
'High computation, modest input values'? I can think of a few fields where that may be necessary but mostly in science, education, engineering etc, but in those cases the machine you have locally acts merely as a console. As such, most modern PCs would do just fine. Even some older ones if institutions allowed them. More often than not though, older systems are forcibly removed from the pool on security grounds. Either way you certainly wouldn't need a MBP for those tasks.
For what you are suggesting, I'll stick with my estimate of less than 1%.
<(2) utilitary appliences designed usable life is different, compare to computers. actually, over the years, the designed usable life went down a lot... due to both user behaviour, software requirement, and technology roadmap (moores law of 18 month got a lot to do with it). Use to be 3 years warrantee with 5 years zero field return, currently the warrantee dropped to 6 month... material selection on the durability changed accordingly. price reflect it as well - even pricy mac, price per function and function per weight still out there as good buy." "2. Sorry but I didn't understand much of that." product, depend upon its application got a defined usable life.... wash machine usable life is normally expected to be longer than laptop...computer used to be long usable life, because it was pricy (mainframe)... that all changed when pc take over. Intel double the transistor counts (moores law) every 18 month of IC put product in acceleration to be obsolete faster... that impact the usable life..(the software utilize the 18 month cycle crank out more memory hungry and faster software- 32 bit vs 64 bit)... the price went down great deal. the designer expect you use it for say 3-5 years (upgrade every 2 generation or 3 of the hardware) instead of 10 years... the components and material selection start from the design reflect the intend product life (replacement part, usually are designed to be plug and play by repair house, but not the other part). washing machine with a computer is a new animal regarding the usable life mismatch... TI got some comment on it. more see book "product integrity and reliability in design". hopefully, it helps. IMHO ;-)
Thank you for explaining. I think we are talking about different subjects. I am not referring to the life of the machine but some of its components. In that respect it doesn't matter if the machine is an electrical kitchen appliance or a computer.
If the manufacturer is so confident with the longevity of a particular component so as to offer a specific warranty for that component, it clearly believes that the marketing weight of that warranty will be relevant. For the user it means peace of mind. A win, win situation for all involved.
If these latest generation SSDs are designed to far outlive the expected lifespan of the computer, Apple could offer SSD specific warranties and Gove users peace of mind. Of course it won't, but that isn't the point really.
'High computation, modest input values'? I can think of a few fields where that may be necessary but mostly in science, education, engineering etc, but in those cases the machine you have locally acts merely as a console. As such, most modern PCs would do just fine. Even some older ones if institutions allowed them. More often than not though, older systems are forcibly removed from the pool on security grounds. Either way you certainly wouldn't need a MBP for those tasks.
For what you are suggesting, I'll stick with my estimate of less than 1%.
Well, if you want to claim that only 1% of pro users need the stuff you claim is required I guess that's fine. Especially since you have been unwilling or unable to provide any use cases where the 2016 MBP falls short and the cloud isn't a viable alternative that isn't 4K+ video related.
Comments
If they had taken six Apple based SSDs it would have been a little more relevant.
Time will tell how drives (and specifically Apple drives) perform.
As I have said, if Apple has conducted lifespan testing based on actual user hardware (and it surely has) and found drives to be an order of magnitude more reliable than anything before it, it would be great if they slapped a 10 year warranty on them. It would be great marketing material too.
And not making them user upgradeable is what people are complaining about. Not to move them up to a new Mac but to swap out any failed drive (for whatever the reason) and put in a higher capacity one.
Why should I have to buy all the internal capacity I might need for the lifetime of the machine and at Apple's prices at the time of purchase? A better approach is to let me upgrade down the line - and at cheaper prices. The same goes for RAM although in my opinion, no Pro machine shipping in 2017 should be shipping with 8GB RAM or just 256 SSD.
Storage space is still just as important as it's always been. Cloud storage is a useful step in the right direction, but it's nowhere near ready for prime time yet.
I've barely used it, but the few times I have it hasn't been warm. 15" model.
Speaking only to that specific comment and not the rest of your post, I don't understand what you mean. The Touch Bar can still be used as standard row of F-keys, just like before. It CAN be used for other things but it doesn't HAVE to be.
Losing data isn't the point. I have backups, but what do I do when the SSD in this new MacBook Pro starts failing in three or four years? I can't replace it. I'll have my files backed up, but will I have to buy a whole new computer to restore them? Or at least be faced with a costly-due-to-complexity third-party repair and the downtime associated with that?
I don't know yet, but I'm at least a little concerned about it. When they're cheap and thus easily replaced it's easy to tolerate the fact that consumer electronics can't be serviced. When the device costs five-thousand bucks like this thing did, it's harder to swallow losing an entire computer to a failed storage system.
I haven't supported that sort of thing in over a decade but we used to do them on multi node ROCK clusters using OpenFOAM so still not a single uber workstation workflow. Today you can just do this: http://cfd.direct/cloud/
Do spot purchases of time when the rate is low. It's much more efficient than cobbling together 40 antiquated workstations to form your own cluster. Likewise, its cheaper to put your build farm on EC2 than local workstations and do on demand pricing.
Second, with USB-C you have access to external storage at very high speeds. You are better off sneakernetting large assets via external drives between your laptop and desktop than transferring over the wire.
Third, cloud storage include local cloud assets like SAN and NAS.
So really the only existing workflows impacted are folks with small/mid sized projects that normally fit on internal storage with moderate speed needs. That may be a lot of independent pros that are one person shops but not nearly the bulk of pros or pro sales.
That's what your iMac is for. If your business is so small you only have a MBP then what a Mini or your old laptop is for.
Larger shops have extra equipment to FedEx to a site. Smaller guys better be smart enough not to have a single point of failure whether it's a body, lens or laptop. If LensRentals is your backup for smashing a $5000 lens then a place like meetingtomorrow.com is your same insurance for a $5000 laptop until you figure out a permanent solution.
Note that lenses are not user serviceable either. Or for that matter a D5, Scarlett or Ursa.
I would guess most Pros if the definition of pros is "folks who depend on their computers to make money" as opposed to "folks with the same needs as myself".
Given you can edit 4K on a MacBook (using proxies) there's very few pro tasks that the new MBPs can't handle. Those that it can't (like build farms and multiple VMs) are often low data transfer (upload new source code) far more easily done in a cloud environment if your devops support is circa 2015 vs circa 2005.
The remaining are all extreme and rare edge cases where folks want to natively edit and grade 5K r3d on the road with just a MBP.
4K without proxy is doable on a MBP in FCPX. 4.6K and 4GB video ram becomes the limiter in DaVinci.
Ever since Apple's batteries became non-user-replacable seven or eight years ago, they have offered battery replacement via authorised service providers for $129 (plus taxes) for all models regardless of age, and $199 for the glued-in retina MacBooks Pro (since 2012).
https://support.apple.com/mac-notebooks/repair/service/pricing
As for the definition of 'pros' I would say most try to make money through using their computer. My needs though are irrelevant. We are generalising but would you try to put a percentage on the amount of MBP users that need access to cluster based cloud computing?
Would less than one percent be a reasonable guess?
I would say that 90% of the developers complaining about the 16gb limit is better served with better devops that leverages a cloud infrastructure. Unit test locally then push the code to the test branch (or trunk...whatever floats your boat) and it gets deployed into a test environment located in the (private or public) cloud. Any differences between your local dev environment and the deployed environment is exposed and testing is done with a more representative data set. None of this needs fast uploads or downloads. Just sufficient bandwidth to SSH or RDP into your test environment.
Likewise the majority of folks with high computation needs but modest input values are better served with using either a private or public cloud infrastructure rather than single high performance laptop. Most modern tools of this kind are already built for distributed computing across nodes. Even those apps that cannot be used across multiple nodes are better served with a high performance instance than a 4 core local laptop. A single c4.xlarge instance $0.312 per hour using the current spot instance pricing. That's a Xeon E5-2666 v3 3.5 Ghz instance with 18 cores and 60GB of RAM.
Folks with both high computational needs and large local data sets (like 4K video editing) are better served by desktops and not laptops. They are not served well by either the cloud or the laptop.
So who's left? A small subset of use cases that have high computational needs, large local data sets and need to do so in the field where it is more efficient to do so on a small 15" screen rather than shipping a 27" iMac in a Pelican case to the site[1].
Your 1% estimate is more applicable to folks that need 32GB and more than 2TB internal storage than the general pro use of MBPs.
--
[1] Having helped a few local pro photographer out I've seen them bring the iMac so potential buyers (aka parents) can see the photos of the shoot (ballet, etc) on a bigger screen and buy them. Which is where they make the majority of their money.
For what you are suggesting, I'll stick with my estimate of less than 1%.
"2. Sorry but I didn't understand much of that."
product, depend upon its application got a defined usable life.... wash machine usable life is normally expected to be longer than laptop...computer used to be long usable life, because it was pricy (mainframe)... that all changed when pc take over. Intel double the transistor counts (moores law) every 18 month of IC put product in acceleration to be obsolete faster... that impact the usable life..(the software utilize the 18 month cycle crank out more memory hungry and faster software- 32 bit vs 64 bit)... the price went down great deal. the designer expect you use it for say 3-5 years (upgrade every 2 generation or 3 of the hardware) instead of 10 years... the components and material selection start from the design reflect the intend product life (replacement part, usually are designed to be plug and play by repair house, but not the other part). washing machine with a computer is a new animal regarding the usable life mismatch... TI got some comment on it. more see book "product integrity and reliability in design". hopefully, it helps. IMHO ;-)
If the manufacturer is so confident with the longevity of a particular component so as to offer a specific warranty for that component, it clearly believes that the marketing weight of that warranty will be relevant. For the user it means peace of mind. A win, win situation for all involved.
If these latest generation SSDs are designed to far outlive the expected lifespan of the computer, Apple could offer SSD specific warranties and Gove users peace of mind. Of course it won't, but that isn't the point really.