Prediction: Coding will be no more by 2040-2050
Continuing the trend of blockchained predictions, I'll make one that sounds "absurd" - but I will make it nonetheless.
Context
As of 2016 coding / programming / "developing", is a very active field where the market demand exceeds the "supply" of human coders/programmers. Yet, according to my prediction, human coders will be redundant in just a few decades - if not sooner.
The Prediction
AI will take over the role of coding in 5 steps:
1) Primitive AI will assist coders to make programs even if the human coders don't know programming too well.
2) Primitive AI will evolve to accept all kind of coding input, of any language, in order to be able to compile to very optimized machine code for a given hardware - code that will be extremely efficient compared to 2016 compilers.
3) AI will start to interface with the coders on a more human level to create the desired programs. Programming will take the form of entering a sequence of human directives of what the program should do. These directives (in human language) will then be composed into a functional program by the AI and compiled with the most efficient way.
4) AI will simply request the specifications of what a program should do and create it, after the human inserts the specifications.
5) AI will not need to "make programs". AI, as an extremely intelligent "consciousness" with immense processing power, will be able to deliver seamlessly and on demand, whatever the human asks - without the necessity of creating intermediate programs. At that point coding will be effectively "dead".
8^D love it!
lol...
hahahahaha :D
always an appropriate post when it comes to innovation
fully autonomous software will change our world.
in some ways many of us can barely imagine, and in many ways none of us can begin to imagine.
I'm kinda happy I'll be outta here by 2050 most likely {:^)
Born in 1964, a product of the sixties - hallelujah !
Have fun kids! Love ya, DM
It could happen as early as 2035 though :D
doh!!
you are correct!
Not even step 1 by then, unless the AI people actually change their tactics. They block too large.
I don't think that human will ever give control such AI. As it may learn from human and cause destruction like in most futuristic movies. But however I think visual scripting can be the future @alexgr. By the way nice post.
They won't consider it control... rather that the AI automates some of their tasks. It will be "assistance" with ever-increasing "quality"...
Step 6) Skynet decides it will be better off without those meddling humans, and the robopocalypse begins. At that point, we all will be effectively "dead" (At least I'll be a solid 55-65 years old at that point, haha) Seriously though, it's fascinating to watch technology progressing at the rate it is. I'm curious at what point (year) do you think AI will be effective enough to start assisting coders?
Lol, for (6)... actually humans will prefer to "merge" with the machines, so the machines will have it easy - except for the few that don't want to "merge".
As for the AI assistance, I suspect the first primitive assistance could start around 2020 give or take a couple years - depending corporate priorities.
@alexgr
Why not humans becoming cyborgs so that the A.I programming would be indistinguishable from our own mind.
Initially AI will be externally interfaced but ultimately the interfacing with the human mind will be increasingly integrated or internalized - so yes, in a way, this will happen. But the human will / intent of (I want to do this) will continue to be the activating force of what needs to be done.
I agree on points 1 to 4. However,5 is highly controversial. We don´t know what consciousness is. It follows from physicalist functionalism,
that consciousness could be created in a computer program,but there is no evidence for this claim.
It has been popularized by popular science journalists who are unaware of the philosophical implications.
Personally I belive that A.I will reach a dead end when it comes to reaching sentience. But we will see. If you are right,it should happen by 2040 I believe. After all,we don´t need to mimick human level brain capacity for sentience. Rats,for instance are clearly sentient beings. An A.I as smart as a rat should then reach sentience,if it´s possible. The reason I don´t believe it´s possible refer to the explanatory gap.
A turing machine, even with effectively infinite capacity, will never have volition like a cat.
If you notice, I used "" in the word. It's because it's not exactly a consciousness. But it will resemble one from the human perspective.
I guess programmers will survive with steemit help :)
The last point will not happen on Turing machines even with effectively infinite capacity.
And, honestly, I doubt an AI would bother waiting about on the whims of anyone else.
The best part about difficult predictions is that they are rejected on the spot as impossible for X or Y reason. This is what gives them value (the more difficult it seems, the more value the prediction has).
I might be convinced if you could describe a methodology that would lead to it; what you mean by "consciousness" or even what you mean by "Artificial Intelligence".
I have working definitions of these things which probably owe too much to Godel, Escher, Bach.
"Consciousness", in the above context, is exhibiting sentient-like behavior. Obviously, it's just code running in a machine.
As for AI, I'm covered by the wikipedia entry of 2016 on the matte (wikipedia snapchots of 2016 would suffice in the future).
I don't really want to convince anyone. There is no reason for that. It's just a prediction. Time will pass and it will either be proven or falsified.
I used to make a lot of predictions in forums that have now vanished and along with them, my (what I thought would be provable and time-stamped) predictions. So, hopefully, this blockchain will still be around - or its data won't be lost.