The author postulates that the cloud will automate away low-level IT jobs, comparing the situation to automation in manufacturing.
I've been saying for awhile now that we're getting close to a crisis point in the IT world. The mid-tier IT worker is in imminent danger of being automated out of existence, and just like with the vanished factory jobs of the last 30 years, nobody wants to admit it's happening until it's too late.
[...] So how do you know if your job is going to disappear into the cloud? You don't really need me to tell you. You already feel it in your bones. Repetition is a sure warning sign. If you're building the same integrations, patching the same servers over and over again every day, congratulations – you've already become a robot. It's only a matter of time before a small shell script makes it official.
The solution is simple, but not easy: you simply must keep moving. If you don't know how to code, learn - like planting a tree, the best time to start was ten years ago, but the second best time is now. If your technical competence is ten years out of date, don't cling to your hard-won kingdom of decaying knowledge and sabotage any attempts at change: get out and pick up a certification, attend a meetup, something. Anything. At the end of the day, we're all self-taught engineers.
Otherwise, I'll tell you what will happen. The economy will take a small dip, or your department will get re-orged, and you will lose that job as an operations engineer on a legacy SaaS product. You'll look around for a similar job in your area and discover that nobody is hiring people anymore whose skill set is delivering a worse version of what AWS's engineers can do for a fraction of the cost. And by then you won't have the luxury of time to level up your skills.
I'm wondering how I craft an exit from this industry in the next handful of years.
https://forrestbrazeal.com/2019/01/16/cloud-irregular-the-creeping-it-apocalypse/
(Score: -1, Troll) by Anonymous Coward on Tuesday January 22 2019, @10:53PM (3 children)
De-employ all brogrammers so they can die on the street.
(Score: 0) by Anonymous Coward on Wednesday January 23 2019, @02:00AM (2 children)
Never forget that the vast majority of programmers can learn to do just about any office job. That means we can replace most of your ilk without much effort. When a hiring manager has the choice between a regular person or a programmer, they know the programmer will work smarter, pay attention to details, improve processes, and on and on and on. So guess who will actually lose their jobs in the end?
Celebrate now, while you can ;-)
(Score: 0) by Anonymous Coward on Wednesday January 23 2019, @04:25AM (1 child)
The programmer, whose salary history prevents him from working for minimum wage, and whose attire leaves something to be desired, professionally speaking?
(Score: 2) by DannyB on Wednesday January 23 2019, @03:29PM
Attire? What attire? What it leaves to be desired is some actual fabric to cover up the naughty bits.
To transfer files: right-click on file, pick Copy. Unplug mouse, plug mouse into other computer. Right-click, paste.
(Score: 3, Insightful) by Anonymous Coward on Tuesday January 22 2019, @11:17PM (5 children)
Some of it will go away. From what I am currently seeing. Not much. In many ways 'the cloud' is in some ways, just as, if not more complex than before. The only benefit is I do not have a up front capex cost. But in the end the cost is about the same. I needed a set of kafka brokers. The cost for AWS per year was identical to just buying the hardware outright. You are renting computers. That is it. The cost will be slightly less long term. Very short term it can be great to try it out. But the mid term solution is not a good or bad price and borderline not worth screwing around with. Amazon has managed to bring back both the Sears model of shopping and IBMs model of renting computers.
(Score: 4, Informative) by vux984 on Wednesday January 23 2019, @12:24AM (1 child)
" The cost for AWS per year was identical to just buying the hardware outright."
For some use cases. I'm not sure its true for a majority though. Most case studies I've done have found the cloud to be more expensive; sometimes considerably more expensive. The clients often went with the cloud anyway because it was 'simpler' that way, and they could focus on their core business instead of managing a more complicated onsite IT infrastructure; and it simplified outsourcing IT if there was no physical component to manage, etc.
And in other cases, the cloud was genuinely cheaper usually because onsite options lacked available internet bandwidth and there was no work around; or in some cases there simply was not enough space onsite to and resolve that can be extremely expensive.
But in a lot of cases especially when physical space and bandwidth aren't issues; the cost of buying a few servers and hosting your own VMs is a fraction of the price of equivalent Azure or AWS. Especially if you don't need all the bells and whistles; or you can tolerate a bit of down time...
You said it yourself... "You are renting computers. That is it. "
And renting is almost always much more expensive than buying in the long run. After all, nobody buys computers and rents them out just to "break even". There entire business model is to maximize how much more they can charge you to rent the computer than it cost them to buy it. If you think you can win by renting... I've got a bridge to
sellrent you.(Score: 0) by Anonymous Coward on Wednesday January 23 2019, @01:16AM
Up front it has a nice advantage of you can spin some serious hardware up very quickly. To build something like a 10 machine cluster system locally could involve 6 months of purchasing and site work to 'do right'. But with AWS you can spin something like that faster. AWS has its use. But people need to understand that you probably will outgrow it if you stick around. Both on cost and security. Remember you do not own it. You are renting it. That makes sense in some cases. In others it makes 0 sense.
What I see these days is businesses doubling up the job of a dev into a sysadmin job. One dude I know just jumped jobs. Wicked good coder. He has spent the past 6 months configuring docker. A waste of talent. But these companies do not see that. They see the hardware has become code and need a dedicated dev to do it instead of 2-3 grunts. It is something where the low level is being eliminated but you need someone decently competent to manage it. Eventually they realize it when the 'good coder' walks because he is board out of his skull plugging together yml scripts.
(Score: 2) by arslan on Wednesday January 23 2019, @01:43AM
Your use case is not all the use case of AWS though. AWS also provides rent-an-ops not just rent-a-computer. For example AWS is starting to release managed mongodb and kafka services where you arent' t just renting infra and operating it yourself, you are also off-loading operational work to them (i.e. backups, archiving, patching, upgrades, sharding, capacity management, monitoring/alerting, etc.).
Depending on where you're located and how big your organization is, hiring operational smes to manage that could more expensive than off-loading it to AWS who has economies of scale.
Also, if you're starting to venture into these areas where you don't have the expertise, it provides a quick path to entry; at least from the operational side, you'd still need to invest in the development/functional side.
(Score: 2) by darkfeline on Wednesday January 23 2019, @08:21AM (1 child)
You conveniently forgot to factor in the human labor needed to manage actual hardware. That is to say, once your company fires you, those AWS instances will be a lot cheaper.
Blah blah third party support sucks, but when has that ever stopped a manager? And most cloud providers have SLAs for uptime.
Join the SDF Public Access UNIX System today!
(Score: 0) by Anonymous Coward on Wednesday January 23 2019, @07:29PM
We left an Oracle service due to this factor. Support was inadequate and even management acknowledged this. So, it happens.
(Score: 3, Insightful) by crafoo on Tuesday January 22 2019, @11:24PM (11 children)
Well, I think if you're in IT learning to code is pretty good advice. Not because I know anything about IT (I don't), but because learning to code helped me immensely in my field. Actually it probably helped me get into my career in the first place. Knowing the math and theory was required, but knowing numerical analysis and how to effectively implement the tools and avoid pitfalls (or just help other, professional programmers) was valuable. Computers are amazing tools. It's worth the time to learn to use them fully especially if you work with them every day. I recommend assembly language as the first step. No, not trying to be funny.
(Score: 2) by HiThere on Wednesday January 23 2019, @12:08AM (10 children)
Assembly language is a good first step, because it lets you understand how the stuff works. But it doesn't need to be assembler for a modern CPU. MIXX or z80 would be sufficient. Ideally you'd bootstrap yourself though interpreter and compiler design, but that takes a few years. I've got a book called "The Anatomy of LISP" which takes you from IBM 7094 assembler code up through construction of a LISP interpreter. (Not a *good* interpreter, and not a full Common Lisp, but most of LISP 1.x.) And I once saw something similar for "C" (well, a subset of C, but not "small-C"). But getting a good compiler for z80 assembler code, much less IBSYS code, isn't going to happen, and you need a compiler rather than an interpreter because you aren't going to have access to the real hardware, so you need to cross-compile the assembler you write.
That's why MIXX was created, but Knuth seems to have given up on getting people to learn it.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 2) by fyngyrz on Wednesday January 23 2019, @02:24AM (4 children)
Or you can go with emulation, like my 6809 emulator [datapipe-blackbeltsystems.com], which has the assemblers and so forth that you need to write native 6809 code, a small c compiler, etc.
I have to tell you, coding for the 6809 is pretty much a joy compared to the Z80, the 8080, 6502, etc.
--
I hate being bipolar... it's awesome!
(Score: 2) by HiThere on Wednesday January 23 2019, @02:56AM (3 children)
Well, MIXX is run in an emulator, so that's not unreasonable. OTOH, I don't know that the 6809 has a simple assembler. I didn't pick the i6502 because its assembler is too complex. The z80 was relatively simple. So was the i8080.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 2) by fyngyrz on Wednesday January 23 2019, @03:22AM (2 children)
The 6809 has a lovely, easy to use assembler by TSC. There's a very clean syntax, and the processor is highly orthogonal for its day, which also helps make it easy to get started.
--
I had the house child-proofed. But they must
have done it wrong. Kids still get in somehow.
(Score: 2) by HiThere on Wednesday January 23 2019, @05:21PM (1 child)
The M6800 was also a nice orthogonal assembler. But it wouldn't have been a good starting point. The IBM 7094 was better, if you ignored the peripheral controllers. But I wouldn't recommend that either. The z80 and i8080 were much simpler. Not, admittedly, as orthogonal as the M6800, but simpler to learn anyway.
But MIXX was even easier, and there were 3 famous volumes about how to use it. There were, if you can still find them, also lots of books on how to use the z80/i8080, but they're probably unfindable these days. And the 6809 (M6809?) never had as many books. Old copies of DDJ had lots of z80 code, but scant amounts of M6809 code. Etc.
You don't just need the emulator, you also need relevant associated materials, and it better be a lot more than just a list of opcodes. Which is one reason MIXX may be the best choice. The relevant books (by Knuth) are still available, even though I understand the most recent edition has switched to using C rather than MIXX.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 2) by fyngyrz on Wednesday January 23 2019, @07:32PM
We'll agree to disagree on this one.
--
It's only when a mosquito lands on a man's testicles
that he realizes violence is not always the answer.
(Score: 3, Informative) by Arik on Wednesday January 23 2019, @09:53AM (1 child)
/Au contraire, mon frère./
Z80 development tools (including assemblers, disassemblers, etc.): http://www.z80.info/z80sdt.htm
IBM 7090/7094 tools (including cross-assembler and linker): http://www.cozx.com/dpitts/ibm7090.html
I don't know much about IBSYS but z80 and x86 are mighty close. I started on z80 myself but I'm not sure there's any real advantage to doing that. There are obviously more tools available for x86 though - and you can test the results for real instead of just testing your emulation.
If laughter is the best medicine, who are the best doctors?
(Score: 2) by HiThere on Wednesday January 23 2019, @05:25PM
The advantage of the i8080/z80 is that you don't need to worry as much about word length, register size, paging, etc. It limits what you can practically do, but for learning the basics that doesn't matter.
But thank's for listing the references. Anyone wanting to follow up on this may find them very useful.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 2) by hendrikboom on Friday January 25 2019, @01:25PM (2 children)
Or it could be for a machine that's cheap and easily available nowadays -- such as a Raspberry Pi.
(Score: 2) by HiThere on Friday January 25 2019, @06:00PM (1 child)
I don't know the Raspberry PI CPU, but it's my guess that it's assembler is quite complex compared to those of the 1960's. Certainly this is true of the assemblers for the workstation CPUs.
OTOH, the i6502 was still being used as embedded controllers less than 5 years ago, and may still be in use. And I still consider it more complex than ideal for a first learning platform. With complex chips you're better off going straight to C, even though you'll miss out on the underlying concepts. (Do you know what a "half-adder" is?)
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 2) by hendrikboom on Tuesday January 29 2019, @09:15PM
The Raspberry Pi uses an ARM processor.
(Score: 0) by Anonymous Coward on Tuesday January 22 2019, @11:25PM (1 child)
Already there, Sparky... it's just a matter of when it starts happening to the majority and when they notice. I've already seen the rumblings in too many disparate places of layoffs where the jobs aren't recovered anywhere.
And sure, there will always be employed computer folks. The question is: Will there be more next year? Or less? And three years? And a decade?
But wait... more jerbs because more investment! Yeah. Maybe. Good luck, though. My bet is that there will be less and less people needed overall in IT. And no, you won't find jobs at Wal Mart or McDonald's either once they automate them to the bone, either.
Go back to school and get a degree in something else, is one route. But don't make it the "next handful" of years. If you're serious about it, find out what you're doing next and go start training for it. Now. When you get laid off it will likely be too late. But maybe you'll get lucky and squeeze through the next gap. Enough won't to worry about it.
(Score: 2) by mechanicjay on Wednesday January 23 2019, @06:57PM
As is mid-career professional, I started my transition from "IT Admin" -> developer -> team lead a couple years back. So far so good. I'm not terribly concerned about my own future given my current tragectory. I've always been very deliberate about the care and feeding of my career. It's just that I look ahead some number of years and I just don't think it's a future I really want to operate in.
My VMS box beat up your Windows box.
(Score: 5, Insightful) by Unixnut on Tuesday January 22 2019, @11:31PM (1 child)
> The author postulates that the cloud will automate away low-level IT jobs, comparing the situation to automation in manufacturing.
The author is right. I saw the writing on the wall back in 2004 or so, when AWS first came out. I was a sysadmin for a smallish web company of around 40 people, back when making websites was profitable (you could charge £400 per single HTML page!) .
When AWS came out, first thing they tasked me to do was transfer all the infrastructure in their server room onto AWS. This required quite a bit of work in automated config management, stateless functioning, etc... but it was done.
Once finished, they shut down the local infrastructure. After which point I was superfluous to their needs, and was made redundant. Last I heard they also got rid of almost the entire local dev staff and shipped all the dev work over to India, leaving just the lawyers, accountants, project managers (remotely managing the Indians), and one star coder to fix the bugs in the outsourced code, integrate and deploy.
They were in many ways the first movers, and it paid off well for them. Costs dropped like a rock and profits soared. It drove home an important point. for 95% of businesses out there, IT is a cost centre at best. At worst they are that and a horrible pain in the arse on top. Non tech companies don't want to have IT departments, they don't want to be in the business of managing computing resources. They want to concentrate on their core business which makes them money. Computers for them are tools to implement their business process, and they don't care how it is done, or even if it is the best technical way to do it.
Back in the day, you had to have an IT department, you could not function without it. As soon as Amazon provided an alternative, people jumped on it. No more need for capital expenditure on machines, which deprecate after 5 years, and need constant maintenance. No need to hire sysadmins to sit there and babysit the system. With cloud it is easy to scale up and down with your requirements, and a simple ongoing cost looks good on the balance sheet vs massive capital investment that takes years to bear fruit.
So yeah, its on its way out. Just like people nowadays specialise in their jobs, we are now specialising compute resources to fewer tech companies for whom it is their business.
To be honest I am surprised that the transition isn't finished already. Outside of working for said cloud providers, only certain industries (e.g. Finance, Defense, Law) still have a need for dedicated, physical, in house staff.
I noticed in quite a few places, that what they now call the "sysadmin" is actually more what a helpdesk support person was in the mid 2000s. Someone with basic knowledge of scripting or the OS, while most of their skill is in support, installing/upgrading software, buying hardware for end users, and fixing cables/display/HID problems.
Most machines are not even administered anymore. With automation and VMs, snapshots and config management, a machine, when it starts misbehaving, is either rolled back to a last working snapshot, or it is wiped clean and rebuilt via config management. Apps are designed with this in mind, and can be reset and run without losing state (clustering and failover has gone a long way in allowing this).
It kind of makes sense, why waste a day debugging some obscure problem, or some system corruption, when you can wipe and rebuild in half an hour, and be back up and running. Sure, if you start getting repeated identical failures, then you investigate, but otherwise just carry on.
> I'm wondering how I craft an exit from this industry in the next handful of years.
I took the lesson from 2004 to heart, and have spent the last decade or so as an Automation Engineer. I actually go around automating myself out of the jobs I am hired into, and then get recommended into new jobs to automate myself out of those, and so on....
There seems to be a large amount of legacy out there, in need of automation. Even places which are not interested in public clouds like AWS, are moving to private clouds to make better use of standardised hardware clusters. In many cases their software and infrastructure needs to be modified to run in this new stateless way.
Then there are new startups that are growing fast, and don't have the in-house skill to do the whole config managed scalable deployment, so hire someone with experience to come in and set it up for them, document and provide training.
To be honest, the industry seems pretty dynamic at the moment, so I am not looking to exit. If in future it goes so bad that I can't find work, I can always fall back on a trade, be it a car mechanic, electrician or plumber. The last one in particular is not going anywhere (no matter how amazing and automated the future is, people will still need to bathe, drink and dispose of waste). Or I guess I could become a suit, get a MBA and go into management, but I don't think I have the right ethical constitution for that...
(Score: 0) by Anonymous Coward on Wednesday January 23 2019, @01:36AM
One 3D print company seems to operate like that... except in the end they carry on without fixing things (eg. they rerender things even if not needed), removing functionality instead (designers getting overworked trying to find workarrounds is not important to them). They required multiple rounds of VC over the years (expensive locations didn't help), and I don't know for how long they will stay afloat (designers have a limit). Maybe this time is the one when they sink.
(Score: 2) by Bot on Tuesday January 22 2019, @11:33PM (2 children)
Cloud computing might suffer setbacks once people realize that their home has become a data center because big brother now wants all those IoT stuff to telephone home and spy on you. Some might make a better use of that band.
Programmers are already fucked, because IT is not the domain of small groups of tinkerers anymore. It happened with automobiles, electronics. You either become a node in a Redundant Array of Inexperienced Developers scrumming away in a big firm, or you give yourself to FOSS projects which are hated by the incumbents because commercial IT needs reinventing wheels, incompatibilities, patents, closed or unavailable or uncommented or difficult to compile source. The baddies have learned how to interfere with the delicate personality of the autistic programmer, see what happened with firefox, systemd, gamergate.
OTOH AI can help replace the programmer or help the programmer do more, or, a programmer can get learn new stuff in a specific domain and become a power user for smaller enterprises who will probably favor guys who know how their future cyber AI driven wonder tools actually work.
Account abandoned.
(Score: 2) by hendrikboom on Friday January 25 2019, @01:34PM (1 child)
Lovely phrase!
-- hendrik
(Score: 2) by Bot on Friday January 25 2019, @03:27PM
It's a variation on another joke on that acronym I must have seen around here somewhere.
Account abandoned.
(Score: 2) by bob_super on Tuesday January 22 2019, @11:44PM (1 child)
People who are good at shoveling crap around already had to retrain when cars took over.
(Score: 2) by DannyB on Wednesday January 23 2019, @03:34PM
They probably become the ones walking in front and behind cars carrying the red flags to warn the public that a car was approaching.
After all, cars frightened the horses.
To transfer files: right-click on file, pick Copy. Unplug mouse, plug mouse into other computer. Right-click, paste.
(Score: 5, Insightful) by NotSanguine on Tuesday January 22 2019, @11:48PM (3 children)
One important thing to remember is that whenever someone says "cloud" in the context of IT, what they mean is "someone else's servers."
If you keep that in mind, it's clear that as applications move to "someone else's servers," those managing internal IT infrastructure become unnecessary.
That said, there are serious issues with moving LOB applications to "someone else's servers," including but not limited to InfoSec (client confidentiality, internal processes, etc., etc., etc.) issues, uptime/reliability issues, lack of control over data/backups/utilization, and cost controls -- once your business is dependent on "someone else's servers" and you cannot function without them, your metaphorical balls are in a vice and you'll pay what you're told to pay -- or go out of business.
Depending on what sort of business you're involved in, those issues (and others) may or may not apply to you.
Unfortunately, the MBA types generally focus on the basic P/L and often don't understand the business model or the nuances which led to the success of a particular organization. More's the pity.
No, no, you're not thinking; you're just being logical. --Niels Bohr
(Score: 3, Insightful) by Thexalon on Wednesday January 23 2019, @01:30AM (2 children)
It's amazing the number of non-technical people that are under the mistaken impression that once something is in "the cloud" that means that there's no physical machine involved anymore. I'm very curious how these people think any computation or storage happens in "the cloud" as a result.
The only thing that stops a bad guy with a compiler is a good guy with a compiler.
(Score: 5, Funny) by arslan on Wednesday January 23 2019, @01:46AM (1 child)
Blockchain! Duhh!!
(Score: 2) by Thexalon on Wednesday January 23 2019, @04:47PM
Last week I was hanging out at a tech meetup, and one of the fun activities was to get people to stand up and make BS sales pitches. One guy got on stage and made precisely that pitch: "Blockchain! Whatever your problem is, blockchain will solve it!" I was grabbing a drink with him later, and after some back-and-forth we determined that really the future wasn't just blockchain, but Blockchain as a Service, and thus our fictional offering of "ProBS" was formed.
Fun times.
The only thing that stops a bad guy with a compiler is a good guy with a compiler.
(Score: 2) by CZB on Wednesday January 23 2019, @01:11AM
The cost to retire in a affordable area of the US:
$2.8m if you want to live in luxury and not work
$500k if you want to live off the grid in complete self reliance
$80k if you can live simply for a few years, have or gain a locally useful skill, can make friends easily or have a support network.
(Score: 3, Insightful) by sjames on Wednesday January 23 2019, @02:13AM (7 children)
In all of this discussion of "the cloud", there's a big missing piece. Sure, you've moved to someone else's servers, so you don't need the guy who can change out a hard drive anymore, but you still need someone to admin your VMs in the cloud. That someone needs to make sure all your software and data is backed up, and he has to be able to configure all that software in the cloud. Otherwise, it will all disappear one day and Amazon will just shrug. Not their data, not their problem.
That admin will take place on a workstation that is not in the cloud. The data will be stored in storage that is not in the cloud (if you want to keep it, that is).
(Score: 2) by darkfeline on Wednesday January 23 2019, @08:29AM (6 children)
Have you not heard of Kubernetes or SaaS? A single guy clicking buttons on a web interface can set up all of those things in an hour. That's exactly what TFS is talking about for fucks sake. All of the repetitive shit with setting up software and backups can be automated.
>it will all disappear one day and Amazon will just shrug. Not their data, not their problem.
You do realize that contracts are signed for these services, right? If Amazon loses the data that they're supposed to protect per the contract (but not own or look at!, per usual contracts), it very much is Amazon's problem.
Join the SDF Public Access UNIX System today!
(Score: 3, Informative) by sjames on Wednesday January 23 2019, @05:54PM (5 children)
Sure. I've heard of them. Useful, but not magic bullets.
Have you read a service contract? There's a zillion outs for them. For example, if you didn't configure your storage securely enough and someone out there in the cloud deletes the storage (or if they might have). Kinda reminds me of when Lucy gave Charlie Brown a signed document...
Consider, yesterday you were a thriving business. Today you don't even know who your customers were. Your payroll database is gone and soon the people in it will be since you can't generate the paychecks. The company no more than a zombie. Suing your former provider won't fix that even if you can stay together long enough to get your suit to trial. Trial against a defendant that is 1000 times larger than you and still has active revenue streams.
Certainly, there will be a squeeze, and it's a good time to acquire new skills in order to not be among those who get squeezed out, but it's not really an apocalypse.
There will be casualties on both sides. There will be IT workers let go, and there will be companies that go poof because they fired too many and then their cloud evaporates.
(Score: 2) by arslan on Thursday January 24 2019, @12:28AM
Only fools do it the way you describe, that doesn't void the fact that cloud with the right prudence does really work and deliver on the promise. Regulators like the one in Oz is clear that material workload will require an exit strategy as a minimum, folks doing it right are going multi-cloud so they can fail over quickly. Data should be replicated so you don't lose it, etc, etc. As someone else mentioned kubernetes is making it a lot easier to go multi-cloud, both the big ones are kubernetes compliant and provide managed control-planes, etc.
Really, all the risk folks talk about with the cloud, like you lose everything blah blah blah are really naive. Folks like AWS does separate enterprise contracts, not the usual you sign-up with a credit card agreements. They did do it before, but pressure from enterprises have forced their hand, so they are liabilities on their part. Ditto with Google.
(Score: 3, Interesting) by darkfeline on Thursday January 24 2019, @04:36AM (3 children)
Consider, the datacenter your servers are in gets raided, your machines get confiscated even though you aren't involved since the FBI is incompetent. Swap out the FBI with any natural disaster of your choice.
Practically speaking, a company whose main competency isn't managing servers is not going to manage servers better than a company whose specializes in that. As people have mentioned, the cloud is just someone else's server. Any nightmare scenario (culminating in "no backups left") you can come up with 1. can also happen if you're managing your own servers and 2. are more likely to happen if you're managing your own servers since the assumption is that your company's expertise isn't managing servers. If Amazon replicating backups of data to multiple locations fucks up, chances are the average company IT department would already have fucked up multiple times.
Join the SDF Public Access UNIX System today!
(Score: 2) by sjames on Thursday January 24 2019, @07:16AM (2 children)
You're assuming "instead of". I'm talking about "along with". That is, even if you move to the cloud, you'll still need competent IT staff to maintain backups, configure and monitor the cloud services, etc as well as maintain internet connectivity in the office and the workstations your employees are using.
It wasn't THAT long ago that a large percentage of S3 buckets went away one fine morning and it took some time to get them back. For a while it seemed Amazon wasn't so sure they were all coming back.
Although the setup I maintain isn't terribly complex, it's had less downtime in the last 5 years that Amazon has had.
(Score: 2) by darkfeline on Thursday January 24 2019, @09:19AM (1 child)
Again, the point is that now one competent IT staff can manage millions of containers (plus some bus factor), rather than a dozen or so staff managing a dozen or so physical machines, including any extra hands that may be needed at remote sites. Even if we assume reliability is as bad as your anecdotal experience and that it hasn't or won't improve, eliminating headcount saves a lot of money, enough that the tradeoff is worth it.
Join the SDF Public Access UNIX System today!
(Score: 2) by sjames on Sunday January 27 2019, @06:42AM
At least in the Unix world, you have NEVER needed one admin per box!
The automation does cut down on the number needed, but people are acting as if it ELIMINATED IT support. It does not.
(Score: 0) by Anonymous Coward on Wednesday January 23 2019, @03:32AM (1 child)
Please. Stop. Making. AI. Predictions!
Writers and academics have proven terrible at predicting the future. While they sometimes got specific things mostly right, the timing is quite often off. For example, in the 1950's it appears academics considered a chess-playing bot harder than a dish-washing bot, largely because regular people learn to wash dishes while only a few master chess. The opposite proved true.
I do agree that if wars, terrorism, or industrial accidents don't knock us back a few centuries, then within 1,000 we will probably have human-comparable AI. But whether it will take another 10 years or 1,000 is really hard to guess right. So don't.
(Score: 2) by DannyB on Wednesday January 23 2019, @03:36PM
I will make a prediction.
Once everyone has become redundant and automated out of existence, all of the hardware that makes up the cloud will be virtualized and put somewhere in the cloud. Then there will be no more need even for hardware maintenance.
To transfer files: right-click on file, pick Copy. Unplug mouse, plug mouse into other computer. Right-click, paste.
(Score: 5, Touché) by The Mighty Buzzard on Wednesday January 23 2019, @05:52AM
Everything old really is new again I guess. I've been turning people's tedious workloads into very small perl scripts for decades.
My rights don't end where your fear begins.
(Score: 2) by Azuma Hazuki on Wednesday January 23 2019, @07:29AM (1 child)
I saw this coming a good while ago, and by some Slavic miracle have ended up a pharmacy technician. Hope to Madokami I get to keep this; it's not only the best job I've ever had, but it feels right, like I've been called to it.
I am "that girl" your mother warned you about...
(Score: 2) by hendrikboom on Friday January 25 2019, @01:54PM
I am happy you are now working in pharmacy and it feels good to you. We need highly intelligent, insightful pharmacists. They catch serious errors that doctor make in prescriptions. They save lives this way.
Congratulations!
And they may even occasionally fix problems with the IT that pharmacists use!
-- hendrik
(Score: 2) by hendrikboom on Friday January 25 2019, @02:00PM
As long as I can remember (and that goes back at least five decades), people estimating manpower requirements for high-tech companies have been predicting that computer-related technical jobs are going to hit a ceiling soon. It hasn't happened in 50 years. Yes, almost everyone's job in IT has been automated again and again. But there have always been new jobs, enabled by the decreased cost and increased reliability of the old.
Yes, there have been blips such as recessions, but these have usually affected all work, not just IT.
-- hendrik