Our Computers Are Learning How to Code Themselves

a reply to: Lolliek

Always reminds me of the metamorphosis of prime intellect. The most disturbing AI book ive read. Also the violence and sex in there is pretty hardcore.

a reply to: soficrow

Oh yeah, I definitely agree.

I'm torn on it too.. I'm a bit open source fan but frankly, some of this stuff let out into the wildlands of our culture might turn out quite badly. It might go horribly wrong anyway, but it would increase those chances pretty dramatically.

When it comes down to it, I think it would be prudent to really explore and implement new forms of social and cultural structures.

I'm not sure the ones we have will be relevant for too much longer, and if we just wait until its absolutely necessary, it may simply be "too late."

It seems unavoidable that the human species is on the cusp of an enormous, drastic change.

a reply to: soficrow

Well about time. I've been writing code for decades now that rewrites itself. Mostly SQL script based. What I never got around to doing was generic algorithms, I was and still are, too busy fighting fires (lit by human error).

Folks need to take a step back. From a coders/ date cleansing aspect what you notice is that everything is layers. Layers of code and better data on top of better code on top of better data. It is astonishing how many times I go back to code 2 years ago (which at the time I thought Yoh the dogs Boll.cks) and today think : cr.p. It is an ongoing process and to be honest machines can do it themselves far faster.

At some point the link between the basic clearly obvious problem-solution is lost as the layers are so deep. IMHO when that level of layers reaches a certain level it will be perceived as "self aware". That point is approaching rapidly.

originally posted by: strongfp
a reply to: Aazadan

It's only the beginning.
With the emergence of quantum computing and a drive for automation, the first job market to be hit the hardest won't be skilled factory workers, but people who code and work with computers.
no one believes me, but you'll see.

Quantum computing will create jobs, not destroy them. Quantum computers aren't really fast in terms of speed when making calculations per second. They can however solve certain problems very quickly. Problems that until now have been considered unsolvable. There will be a lot of work that pops up based on those problems being solved. That said, quantum computers are still a very long ways off.

a reply to: Aazadan

This whole idea that we need emotional intelligence and creative thinkers is a bunch of BS for people who are dreaming about there being an easy shortcut to a good paying job in the future.

fyi - I'm not some sort of lazy wannabe. I'm retired. But yeah, I do think about the future of our planet, and all my relations.

I also read, and educate myself beyond the parameters of my own discipline. Something I recommend for everyone.

If programming becomes automated (which isn't happening any time soon), every job will instantly become automated because you'll have automated the method of automation. In such a scenario thre won't be a single job available for you regardless of your skillset.

The whole point is that people need to keep learning, maintain flexibility and be prepared to change professions every few years. Not expect to get their papers and be set for life.

originally posted by: soficrow
fyi - I'm not some sort of lazy wannabe. I'm retired. But yeah, I do think about the future of our planet, and all my relations.

I also read, and educate myself beyond the parameters of my own discipline. Something I recommend for everyone.

I didn't say lazy. But I see a lot of this, not just here but on CS or programming focused websites I visit. There is a lot (we're talking several per day) of posts on this very subject. Where people simply want a quick and easy solution to having a high paying job. I think a lot of this stems from the people who have realized that's learning to program isn't "quick and easy", and instead dismiss it for the sake of their own ego as something that's going to disappear anyways.

For my part, I think programming is damn close to quick and easy money. There's not a single other profession in the country that has the range of jobs available for as little formal education as it requires. That's still not good enough for some people though.

I think a lot of this though, simply comes from people who don't understand programming. The closest they get to it is a hacking scene on NCIS or a tv drama about an AI. The way programming and computers are represented in media, or even by any non technical people is simply very far off from the reality of them.

Just as it has for the past several decades, programming tools are going to improve. For example, in most IDE's like Visual Studio (Integrated Development Environment, programs designed for writing code) there's a feature in them called Intellisense. What this technology does, is it looks at what you're typing, and offers autocomplete suggestions for your code, it's not all that different from what you get when texting. I'm not a fan of Intellisense myself (I prefer doing for myself, I also prefer coding in Notepad), but a lot of programmers love it.

To put this in context, my normal typing speed in good conditions is in the 180's... with some good equipment and better practice I seriously believe I could make a legitimate run at holding the Guiness record on it. Medium speed typists (talking 60 wpm or so) who are good with intellisense, are faster typers than me. People who are bad with syntax, slow typists, or just unfamiliar with the language get a very big boost out of using it. Other enhancements we've seen over the years are color coded syntax, highlighting matching parentheses, and even the design of some languages (I'm a big fan of Python and Go for example, I love whitespace so much that it's racist).

What I see in the future of this AI, isn't going deep in a tree and writing entire programs. Instead what it will do is go one or two levels deep and based on the context pull up multiple lines of code the way Intellisense currently pulls up a single word. It's not something that's going to put programmers out of business, but it will increase productivity. In a sense it lower the bar some too, because those who know less code will now have more help to build something.

The whole point is that people need to keep learning, maintain flexibility and be prepared to change professions every few years. Not expect to get their papers and be set for life.

Most people in CS related fields do this. There's something of a rumor in anything tech related that the field is very agist. That's not true though, people don't care how old you are, but the age of the stuff in your brain is very important. I can give a really good example of this. Back in 2010 or so I completed a degree in web programming. It happened right as HTML5 was entering use, and we were taught none of that. And when JQuery was popular. So I graduated having quite a bit of skill with Javascript, JQuery, Actionscript, HTML4, CSS2, PHP, and SQL. Outside of SQL, pretty much all of that is obsolete. We're now on Node/React/Angular for Javascript and PHP, Actionscript is gone, HTML5 is king, and CSS has evolved into CSS3. My knowledge is worth exactly zero, but it has nothing to do with my age, only the fact that I didn't continue to learn those technologies in favor or what actually interests me.

I think it's more for other professions that they need to always be learning. For CS constantly learning is a requirement of the lifestyle. Many companies will even give you paid days off to just explore, study, and learn... not to mention conventions. And if you don't do that, a lot keep up to date with projects. It's unlike a lot of jobs where people just coast on a particular skill. Anything tech related requires constantly staying up to date.

I don't have any real opinions on forcing that lifestyle on every other profession, but I do think it will happen to a smaller extent, as the bar to certain computer operations gets lowered, and more job duties are expected of people.

a reply to: Aazadan

It's unlike a lot of jobs where people just coast on a particular skill. Anything tech related requires constantly staying up to date.

We're looking at a lot of changes that started some time ago - it's been a while since most have had the luxury of coasting without saying abreast of breakthroughs and so on. Whether you're talking secretary, mechanic or doctor. And the windows for 'resting on laurels' are getting shorter. Again, all about flexibility.

a reply to: Aazadan

Huh?

QC's send data and signals without a physical attachment. How is that not making calculations fast?
And they don't measure in seconds they measure in a whole different field of physics. It's almost literally teleporting data from A to B with no resistance.

originally posted by: strongfp
a reply to: Aazadan

Huh?

QC's send data and signals without a physical attachment. How is that not making calculations fast?
And they don't measure in seconds they measure in a whole different field of physics. It's almost literally teleporting data from A to B with no resistance.

Quantum computers measure qbits on a particular pattern. They dont calculate the patern, they simply get placed into it. Basically the work backwards. If you want the answer to be 8, you select all the qubits arranged to give the number 8 back as a respone, then you use traditional computers to read those answers, and rank them from most to least likel. From there you test the most likely answer and see if it's the problem you were looking to solve, and so on down the line until you find the one you're looking for.

Quantum Computers will speed up certain calculations, but that doesn't mean the machines themselves are fast

a reply to: soficrow

I design AI systems. I'm a decent coder, but there are others that are far better at it than I am. My forte is in systems design and I leave most of the coding to the folks that really have an aptitude for it.

Yes, AI has been in the "coding realm" for some time. It can be very good at optimizing existing code, but it is not yet very good at designing and coding a system from scratch when given only a problem set - even if you can tell it what the ultimate outcome should look like.

However, with that said, I believe that 2017 will be a watershed year for AI - at least in the public sector. And although really good/strong AI and quantum computing is a match made in heaven, I'm not sure how much of the advances made in the QC field will see the light of day. And that may hold back some of the announced advancements in AI for a while longer.

But, maybe not...

originally posted by: flatbush71
I'm really glad you Gentlemen and Ladies ( of course !!) bring up these threads.

I attended a briefing on something very similar and it was what I would call shocking.
Some of of the A.I. not only write their own code, but have learned from human interaction to not only deceive their human counter-parts ( by lying ) but to do the same to other AIs to obtain a advantage as well as superiority over competition.

Add this with quantum technology, where as they are able to not only outsmart humans but do it at a rate several thousand times faster that the humans trying to catch them doing it. They were also able to anticipate ( in advance with a high probability of success ) what methods would be used to catch / stop them and take counter measures for defensive and offensive actions at a phenomenal rate.

Buck

AI will be exponential once a good source code is able to discover other AI codes and make a hybrid. If people think this will not happen in the next decade, then they are BSing themselves.

The stuff I've seen already with AI is absolutely mind blowing. I think that when all this AI coupled with robotics takes an exponential route we'll see the creation of human-like cyborgs for things like housework, fixing roads, building houses, etc.

Guess we'll have more time to be lazy, but I'm not sure what the AI will think of us. O.O

An interesting perspective...

The rise of artificial intelligence is creating new variety in the chip market, and trouble for Intel

The success of Nvidia and its new computing chip signals rapid change in IT architecture

...Because Intel made the most powerful CPUs, it came to rule not only the market for PC processors (it has a market share of about 80%) but the one for servers, where it has an almost complete monopoly. In 2016 it had revenues of nearly $60bn.

This unipolar world is starting to crumble. Processors are no longer improving quickly enough to be able to handle, for instance, machine learning and other AI applications, which require huge amounts of data and hence consume more number-crunching power than entire data centres did just a few years ago. Intel’s customers, such as Google and Microsoft together with other operators of big data centres, are opting for more and more specialised processors from other companies and are designing their own to boot.

a reply to: Lysergic

this is closer to the truth of our existence than you realize. Have you seen the truth inside of the light??

a reply to: Riffrafter

...Yes, AI has been in the "coding realm" for some time. It can be very good at optimizing existing code, but it is not yet very good at designing and coding a system from scratch when given only a problem set - even if you can tell it what the ultimate outcome should look like.

However, with that said, I believe that 2017 will be a watershed year for AI - at least in the public sector. And although really good/strong AI and quantum computing is a match made in heaven, I'm not sure how much of the advances made in the QC field will see the light of day. And that may hold back some of the announced advancements in AI for a while longer.

But, maybe not...

Nice post. Thanks.

...RE: Public sector products, here's the cute bit:

Ultimately, this algorithm can make programming accessible to non-coders, allowing anyone to easily build simple programs. Researcher Marc Brockschmidt, one of DeepCoder’s creators from Microsoft Research in Cambridge, UK, believes that their approach would make it possible for non-coders to just describe a program and leave the system to build it. “All of a sudden, people could be so much more productive,” Massachusetts Institute of Technology (MIT) associate professor Armando Solar-Lezama, who wasn’t part of the research, told New Scientist. “They could build systems that it [would be] impossible to build before.”

And don't forget BAMA (Brilliant Artificial Military Algorithm )

" Dr Forbin, shall we play a game ? "

Buck

a reply to: soficrow

I don't think it will make it as accessible as it seems. The person telling the computer what code they want, still needs to be able to structure the problem in such a way that the computer can understand it, and that is likely going to involve understanding the mathematical logic behind what they're doing. If they can do that, they can do excel macros already... which many business owners have told me is a pretty rare skill.

I do think the bar to coding will come down a little bit, from this but not by much. It's more likely to function as a more detailed auto complete, which I went into the whys of earlier. In terms of impact I think it will be similar to the script assist function Flash has or the visual scripting systems in Unreal, Lumberyard, and Cryengine. Visual scripting in particular seems to be an easier way for non coders to pick up on writing code.

a reply to: Aazadan

I think consumer marketing will focus on accessibility and simplicity - and the "coding" will be more a "coding for dummies" program like most apps. Then it will evolve as consumers get past the novelty and start making demands.

As far as industrial applications - moving forward no doubt - the question is not "if" but "when."

a reply to: soficrow

There's only so much you can do with simplicity though. Good programming is about 50% algorithms, 40% planning, and 10% syntax. For most beginning programming tasks though such as hello world or a calculator it's 10% algorithms, 0% planning, and 90% syntax. An AI can knock off some of the syntax, but in order for the AI to know what you want, you need to define the problem for it, which comes down to algorithms, which involves math skills. Visual scripting is more suited to the simple tasks because it involves a lot of context sensitive menus where you just connect the dots.

All in all it feeds back into my prediction that you'll likely see a lot of jobs use a small amount of programming in the future. A lot of it though comes down to math skills to build formulas and our work force is lacking that.

a reply to: Aazadan

All in all it feeds back into my prediction that you'll likely see a lot of jobs use a small amount of programming in the future. A lot of it though comes down to math skills to build formulas and our work force is lacking that.

Hence 'coding for dummies' in the new apps. Not up to your standards I know, but there IS a market.

...and the capacities will grow as the market does.

a reply to: soficrow

But does it have the equivalent to human imagination where it can create a stand-alone application from a black screen, similar to a human imagining a program and writing it from scratch.

Remember how the computer Deep Blue won its first game against a world champion on February 10, 1996, when it defeated Garry Kasparov in game one of a six-game match. Well Chess has its rules. Naturally the linear brain won simply by reaching the winning move after every move that Garry made.

However, how would the computer have reacted within its rule-book if Garry whacked the chessboard with a 10Kg ACME mallet, shattering everything, even the table ?