What are you going to do when the tech bubble bursts, and suddently that CS degree you got because your mom forced you to go to the top college in your area, screaming at you that "You have to be better than your father!", and would only be followed by work, disappointment, fustration, shame, and heartbreak because your only girlfriend in the world, the only person who could ever get past your innate ability to run off anybody within a 5 mil. radius because you said the wrong thing at the wrong time, died while trying practice for some stupid-ass swimming competition she never wanted to do anyway either, and you know she never wanted to do it because you seen how her father talked to her, you seen how he never gives her anything but shit, yet you can't say anything because you don't want to lose her, and even still he never liked you, so now he won't even let you go to the funeral to say a proper goodbye, but why should you be allowed there anyway, because it's not like you tried to stop her from dying, you knew you should of went there with her, but you didn't because you had to fucking """study"""?
do you think tech will be cheaper, or more expensive?
>tech bubble
>bursting
Nigga it just started, IT is everywhere nowadays, just wait for all those selfdriving cars, computer controlled toilets with automatic fecal scans and networked fuckdolls with brainwave VR headpieces
like I'll ever have a gf
>>55545459
What suddenly happens to my CS degree? Learn how to write a sentence, you fucking idiot
>>55545498
>implying you will be doing any of that shit when intelligent software and AI will do it ten times better for free
>>55545459
I think tech will be fine. Also probably bait but >projecting
>>55545459
>girlfriend
Stopped reading
>>55545459
Get a therapist brah
>CS degree
I have an IS degree
>girlfriend
She's an engineer considering becoming a patent litigator that is guaranteed a job because her dad has is own firm
Nice meme loser
>>55545498
So? The bubble isn't that tech is being used. It's that tech is paying well.
Programming used to be hard. It used to require a lot of higher mathematics and intense precision. That isn't true today; any schmuck who can follow directions can produce useful code. Wages aren't based on how productive you are, they're based on how difficult you are to replace. When software engineers were brilliant men with years of training, they were hard to replace, so their salaries were high. Today, colleges churn out CS grads by the thousand every year. Coding bootcamps can manage even higher throughput, and the graduates from those are usually desperate enough to take whatever they can get. Meanwhile, tens of millions of Asians will happily do any programming job for just over poverty.
Tech wages have been on the decline for a while, and that is not going to stop. It's going to speed up. High-end computer science research will still pay mountains of dosh, but there's already hundreds of applicants for every research position, and the schools keep awarding Ph.Ds. If you're not a literal genius, that won't help you.
Plus, we're at the cusp of a real revolution in the way computers are programmed. Learning machines are emerging; general purpose AIs are coming to life. Most of the big research firms have shifted focus from trying to attract as much talent as possible to trying to acquire as much data as possible, because with sufficient data you can replicate the performance of any expert.
Tech is here to stay, but high salaries in tech are not. All the wealth that tech creates goes to the billionaires that own it, not the engineers that build it. When the billionaires don't need the engineers anymore, they will hang us out to dry. It's already happening. Welcome to the future.
>fell for the oil meme
>oil price crashed and can't find a job
>going back to school for the CS meme
>bubble will probably pop when I graduate
such is the meme life
>>55548121
Should've been a freezer technician :^)
>>55545459
Oh look, another mentally ill jojo faggot.
>>55547932
There were so many assumptions, biases, and general false information in three paragraphs.
>>55548121
Fellow meme life brother.
>>55548121
Your problem is you keep getting into fields when they've already peaked. That's a trap for dipshits (like you), the only way to make big money off of a skill is to get in when it's still rising, not when it's stagnating or rotting away.
Stop following meme careers.
>>55548283
The oil industry isn't a meme, he just has bad timing. And the software industry is still in the infant stages.
>>55547130
But when you get that advanced tech, new things need to be taken care of, so new job spots open in 'tech' industry.
>>55548169
Lol no joke my old boss paid $3000 to repair one of the foodservice fridges.
>>55548323
>The oil industry isn't a meme
Point at him and laugh!
>>55548425
Freezer Technician is so underrated a job, people forget that food is essential to living.
I'm going to study to be one the minute i finish this CS degree
>>55548391
> He's never heard of self-repairing machines
We can automate routine maintenance. There'll still be room for highly skilled technician work, but that only requires one or two people in place of several hundred. And then what happens when we can automate complex maintenance too? What happens when we build AIs that are better at building AIs than we are?
>>55548482
I bet you couldn't even if you wanted too, those print money niche jobs would be passed down father to son like elevator repair
>>55545459
there seems to be a lot of pent up anger here.
>>55548169
anything airconditioner, refridgeration, cable t.v. (and its alternatives) pays really well
>tfw didn't go to college
>tfw I am a manager over people with bachelors degrees now
>tfw have to train people with CS degrees on how to actually do real-life tasks
>tfw make more money than CS college graduates
lol
>>55545459
>you seen
Nigger
>>55548567
>nigger
>degree
your autistic to believe that
>>55548489
That's not going to happen. There is no general AI and never will be until we develop a mathematical model that gives an Intel 8080 the intelligence of a cat.
>>55548573
>implying he isn't an oreo
>implying you don't see those in tech fields
>>55545459
>Implying the bubble will burst
There will always be a talent shortage because technology is hard for normies and every business needs technology. That is, unless they start teaching Elements of Computing Systems in Kindergarten alongside Reading and Social Studies. But we don't fund schools in this country because that's soshulizum.
>>55548489
If we get to the level of advancement of Star Trek replicators and self-replicating nanomachines constructing our Death Stars and Dyson Spheres then materialism and money is redundant anyway because everything will be virtually free.
>>55548608
you have me there, I withdraw my insult
>>55548584
You have no idea what you're talking about. We HAVE such a model. We have several dozen, in fact, with various minor advantages/disadvantages between them.
> There is no general AI
This is simply not true. We have many, many machine learning models that are capable of adapting to any problem domain as long as you find a way to structure the data in a way it can use.
Animal intelligence is not magic. We understand the mechanics behind it quite well; if you understand cellular biology at all, look into introductory neuroscience courses online. We could easily create a synthetic cat. It would take a very long time to gather enough experiential data to evolve it to any meaningful way, but once there it would be superior to all existing cats. If artificial life interests you, here's a paper about virtual fish that was published in 1995: http://www.nada.kth.se/kurser/kth/2D1381/ArtificialLifeMaes.pdf
The really earthshaking property of machine learning is that it transforms problem solving from an expertise-based field to an experience-based field. To solve a difficult problem in the conventional way, you need engineers smart enough to break the problem and address every case in a comprehensive way. For most abstract problems, no human alive is smart enough to do that. Machine learning flips the table on that: now, all we need is engineers just smart enough to create an appropriate model that can be trained to solve a problem by learning on its own. It is not a coincidence that around the time GPGPU systems became advanced enough to run very complex graphical models quickly is the same time that all the major CS firms shifted focus from attracting the world's best talent to aggrigating the world's data.
>>55548936
Animal intelligence is not magic. Implementing that intelligence in a "brain" with the computing power of an Intel 8080 is magic.
>>55548965
It's a good thing that it's not 1974 then, isn't it?
>>55548936
What is a major CS firm? I didn't know there were private businesses outside of academia/R&D that did CS research.
>>55548999
The point is that with the models we have today, we cannot adequately simulate human level general AI on any of our computers.
http://motherboard.vice.com/blog/for-one-second-a-supercomputer-mimicked-the-human-brain
>>55549002
Oh, there are plenty. You see @ibm.com, @intel.com, and @google.com as often as @dickbutt.edu in the author list on most big CS papers. Most CS Ph.Ds go into corporate R&D; only a small minority go into academia.
>>55549021
You don't NEED to match the human brain.
For the vast, vast majority, you only need something that can make the correct decision as often as a general person would.
That's a dramatically more tractable problem space for most industries.
>>55549021
We don't need to simulate everything a human does, only one thing. No one wants a robot that both builds airplanes and cooks contemporary French cuisine; that'd be stupid. Not only would it be stupid, but training an AI on the union of airplane construction and french cooking would be monstrously more time consuming than either one individually.
This bait thread shouldn't have this many replies
>>55547815
>>55547849
>>55547721
>>55548531
>>55548567 (forgot the 've)
>>55548584
>>55548622
>>55549021
:^)
>>55549079
>>55549002
Google is the really architypical example, and they're the one that inspired
>It is not a coincidence that around the time GPGPU systems became advanced enough to run very complex graphical models quickly is the same time that all the major CS firms shifted focus from attracting the world's best talent to aggrigating the world's data.
Google wants to solve all of the world's most interesting problems, and they want to solve them in ways that keep people using Google products. Up through the 2000s, that meant attracting the smartest people, lavishing them in luxury so they never leave, and trusting them to create great products. That was fine. Then, it wasn't. The world changed; Computer Science changed, and so Google changed along with it. Google began to farm data heavily. (So did everyone else.) They did this because they had solved most of the problems that they were interested in and capable of solving. The new horizons aren't reached by hiring the smartest people, they're reached by training the strongest AI.
>>55549090
This brings me back to the point about AI being able to do complex repairs e.g. a robot deduces 40 lines within its data bus is faulty so it goes back to the dedicated repair/maintenance robot to get fixed where the repair bot executes all kinds of complex processes to request the right materials to be used to patch the necessary repairs.
I sincerely doubt that such a thing will exist in the near future (20 years). 50 years is a long time away and maybe these things would happen in 2060. I'm also sceptical that it is possible to program an AI whose purpose is to program another AI or even itself to do useful things.
>>55549163
>They did this because they had solved most of the problems that they were interested in and capable of solving. The new horizons aren't reached by hiring the smartest people, they're reached by training the strongest AI.
Other than self-driving cars, image identification, and Youtube recommendations what does Google do?
>>55549175
See for yourself. Here's the current openings in Engineering & Technology:
https://www.google.com/about/careers/jobs#t=sq&q=j&li=20&l=false&jc=DATA_CENTER_OPERATIONS&jc=DEVELOPER_RELATIONS&jc=HARDWARE_ENGINEERING&jc=INFORMATION_TECHNOLOGY&jc=NETWORK_ENGINEERING&jc=TECHNICAL_WRITING&jc=TECHNICAL_SOLUTIONS&jc=TECHNICAL_INFRASTRUCTURE_ENGINEERING&jc=SOFTWARE_ENGINEERING&jc=MANUFACTURING_SUPPLY_CHAIN&jc=PROGRAM_MANAGEMENT&jcoid=7c8c6665-81cf-4e11-8fc9-ec1d6a69120c&jcoid=e43afd0d-d215-45db-a154-5386c9036525&
>>55549221
Those Google jobs don't give me information about the information problems they working through using AI technology.
>>55549164
Have you ever worked electronics repair? I have. It would probably shock you how simple the job really is. You could program a repair bot sufficient for ~80% of all jobs without using any kind of modern ML models; just using old 1980s expert system design. The hard part isn't, and has never been, figuring out how to fix a given problem. The hard part would be building the hardware to manipulate the physical machine.
>>55549252
*shrug*
Poke around google.com/careers and look for research positions. You'll probably find them looking at an individual google office, rather than google as a whole. When I was graduating college, someone from google engineering invited me to apply, and while looking around I found that google divides research and engineering very distinctly. I've heard that Microsoft, IBM, Intel, and most other big companies do too.
>>55547932
Algorithms are getting more and more complicated as new optimisations are found.
Things like AI/voice recognition/facial scanning etc are all way harder than making a calculator in assembly code.
You're legit retarded if you think programming is "easier" just because we have high level programming languages now.
>>55549262
> The hard part isn't, and has never been, figuring out how to fix a given problem.
If you're referring to decision support systems, those systems aren't trusted to give definitive answers, only to narrow the database of possibilities into a smaller list of possibilities.
>start tech uni bacuz it's the new wave of moners
>finish uni by the time the market is saturated as le fug
:^)
>>55549319
There are several million people doing software development in the US right now. Do you honestly think they're all advancing cutting edge algorithms? Don't be stupid. Software jobs are about maintaining giant, multi-year long projects, building web apps, and writing firmware for mundane devices. All of that is getting easier over time. For every researcher pushing the limit of what mankind is capable of, there are ten thousand engineers and code monkeys pushing sticky notes across a white board.
>>55549344
Let's say that when you graduate, the employment field in CS is indeed saturated and you can't get the job you need at the price you need it. The answer is to expand yourself. If you study law, you can apply CS knowledge to datamining case law or analyzing legislation with mathematical models. If you study medical science, you can join the ever complex medical science field. If you study education, you can train the next generation of Hitlers to progress the white race.
>>55545459
I'll retire in my castle at 40.
>he fell for the bubble meme
>>55549262
The difference is that ML algorithms offer the option to build generic, domain-agnostic solutions with adaptable knowledge bases that can be built online and generalize on unforeseen data in useful and often surprisingly correct ways; while the expert systems requires the involvement of many field experts along with large teams of software engineers (it's almost entirely a menial task, not a technical task), is specialized for the domain, doesn't answer out-of-domain at all, and can only be expanded offline. When you include natural language interfaces into the mix, expert systems are outright no longer buildable because the sheer complexity of the model you'd have to build by hand for this would exceed the entire workforce or exceed your company's lifetime (depending on how you allocate your resources).
>>55548121
>tfw got a br job offshore in hopes of getting into big oil
>be back in school for CS
>hoping the bubble doesn't burst
:^)
>>55545459
The thing with CS tho is that it's applicable to a wide area of fields if you have some sort of additional knowledge like anon said >>55549385
>>55549367
Are you employed in tech? If so what kind of work do you do
>>55549452
Okay, none of that addresses what I said, which is that the impediment to automated maintenance is hardware, not software.
When I was repairing electronics, for a while I paid special attention to the actions I made, trying to put together everything a robot would have to do. There's a lot! I couldn't imagine building a specialized machine to do it all, and I've seen general purpose robots and am unimpressed: they don't seem to have the precision of movement required to fix the little fiddly bits I dealt with most days.
>>55549520>Are you employed in tech? If so what kind of work do you doI'm an engineer pushing sticky notes across a white board. I work on automated factory assembly lines.
>>55549175
Virtually everything, most of it leading nowhere. They have quite a lot of language modeling and image modeling experts. They also have RL experts. DRAW and their go AI are two recent examples in these areas. They also did GoogLeNet but that's just "throwing moar hardware at the problem" along with scalability issue fixing for training schedules on the scale they used (namely with distributed weight updates).
The thing with google is that there is no clear "research" and "software engineering" division. There are "research-focused teams" and "software engineering-focused teams", but they intermingle, and much more importantly, you don't get to choose where you work or on what, and you don't even get to state a preference. They assign a project and team based on your resume and interview, so you can as well be stuck working on a shitty website's frontend instead of doing research for the part of the backend that does all the work and is actually a hard and interesting ML problem.
That's why ultimately I turned their offer down. Lots of startups going around offering the same or even better salaries than google, which have arguably even more interesting and legit people working for them, but actually separate research and software engineering, and feature much clearer research interests, so you don't get cucked.
That said, the other faggot is full of shit. He has no clue how ML is done and what ML needs to work. He actually unironically think you can just throw more data at a problem and gain something significant from it. Protip: it's exactly equivalent to just throwing more hardware at a O(n^n) algorithm instead of switching to an O(logn) variant. It's no coincidence that the only successful effort in AI from google comes from deepmind.
>>55549580
what exactly are you working on and what education do you have
>>55545459
Give zero fucks because I dropped out 10 years, and again 8 years ago.
Been doing this shit for near 20 years. If the bubble actually bursts and I get let go, I get:
- Severance in cash (because fuck stock as a safety net)
- Option to cash out stock without tax penalty
- Extended unemployment (because economic down turn)
- Option of trade school paid for
- No need to drive to and from work
- Give zero fucks
>>55547932
never on /g/ have I seen so much bullshit in one post
>>55547932
t. webdev losing job to baljeet
>>55548547
Too bad they're more knowledgeable than you at actual science instead of being a code monkey and you will always have to tell people you don't have a degree
>>55548323
>the software industry is still in the infant stages.
It's going through an awkward teenage phase where it comverts to Hinduism and then overdoses on curry. That's how the bubble bursts.
>>55549779
what happens after that
>>55549811
It gets interested in women, and then robots.
A-alex?
>>55547130
>Implying you will be doing any of that shit when india becomes a superpower and warehouses full of $2 a day pajeets steal your job.
FTFY
>>55545459
Tech bubble burst multiple times already. Whatever is left is the dregs and has been for over ten years. The last big shift was when virtually all CS/CSR/Call Centers got offshored before the depression hit in 2008.
>because your only girlfriend in the world, the only person who could ever get past your innate ability to run off anybody within a 5 mil. radius because you said the wrong thing at the wrong time, died while trying practice for some stupid-ass swimming competition she never wanted to do anyway either
What?
I left programming, because quite frankly it sucks now. It was better in the old days, when you could be a neckbeard in your own little hole that nobody bothered much. Now it's all about agile crums and other bullshit.
>>55548521
Not really, you just have to suck up pride and go to a *gasp* vocational college. Yeah I admit they rape you in comparison to the time you actually spend but the world will always need HVAC technicians.
>>55545459
>your only girlfriend
I've never had one and never will. Don't see the point in having one.
>Going to school for anything that doesn't have Engineering in the title
>Expecting to make money
What other dumb shit will /g/ think of next?
>>55547932
>muh FUD
Test post
[spoiler]Bean[/spoiler]
[spoiler]Bean [/spoiler]
[spoiler] Bean[/spoiler]
>>55545498
>>55547932
You morons. The price of tech is already beyond most consumers, which is what causes bubbles: INFLATION. The bubble is well into effect.
>>55545459
I work EE and IT.
I will always have well payed work near my grubby hands.
>>55549101
>forgot the 've
I was pointing out that niggers don't use proper grammar in their speech, so something like' you seen ' is common
>>55545459
you didn't finish your sentence
>>55554404
hello /v/eddit
>>55545459
Fuck you.
What do I do with my cs degree if I don't want to code a lot?
>>55557066
Research in programming-light fields. ML qualifies unless you go with hardware-related research (FPGA ML, CUDA-level optimizations, etc.). Operational research qualifies, though I guess that's not really CS. It should be close enough, though. tCS fields such as complexity theory, game theory and cryptography all qualify as well. You'll still have to program things, but it'll be about 500 lines per project, and you'll mostly reuse the same 500 lines with minor modifications for a few months before starting from scratch. Most of the workload is in actually thinking up cool ideas and planning how to validate them.
I dropped out of that shit for economics (good school)
>>55545459
Unfortunately went through the same thing. In the early 90s I saw that there were jobs offering RNs 20k+ signing bonuses everywhere so got into it and have worked as an RN since. I see new grads all over the place now that either can't find jobs or not in the areas of medicine they want to. And absolutely no signing bonuses like there used to be. This is coming for the tech world.
I have a boring enterprise programmer job, any tech bubble burst won't affect me unless it affects huge and old enterprise applications for banks and insurance companies. Feels good man.
>>55545459
>should of
![maxresdefault[1].jpg](https://i.imgur.com/JHWXb1t.jpg "maxresdefault[1].jpg")
>>55553943
t. Prodigy Programmer
