Slashdot Mirror
Microsoft Research Developing An AI To Put Coders Out of a Job (mspoweruser.com)
jmcbain writes: Are you a software programmer who voted in a recent Slashdot poll that a robot/AI would never take your job? Unfortunately, you're wrong. Microsoft, in collaboration with the University of Cambridge, is developing such an AI. This software "can turn your descriptions into working code in seconds," reports MSPoweruser. "Called DeepCoder, the software can take requirements by the developer, search through a massive database of code snippets and deliver working code in seconds, a significant advance in the state of the art in program synthesis." New Scientist describes program synthesis as "creating new programs by piecing together lines of code taken from existing software -- just like a programmer might. Given a list of inputs and outputs for each code fragment, DeepCoder learned which pieces of code were needed to achieve the desired result overall." The original research paper can be read here.
they keep changing.
If it's learning from the Windows code base.
Can we just hold progress back another 40 years or so? I'd like to be cold in my grave before the world changes so much I can no longer find my place in it.
Also, the massive social upheaval during the transition period between our current system and whatever replaces it is likely to be extremely unpleasant for the average person.
Who is this "Al" and why does he keep trying to put everyone out of work? It seems like the best thing for the economy^w Human race would be to find Al and order a drone strike on him.
---
And before all you ACs leap to eviscerate me for my inscrutable post, consider your glyphs.
I am Slashdot. Are you Slashdot as well?
This might be good at helping people who can't program get their ideas at least functional so they can't be shown as proof of concepts to actual programmers who can finesse it to make it proper.
It really won't. So instead of having to manually code it, you need the exact same type of person to specify requirements with as much precision and detail as possible. An act they were already doing while coding. They will have to do this repeatedly while working out the bugs in their requirements (aka, code), and probably still needing to manually fix things here and there. So like, thanks for AI Snippets?
Given the long history of crappy code from Microsoft, it don't look so nice... more crappy unoptimized code...
https://xkcd.com/1185/
piecing together lines of code taken from existing software -- just like a programmer might.
Ehm, right. Because all coders do all day is cut & paste code snippets from StackOverflow. And those snippets are placed there by the tooth fairy?
The idea isn't exactly new, and there have been some attempts in the past to build code generators which string together finely-grained mini routines from specifications into a working bit of software. But those have proved to be problematic, especially when it comes to maintenance, change requests, and multi-threading. Keep in mind that today's software engineering isn't straightforward input -> process -> output anymore; even the simplest apps have to deal with asynchronous stuff in UI, database and networking code, and the resulting problems aren't solved simply by piecing together the right code snippets. Not saying that there will never be an "AI" that is able to code, and research like this is pretty interesting, but in this form this isn't job threatening in the least.
If construction was anything like programming, an incorrectly fitted lock would bring down the entire building...
Isn't writing out requirements in a way a computer can understand the essence of any programming language that has ever existed? So how is this any different? To truly get rid of programmers, the machine would need to look at the world, figure out what the problems were, figure out the requirements to solve it on it's own, and solve it. Then, yes, would programmers be able to look at kitten pictures all day.
automated software production? this. times a million.
Unless your a programmer, I don't think you can make specifications complete and detailed enough for this to work. I've heard many customers say "always do this" and come back later and say well except when that happens.
Only humans are so dumb as to willingly build machines that will put them out of work.
The real questions are "Who owns the code used?" and "Who owns the final code product generated?"
I can't wait to try and maintain code generated by pasting together random code snippets. And people thought old COBOL mainframe code was expensive to keep going, well hold on to your hats.
I'm looking forward when Microsoft uses this AI to code Windows OS. :)
It will be also the time when Linux will finally take over desktop
Probably longer. But I remember such articles in 1976.
Getting hard to take it seriously.
Oh no, and India was doing so well growing its economy...
The Last One (software)
As someone else said, it sounds like it mimics the type of programmer who spends all his time gluging together copypasta code snippets from StackOverflow without understanding them. So it will clearly put the hack (in the context of "hack writer") contractors from India with fake degrees out of a job first. The kind of coding I do doesn't have snippets to put together.
#naabhaprzrag, #sverubfr-000, #agi-fcbafberq, negvpyr[pynff*=' negvpyr-ary-'] { qvfcynl: abar !vzcbegnag; }
Don't forget we need to build a 2000-mile wall on the Mexican border, and stop all Muslims from entering the country.
Taking those tough but common sense steps will protect your job from automation.
before I believe it will work.
I worked once on a very large project that tried to do something similar for the Dutch tax service: put the (ever changing) tax regulations in some form of specification language, and compile that to C# code. I was a contractor for some time on that project. After a 160 milion EUR budget overflow and some questions about it in the parliament the project was significantly reduced in its ambitions.
The humans that develop these machines have the idea that if they build machines to do the mundane work for them, the humans will have more time to pursue labors of love. Unfortunately, the humans that think this way don't seem to care that the majority of other humans think that if you're not in a traditional 9-5 office job your value to society is null (yes I do realize that retail, service, and infrastructure building positions fall into this null category; this is deliberate).
Nothing new or unexpected here. This is just another abstraction from machine language. Expect bigger, less efficient & less secure code.... same as we have each time we add another abstraction layer on.
Changing requirements aren't a problem. All you need is to define a language where they can be specified precisely, and hire someone who can translate your real world requirements into that language. Once you've got those, you can still do away with programmers because the new magic code generation tools will do everything else for you.
If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
The people who call themselves "coders" already do little more than paste together half-assed open source projects they find on github and snippets copied from Stack Overflow.
Replacing them with an AI would increase the quality of software by orders of magnitude, and increase the productivity of everyone who can't be replaced.
"Is developing", not "has developed". I understand that Dr Frankenstein was developing a perfect human, made of snippets gathered from graveyards.
When all you have is a hammer, every problem starts to look like a thumb.
All you have to do is specify the requirements, then you can just graphically draw it and put the app together! That's it! That's all you'll ever need to do. And it will work, just like it did in the 90s.
Actually, it will work exactly like it did in the 90s. Which is to say not at all, and to store up a load of 4GL take-out projects for the early 2000s...
Bad code is already perversive, it is much more common than good code. So, you write an AI that takes various snippets of, more likely than not, bad code and combines them. And this AI is still written by coders who, more likely than not, are not the best coders themselves, so the sum will probably be even worse than the parts... If Microsoft has great engineers working on fancy stuff like that, perhaps they should instead throw a couple to the Skype team for example (I don't know what they are doing, but they are slowly degrading what was a useful service).
In any case, good coders already reuse libraries etc and end up doing the job of several coders that are inventing the wheel. And it is much easier for the non-technical person to communicate with a coder to describe the requirements, because an AI system would still need a specification language to describe things - it would just be a higher level programming language in a way.
I remember I was told that when SQL first came out (SEQUEL) there was the idea that business users would be able to communicate with databases so db programmers wouldn't be needed... yeah, right.
Violence is the last refuge of the incompetent. Polar Scope Align for iOS
"Called DeepCoder, the software can take requirements by the developer, search through a massive database of code snippets and deliver working code in seconds..."
This I have got to see. By the way, I notice that the first thing mentioned is the proposed name. "DeepCoder" - well, with a name like that, how could anything go wrong? After finding that name, I expect the rest of the project was all downhill. So to speak. Erm...
1. "...take requirements by the developer..." Expressed in what form? As random remarks over a cup of coffee - in which case the usual proportion of incorrect, incompatible and misconceived requirements can be expected, along with the standard quota of perhaps 90% of the requirements not being mentioned at all (because no one has thought of them). Or perhaps in some rigorously defined logical format, in which case we might simply call them "pseudocode" or "Model Driven Design" or perhaps "formal methods".
2. "...search through a massive database of code snippets and deliver working code in seconds..." Ah, the long awaited "Frankenstein IDE"! Now you too can have a loving companion or friends stitched together from offcuts of raw liver and other offal. If only it weren't so easy to pass so airily over real difficulties to conjure up images of working code delivered in seconds. I wonder if Microsoft has thought of providing some kind of validation utility to make sure that the "working code" actually implements the requirements?
I am sure that there are many other solipsists out there.
People have been announcing such things at least since the late 80s, possibly earlier.
Who would have bet on self-driving cars the next 20 years in 2010? And yet, we seem to be on the brink of it.
Who would bet on self-driving cars that work safely and reliably in the next 20 years right now? It's always easy to wheel out some dandy-looking prototype that works fairly well 99% of the time. But that remaining 1% is what hurts you. Given tens of millions of people hurtling about in "self-driving" cars, how many deaths, injuries and other harm does that tiny-sounding 1% represent? http://spectrum.ieee.org/cars-...
How about facial recognition systems for airports and other public places that don't produce prohibitive numbers of false psoitives? How about speech recognition systems that get above that hard-to-improve-on 99% accuracy? (Sounds great until you work out that with 500-600 words per page, 99% accuracy means 5-6 errors per page - randomly distributed so you have to proof-read everything you have just so breezily dictated).
I am sure that there are many other solipsists out there.
Then it is up to the customer to test it, test it again, and make the code meet the specs through a complex evolutionary process where they file bug reports and the developer fixed the reported symptom, without fixing the underlying cause, and also introducing a new bug. This process eventually stops when the customer gives up in disgust and agrees to modify the specs to meet what the code does.
The only difference seems to be they have replaced the diploma mill grad with some artificial intelligence. In some sense it is a very good experiment to see if Artificial Intelligence can ever match the Natural Stupidity.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
This is the first numerical problem I ever did.
It demonstrates the power of computers:
Enter lots of data on calorie & nutritive content of foods.
Instruct the thing to maximize a function describing nutritive content, with a minimum level of each component, for fixed caloric content.
The results are that one should eat each day:
1/2 chicken
1 egg
1 glass of skim milk
27 heads of lettuce.
-- Rev. Adrian Melott
... that can turn the harebrained buzzword/bullshit-laden confused and convoluted descriptions ("specs") of my marketing crew into a working product.
No freakin' way. The AI would probably have a meltdown.
My job will remain for a loooong time. If the rest of the population were able to formulate what they actually want, I'd be out of a job 10 years ago, with CASE tools taking my job. But that didn't happen. However, I might be the one discussion the new software with the AI. Looking forward to that. But then again, research shows that talking uses something like 80% of our brain while writing uses 25%.
So some sort of coding via type, even if it is to talk to an "AI", will probably always remain.
My 2 cents.
We suffer more in our imagination than in reality. - Seneca
More likely HR will expect 10 years of Deep Coder experience the moment businesses start using it.
AI can take over most, if not all, aspects of those functions. What AI, automation, and sufficiently adapted robotics can do is return humanity to its "natural" state: 98% of the people living a subsistence life, while the remaining 2% live in relative luxury.
Where are they getting snippets from? To me, this just makes GPL'ing my code much more attractive.
Those who do not learn from commit history are doomed to regress it.
So you mean high level abstraction is the hard part and once a problem has been stated at sufficiently low level the lower levels can be standardized and needn't be recoded from scratch everytime ?
I'm shocked!
I still wouldn't bet on automated cars in the next 20 years. I invite anyone to prove me wrong, but they'll need to sit their vehicle out in my driveway in a blizzard over night and then in the morning I need it to work after clearing off the windows. I expect that it should pull out of my driveway and drive me to work safely even though everything is under snow.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
This would be the "Do what I mean" instruction that developers have always dreamed of.
We have all these different development methodologies to try and fail ideas fast and avoid wasting time. Often what a business person says isn't what they want. If someone invents an AI that can read minds, then AI writing software will be practical. Until then, garbage in != usable software.
We already have code that automates itself based on short descriptions. We call them high level languages. The resulting code is interpreted by our more concise instructions.
This will end up like a lot of engineering design work outsourced to India, China, etc. Someoneses back in the home country will have to sort out all the bugs and crap (risky, tedious, thankless, unfulfilling work) while management pats themselves on the back and gets bonuses for having "saved money". And of course, competition for the thankless jobs will be fierce since so many will be under/unemployed. The past-promised cornucopia of jobs due to boomers retiring has been supplanted by automation and offshoring. Bye-bye middle class and welcome to the Poverty-Rich Leisure Class!
Who would bet on self-driving cars that work safely and reliably in the next 20 years right now? It's always easy to wheel out some dandy-looking prototype that works fairly well 99% of the time. But that remaining 1% is what hurts you.
The sad part is that we have had the technology for self-driving vehicles since the 1800s. It is called rail.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
A few years ago there was a tool that did something similar for art: you drew a really rough line-drawing that just gave a sense of scale and position to each object. Then you labeled each object like "cheetah" or "motorcycle" and had one special label for the background, like "desert." Then it ran an image search: it would look for an image tagged with the same label, with roughly the same proportions as the outline. It auto-photoshopped it in, and viola! Instant art! One of the demo pictures was a cheetah chasing a motorcycle, and it was pretty good and kinda funny.
However, I notice that the art industry is largely unaffected.
Also, I don't think it is so easy to just "blur and sharpen" between two areas of code. Maybe something like "content-aware fill" would work here? ;-) Not likely.
Does this mean Steve Balmer will have to come out stage yelling "AIs! AIs! AIs!"?
So you're feeding a computer a list of instructions, and it is doing something based on those instructions. Sounds like coding to me, just a more higher level.
This is going to be an inner-platform, not for a platform, but...wait for it...for a programmer.
Deepcode bots are eligible.
No good deed goes unpunished.
Great so now the computer will write the program the customer doesn't want because he didn't really know what he wanted when he wrote the requirements. How will this AI piece cope with software maintenance & bug fixes?
Build a Man a Fire, and He'll Be Warm for a Day. Set a Man on Fire, and He'll Be Warm for the Rest of His Life.
Clippy has now transcended and is going to take revenge on every one that would not allow him to help and disabled him.
Yeah, ... and where do they get code from?
---
Anyway, can't wait, ... so I am off surfing.
So here's a question: What skillset would be needed to use this approach effectively?
What many people don't understand is that describing things to a computer, no matter how, is what coding is. The programming language is just a facade. Managing highly complex tasks accurately enough for computers to understand is where the real skills lie.
Just look at one example from the paper :
A new shop near you is selling n paintings. You have k < n friends and you would like to buy each of your friends a painting from the shop. Return the minimal amount of money you will need to spend.
And the output (modified to fit slashdot):
k=int; b=[int]; c=SORT b; d=TAKE k c; e=SUM d
What it proves is that the AI is great at answering test questions. However, in a production environment, no one is going to write and maintain a description like this.
And it is just a tiny function. To match the complexity of a real-life program, you have to imagine the same kind of description but spanning hundreds of pages...
Compilers didn't put coders out of a job, these AIs may be the next step but they still won't displace coders. Although it may require some skill adjustments, it won't fundamentally change the job.
Don't be so pessimistic! Once the AI figures out 98% of us are consuming system resources in a pseudo-zombie state, it'll code up the equivalent of a unix "kill" program to "release" those resources back to the system. I saw a documentary with Arnold Schwarzenegger that covered this topic a while back.
Good people do not need laws to tell them to act responsibly, while bad people will find a way around the laws-Plato
I have absolutely no doubts this will be a catastrophic failure, as it's done by the same people who made Bob, Windows ME, Windows Vista and Windows 8 just to name a few. So no concerns over losing my job at all, more like in the mood of grabbing some popcorn and watch it crash and burn.
... there was this idea that we let people define their requirements in a formal sort of English, and let the computer write programs from that. That was called COBOL, in 1959, and the computing world hasn't been the same since.
In Murphy We Turst
search through a massive database of code snippets
Does it also search through a database of software patents to make sure that it doesn't infringe?
A better use of the AI in today's world would be to review all existing software patents, then generate as many non-patented concepts as possible and file for software patents on them.
Maybe if you work for NASA, are building life and safety critical industrial software, or writing firmware for a pacemaker you get requirements that are this specific. In the other 99% of use cases a developer will be writing to, forget about it. The cost of specifying a problem to that level of detail is astronomical and also multiplies the cost of any changes you might have to make.
Also, who is writing these detailed specs for the AI to follow? Software engineers. This thing is a threat to code monkeys churned out of 6-month certificate programs, not experienced developers. And if my job in 10 years is to interpret hand wavy customer specs into Lojban for an AI to interpret and then hand the results to some interns to groom into something maintainable, then I'll adapt. It's not that different than what I'm doing today.
Self coding AI...
Shudders. Now it's snippets. Next will be full code. Then it will awaken...
"The avalanche has already started. It's too late for the pebbles to vote." - Kosh
At that point it will be BallmerBot doing the chanting. It has a 2-gallon sweat reservoir, a highly reflective chrome headpiece, and can throw full-sized executive chairs over 50ft :-P
"When information is power, privacy is freedom" - Jah-Wren Ryel
I will consider it a success when I can speak to it and say "I want a piece of software that, without doing anything illegal or immoral, will make me a billionaire in under a year", and it works. Multi-millionaire in under a year is a partial success.
Let's look at 2 basic facts: 1. Programming is at least partially an art. Along with interpreting human requirements which are often the most difficult part of the any software designer's job. An AI can make basic shell classes, but after that, well, even user friendly language constructs have been found lacking after months or years of human involvement. So an AI doing that job for anything other than basic junior level tasks, is not going to be able to do the job at a professional level. Not for quite some time. If ever. (Probably a few sci-fi books on the dangers of this) 2. This is MICROSOFT. Come on. Them make a good AI? They have trouble making a good operating system. If this came from, say, the University of Rochester I might have more faith, but MS? The people who gave us Visual Basic (the worst programming language for quality code), Windows ME, Windows 8, Windows Mobile edition (with a memory manager so bad 3rd parties made money writing software to check your code to be sure yours didn't make it crash) and most recently Windows 10 (still being rewritten, and will be again with the EU "concerns" on privacy)? . No....just...no.
"Imagination is more important than knowledge" - Einstein
Don't forget we need to build a 2000-mile wall on the Mexican border, and stop all Muslims from entering the country.
And the nukes being smuggled in bales of pot.
http://www.thekindland.com/policy/today-in-bomb-weed-arizona-congressman-says-terrorists-will-ship-2654
I took a Software Engineering class where we covered formal specifications. For even a simple piece of code, the formal specifications were longer than the code to cover all of the boundary conditions. It was faster and simpler just to write code. Using code snippets in not new. We call those libraries. I've used those all my life. My code specifies which libraries to use and how to tie those together. Also, random code written by other people are frequently crap. Who is going to vet the quality of the code? I am not scared by this at all.
More like Penultimate. https://openlibrary.org/works/...
in my 30+ years of experience in "coding" 80 - 90% of my time is spent debugging/maintaining code. any idiot can string lines of code together to complete a narrow well defined task. as soon as you give it to a real user, they try to use it in ways you never thought of and voila, bugs... in addition, i have yet to work on a project where the requirements were even close to correct the first time (or the second, third...). schleprock
Yet another documentary: https://openlibrary.org/works/...
Remember when Microsoft Songsmith put Sound Engineers out of work? https://www.youtube.com/watch?... This is really troubling that the same could happen to us programmers.
Like work is a biological imperative.
"In America, first you get the sugar, then you get the power, then you get the women..." -H. Simpson
"search through a massive database of code snippets and deliver working code in seconds"
Thing is, I'm in the business of writing code most of which you won't find in code databases. Also, in a sense, among those that create such algorithms that are behind those like the above mentioned. However, I'd be delighted if a day would come when I could shortly describe what I want and some synthesis engine could create the code for me, even if only partially. Especially the tedious mundane parts - however, most of those we don't re-write but re-use anyway. So no, I'm not really afraid about my job becoming obsolete.
I am putting myself to the fullest possible use, which is all I can think that any conscious entity can ever hope to do.
Indeed this "new" tech (copy pasting junk from Stackoverflow) can't handle simple tasks like division, or modulus.
One programming task I handled, we already had a function which basically computed some averages, using division and a few other simple operations. To simplify, lets's just say the core function was something like quotient(x,y), which would return X divided by Y. That was useful. The customer very much wanted a slightly different version. They wanted quotient(x). That was the requirement for the software, compute quotient(x). I'd like to see any AI produce code for that. I, a programmer, did eventually get the customer what they needed.
My current assignment I'm working on today is similar. It's basically "write an SQL query which returns the list of software products we have in our database, and for each list all of the operating system versiona they can run on". Sounds simple, right. The relevant data is a table of about a million rows in this form:
SoftwareID - ProgramName - Vendor
1 - Firefox - Mozilla Foundation
2 - bash - FSF
3 - jQuery - jQuery Foundation
I'd love to see some IA that writes a query to get, from the above table, information about which OS versions each software package can run on. That's my task as a programmer, the requirements set by the product manager. I'll take care of the need, get the job done. I may also strangle my product manager, but that's a different topic.
"Given a list of inputs and outputs for each code fragment", "can turn your descriptions into working code in seconds,"
Yeah, if I had either of those, I'd already be done. "descriptions" are pseudo code, and inputs-and-outputs are already pretty damned close to almost every business-logic algorithm.
Now, if you can tell me what the client actually wants, figure out how their business runs and what they actually need, then maybe you can spend a week figuring out the inputs, outputs, and code descriptions. Then whether I spend five minutes typing perl, or five minutes using microsoft's tool to find snippets, welcome to the very last 1% of the job.
Much like 3D printers, this only changes the tools. I could have always built a chair from wood with power tools really easily, given a diagram. The hard part was figuring out how I wanted the chair to look, and doing all of the finishing. The 3D printer doesn't do any of that. This snippet-finder doesn't do any of that.
So, to summarize, if by "coder" you meant "translates-pseudo-code-into-java", then congrats on your blue-collar career in a white-collar industry.
All this does is make natural language a programming language.
Most software house management looks down at developers as an expense and not an asset, and MS ALWAYS caters to this view.
Then I can move onto something else.
I never cared about my genes anyways. Now ideas... I can feed that to the technology and it will carry them forward. I welcome our technological replacements.
99% accuracy means 5-6 errors per page - randomly distributed so you have to proof-read everything you have just so breezily dictated
How many humans can take real time dictations with a lower error rate ?
Forget this one at your peril. This will come back and bite MS on the rear end!
Sometimes, real fast is almost as good as real-time.
I don't quite see your problem. You had lots of ways to implement quotient(x)...
One possible: double quotient(double x) { return x/2.0; }
Patents Drive Free Software as Hurricanes Drive Construction Industry
Looking at how FrontPage generated simple HTML markup language behind the WISIWYG interface, programmers have nothing to fear for their job...
There is a fine line between an AI that can program and one that cant, luckily its easy to distinguish one from the other. All you need to do is let the AI loose at its own codebase and if its like all other programming AI-s before it it will muck things up, crash and burn. If however your AI is the very first successful programming AI you will get AI singularity and humankind will possibly be exterminated the next day, hasn't happened yet but there is really no way to gauge how far we are. Could happen tomorrow for all i know, but i wouldn't bet on it. AI that can do programming is just exponential self improvement away from general AI, you could say that writing an AI capable of programming is probably the easiest way to go about creating an general AI. Its also all theoretical because nobody has managed to figure it out yet and it doesn't look very promising so far.
We Lisp people knew we would eventually exploit higher order functions. Problem is that C++ and Java people convinced the world that macros are bad.
Unreadable code is bad.
I should use this sig to advertise my book ISBN-13 : 978-1501515132.
Then you will know you job is safe for a while.
I should use this sig to advertise my book ISBN-13 : 978-1501515132.
Show me a real application generated by it that actually comes close to actually working...
We had self diving cars in the 1980's, a few of which drove all the way across the US. I think they go back to the world's fair in the 1930's. The problem is being able to handle all scenarios that might arise like a police officer standing in the road giving you specific directions.
"I want a program that shows what our competitors' pricing will be tomorrow. Sort it so both name and price are always in ascending order. Email the output as plain text to the following 27 people, but only Joe and I should have the ability to edit it. Make the keyboard give a mild electric shock to managers in stores more than 10% more expensive than their regional competitors. And it should run exclusively in The Cloud, but not require internet access for any of its functionality."
I wish I was joking.
Poverty has always existed and mankind seems to be doing well at over 7 billion people.
So now we know how Windows 10 code was written. It just took a bunch of previously written chunks of kernel and applied a new UI based on a "give me something we already had" command.
Why UNIX?
What do you think we've been doing last few years? Crowd sourcing answers to every bloody coding questions/solution you can think of. Here's how it works. Go to their website (and not just theirs, the Internet hs many websites about this last I checked), type in the coding thing you want done. In seconds it provides you with many different ways to do it, links to knowledgeable folk who actually have done it, and references to libraries and related posts in case maybe someone else actually wrote the requirements of your question better than you did.
By definition, that's impossible. Natural language is too ambiguous, so cannot specify a computer program of useful complexity. Programming has to know what you really want, handle edge cases, handle errors, etc. By the time you have put all the requirements into a form that is useful, you've created a programming language. That's not to say that what they are doing is not useful, since DSLs are really useful and inferring some of the rest is helpful. If you can translate DSL into an existing (or new) programming language, and then have the computer identify what the edge cases / failure mechanisms are and provide ways to handle them, then you've made programming better. Much of modern computer programming consists of converting the user's requirements into a little bit of custom code, putting it in a framework, and adding whatever code snippets from StackOverflow you need to handle your problem.
The more people I meet, the better I like my dog.
Unless you are coding very simple business code, that is never going to work. You will get insecure code with bad or no error handling that does not even work for many inputs and is inefficient in every possible way in addition. Sure, replacing a very bad coder may work that way, but very bad coders have _negative_ productivity, because cleaning up after them is more expensive than coding things from scratch again.
So no, the kind of coding I do (and I do not do it as major occupation, I only do it when the task is difficult enough that our customers fail to find anybody else than can do it and the task is interesting) will not be replaced by AI anytime soon and very likely not ever.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
uneasinnese in their voices
Clueless people are always going to be uneasy. They're usually some of the first to be fired.
...a massive database of code snippets...
I have this mental picture of a robot trying to assemble a working airplane from a box with Legos, Lincoln logs, Tinker toys, Erector set parts, playdoh and silly putty, It may look like an airplane but I ain't gonna fly on it.
The Russians have won. They have made the world a cesspool of distrust, greed, fear and hate.
Comprehensive spec:
http://www.commitstrip.com/en/...
aaaaaaa
I mean, if it didn't write itself... then what good is it? Some developer has to write the code for the thing that writes the code.... right? And since all of the programmers are now out of jobs... who's going to maintain those code snippets that this thing uses? Or are we saying that every code snippet that ever needed to be written, is done... ? No. This is stupid. It's a fun idea... and they can run with it... but I'm not worried for at least another 2 decades of coding. I hope I'm out by then. Not because of age, but choice.
Can I just write a description that ends with "and make it all fit in 64K, with a response time of 1ms, and zero defects." and this just magically happens? If not, then we're not out of a job. Sure *some* programmers will be out of a job, but I think the class of programmers in this position should already know that they're due to be outsourced anyway.
There's been this holy grail of automatic code generation by entering a design in some different way other than text (UML) and then clicking a button. And when something almost works but is amazingly impractical (Rational Rose) people get excited. Someone was actually bragging about using a tool which had "only" 100% overhead in its auto generated code (like saying that your diet is great since you only gained 50 lbs). And how do you even fix bugs in auto generated code? Do you assume that the design must have been wrong and start moving block around on a diagram, or do you dig into the obfuscated mess that the tool spat out?
I have on a few occasions spent time searching for code snippets online for something I needed to do. Hard stuff of course, the easy stuff I can just do in my sleep. But I would find lots of answers that were completely wrong for what I wanted, it was only for windoze, or a broken stackoverflow answer that no one downvoted, or dozens and dozens of repeated answers all copied from each other even though it was wrong, etc. If you're in a windows monoculture, doing some vaguely generic web app that every company in the world does, then maybe this works, but you can have trained monkeys with certificates do that too. Now ask how can you do that *better* than all the vaguely generic web app that everyone else does and put it onto a different platform then you have to do the leg work yourself instead of relying on code snippets from search engines.
So Microsoft have invented the procedure call, except they implement it by copying code around from a library instead of a procedure call, like someone in the 1960s might have done. Riiiiiight.
I'm sorry this just sounds idiotic. Isn't this basically just creating a giant standard library and making it so you write the "program" in a functional spec, which would probably get messed up if the wording syntax isn't perfect, hence the ENTIRE REASON we have highly precise programming languages that follow the exact lines the programmer wrote?
They're basically just trying to introduce either fuzzy language programming so that the same sentence written 3 ways gives you the same code (that you probably don't want) or essentially write a new version of COBOL so that non developers can write terrible code... Yea, not fearing for my job even a little bit, especially given that this "AI" even with long term development will be highly unlikely to architect a system or do much outside of a handful of languages (much less language interoperability). On top of that they straight up say in TFA that the damn thing can't solve much beyond 5 lines of code!
I had your job once, and still do (to a lesser extent). People expect you to "do stuff" and "make it work" even if it is illogical and inconsistent or requires some inference of action based on phase of moon.
Sometimes you have to say "No" in ways that don't sound like "no". It can be tricky, but if you just do as your told without question, you're a cynic. A paid cynic, maybe even a popular one, but still a cynic. And then you can bail when it sinks into the morass of unmaintainability.
If you post it, they will read.
Taking handling errors as an easy to take a stab at situation, once you have figured out general patterns for what goes into programs, you don't have to specify again how to handle the nitty gritty. You can just specify "show a file selection box" without having to specify what to do if a file gets deleted elsewhere and should cease to appear in the box or something.
Looking at the state of the art at any earlier level of advancement, nobody had anything to fear for their job.
Have you tried taking a wild guess at filling in the gap and seeing what happens with the gap filled in?
Good, I can't wait to spend my days fixing code written by a script instead of by a contractor who barely even speaks English let alone whatever language they are supposed to be programming in. It's sad when they are getting paid more than me.
If you're a Window$ coder and jobless, I'm sure the Linux community would love to have you. Just make sure to wipe your feet before you come in. Tux doesn't like it when malware gets on the carpet. Contrary to popular belief, you can make money developing open source software. And I'm sure because it's Window$ developing it, it'll still be crap, just instead of clouds (computing) you'll have thunderstorms. Can you imagine a virus made by a super computer would do? Or, a computer that thinks and has complete Internet access with a bug? Jesus...And because it's cloud computing, the OS in question wounding matter because of universal standards. Our military used to use Linux a lot and not have everything networked, but now it seems only the Airforce and Navy care enough to not replace everything with Window$ 10. And then, you got Russia trying to go back to the way Soviets used to have their own computer formats. I don't blame them. I've said this before, and I'll say it again. AI's only true purpose is to destroy open source ideology by bypassing the need for decryption and backdoors and use digital fingerprinting instead. It is most certainly smart enough. The concept was originally coined and AI tests developed by Alan Turing (Turing Test). His original job in WW2 was decryption. AI was born from military necessity and will always be used (knowingly or not) to cripple privacy in the name of "national security," regardless of its developer. Open source is useless without freedom, privacy, and security, ergo Micro$oft's mission to destroy free and open source and Linux just got even easier. We will need all the help we can get.
Yep, gotta stop those hordes of Mexican Muslims coming in.
"When you have eliminated the unacceptable, whatever is left, however improbable, must be the truthiness" - Holmes
Great so now the computer will write the program the customer doesn't want because he didn't really know what he wanted when he wrote the requirements. How will this AI piece cope with software maintenance & bug fixes?
What do you think pushed Skynet to start the war and obliterate all mankind?
Yes, it's called a compiler.
We went through this when we coded in assembly language, too. It still didn't replace programmers or developers. 8-P
"To err is human, but to -really- screw things up requires a computer!" 8-)
I was told the same thing over thirty years ago. The instructor from the mini computer company (Qantel) we were getting training for informed us that the company was coming out with a "Program Generator" that would makes us all obsolete in a couple of years. Needless to say, it wasn't exactly a big success. So I'm not quitting my day job anytime soon.
Good intention, bad technique. Should I give the idea? They are still in copy paste operations, I complain most free sources are incomplete or not working, gone or lack round corners. BUT we have to try programming AIs, anyway! AIs that program, for non native ENglish stack interpreters. Just do not expect substituting real programmers, though example writers may go. Why did it take hour and a half to download this single page?
Ok, I think a similar technology needs a software programmer or a psychologis to describe properly the customer requirements... ðY
AI can produce something that looks like watercolours or sounds like music, but it can't come up with the concepts itself. Yes, you can get a human to produce a list of requirements (probably in some sort of mark-up language) but THAT IS PROGRAMMING - you've just invented a higher level language.
...but who will be around to maintain it?