Ok. Yet another post which digresses from core style of Pluggd.in. But then that’s the fun part of it! For over decades now we are a part of a society which breathes on technology and even survives by it. Be it over the internet or mobile network or just about anything else. Awesome online bookstores, location-aware stuff, powerful social networks, multi-touch technology and what not. From terminal based access to Lynx browsers to IE/Netscape to emails to IM’s to Mozilla’s & Chrome’s and Facebooks & Twitters we have witnessed evolution of technology from a completely off-line distribution model to now almost a completely online one.
But all this is about past, and let’s not talk about it.
Talk of the town today is real time internet & social media. In fact social media is such a f**ked term over the internet these days that I really don’t feel like using it. But it”s a living present that one could confirm with every Tom, Dick and Harry on Twitter who is busy selling the social media opportunity to the industry. Whatever, let’s not get into that either.
Let’s extrapolate into the future.
Step back and look at last 100 years in development of Computer Science, Electronics and other traditional Technology first. It all started with simple calculators (Wilhelm Schickard), Charles Babbage, Lady Ada Lovelace and then branched out in to specific fields of research which have been sought after & contributed upon by some of the best brains of our society. A branch of science driven by logic.
Through it came basic internet (disruptive in all respects, as we are witnessing it) which has now progressed into real-time internet where Twitters and other services around the world help humans gain insight, knowledge, trends and information in a ‘decisive fashion’ along the real-time-line of our existence.
So what is next?
Paradigm of Machine Intelligence:
Is it not the turn for Artificial intelligence, Genetic algorithms, and Neural networks to take over? Or it is. These fields of research too have been under continuous development by the best human brains across the world and numerous abstractions have already been successfully delivered by passing generations of computer scientists. For roughly fifty odd years or so. Now what if such an ensemble of layered knowledge of simulated intelligence goes out in the open over the internet for a prolong period of time? Does this sound like a tipping point where mass usage of simulated intelligence in technology would take over the regular solutions available?
Entrepreneurship is an indicator:
Rising entrepreneurship in the field of AI is a strong indication of where the subject is headed. Adapting different use-cases of machine intelligence, there are several quaint startups that offer great value already. For example we have Siri – the first intelligence based Mobile app, Gestures from SixthSense Technologies, Minority report act from Oblong Industries and even in Biotechnology there is Genetic 2.0 which talks about “beginning a parallel genetic code with 256 blank four-letter codons that can be assigned to amino acids instead of the available 64 triplet combination that exists in our lifeforms today “. So new, stronger life forms it talks about.
Another awesome example of Genetic Algorithms is that of work by a brilliant musician-cum-emeritus professor from University of California – David Cope – who created EMMY (derived from Experiments in Musical Intelligence) a computer application which composes and plays operatic music. As good as original Mozart or Bach. Having developed an application which qualified the Turing test with indistinguishable musical ability, Prof. Cope confirms that any creative pursuit of man-kind is just a recombination of something “heard” or “lifted” from elsewhere. Call it plagiarism. He even reverse-engineered centuries of music to its roots & forebears to prove his finds. “Nobody’s original,” Cope says. “We are what we eat, and in music, we are what we hear. What we do is look through history and listen to music. Everybody copies from everybody. The skill is in how large a fragment you choose to copy and how elegantly you can put them together.”
Let’s murder the Captcha now, for example.
The test of “humanness” on internet hangs by a thin thread of binary files. There is anti-circumvention clause of the Digital Millennium Copyright Act (DMCA) in the United States to prevent mass scale circumvention of Captcha and an exemplary case of injunction against a violating company too. All major commercial services such as Facebook, Google & Twitter rely on Captcha & Phonetic test for humans.
But technically speaking how unstoppable is it to programmatically parse the binaries that allows the bots to create bot accounts without any human intervention? If one looks at the pace of on-going research in Voice Recognition or Image recognition it seems pretty reasonable to assume that Captchas are soon gonna have to die and something more “undoable” by machines will have to be put in place. Evolution yet again, until the next wave of upgrade!
Well more than the defeat of Captchas or of Voice Binaries is a concern that of applying the Genetics Paradigm to human generated passwords which too are more-often-than-not simply composed. Passwords are easily predictable piece of data for all practical purposes and technically speaking I am not talking about social engineering. Composing passwords is not as much complex or original an activity as is composing music. And when music itself can be genetically attributed to cut-copy-paste syndrome to the aboriginal sounds of animals, birds and other objects how easy it’d be to work out cracks for passwords using genetic algorithms?
Genetics is more than Computer Science
Evolution applies to Nature. So does it apply to industry, businesses, technology, governance, communication and our children too. Let’s hypothesize a Neural Network and open it over the internet to simply tell us a ‘Yes’ or ‘No’ before buying a particular scrip on the Tokyo’s stock exchange! And now if this neural network is trained with data-sets of masses (applying the social media paradigm here) and its corresponding result-sets – frequently called as intelligence of the crowd/market (one can compound training with real-time data inputs also) – then how powerful would such a neural network be? Will it give an answer with impeccable accuracy as to how the scrip’s gonna perform on the stock exchange? Theoretically yes, coz it will have infinite sample size of training over social nature of internet and time, right?
Code named ‘Awareness’
Human brain is claimed to be the only object in the world which is aware of itself. It is powerful, it can think, decide, learn, store, process, collaborate and get tired too. In fact most of evolutionary physical limitations of the human body (as compared to animals) can be attributed to the available alternative of using a powerful brain in our defense. Brain too evolves naturally and slowly over generations & generations of people. But for technology, the realm of evolution is much faster. Our society pulls out best brains competitively and puts them on projects of intelligence. These best brains contribute philosophy over philosophy over philosophy and build on the capabilities of the machines. Now when this process is opened to crowd-sourcing, developer community, open-source paradigm and other contributory concepts what we might eventually end-up with is “accelerated evolution”. Until one day we would have another object which is aware of itself. Or at least more intelligent than a single human brain.
Now the question is: how far is such day?
It should happn in less than a decade.
random pieces but very well written.
Probably human would end creating the next self-aware thing
+1 – Completely Random, but an amazing read.
Makes me proud of @marvindanig
Wow! That’s like a 800 level RC in GMAT
lol @shashi – the whole thing reads awesome with effects of alcohol
…try it.
http://www.youtube.com/watch?v=NLz2oOAFiFk
check
@michael the YT video speaks of reversing evolution, right?
brilliant article.
i love david cope, love it that you mentioned him
and dude! awesome idea there that stock exchange and neural networks one
hmm somebody needs to do something about it
@Anup I agree. It’s a kickass concept that one can practically apply in the future. Awesome post in every respect.
Jyotirmoy Mukherjee
I had to step back a bit after reading the post and think about the gist. I must admit I had to think for some time before totally digesting the article. Here are a few comments –
1. I think the general flow of the post was great (as most posts by you)
2. In terms of citing evidences and links, it was pretty good. I loved all the outgoing links. I would love to quote a few books though which are highly relevant – like “The Selfish Gene” etc.
3. As for my opinion, I think we already have a lot of self aware and self healing systems (which requires a certain level of self awareness). I think things will be better from here onwards.
4. I also see a philosophical dichotomy here – We are making automation an increasing part of our lives (atleast in the west), we are also making a lot of highly intelligent systems – at the same time our (human) population is increasing alarmingly and is raping the planet. Unlike most I dont think those are opposing forces – I think there will be a fundamental reset of the ecosystem (not unlike the financial reset we saw in 2008) which will rationalise the world population (assuming we dont kill ourselves first). At such a point in time, we will be equipped with these intelligent systems to cope to the changes. Maybe I should write a book on this.
All in all – its a very good post. Required a bit of edit – but thats like saying Savita Bhabi has a mole on her foot.
Hey thanks Pratyush. I am waiting for the day when I am able to write in mad-prose like Salman Rushdie
A computer feeling and behaving like a human being is not anywhere near, unfortunately! We need to know the science of the human brain fully first!
Unfortunately, the human brain is a chemical computer that is massively parallel and computes using a million different chemical reactions. The least understood science of all is how the brain works – still in 2010! Check out the literature – there are more things about how neurologists don’t know about how the human brain works fully rather than what they know!!
This does not mean that we have not come a long way. AI research done in the 80′s is what you see today in search indexing and word processing programs correcting your spelling and such. Even in the 80′s we all felt that it was by 1985 or at worst, 1986 when we would have full human intelligence in a computer.
So here we are in 2010 and excited about how close we are.
What we don’t know about human brain and intelligence is a huge amount compared to what we know and can emulate in a computer.
My personal opinion is that as long as we are trying to emulate complex chemical reactions in the brains with 1s and 0s, we have the wrong equipment to do it.
When computers use chemical reactions and do their computing, the story could be quite different!!!
It is definitely a great thing to note that only chemistry can manage the huge processing power of brain. The muscle uses 40% of metabolic energy generated in the body. But there could an alternate solution which consumes same or lesser amount of energy to process with same or higher panache! We have to figure out how fatigue or tiring-out programs that run like cron jobs, help in processing the processing power..:_)
Nice…this may help us rebuild in case the theory of the Mayan’s for the end of the world comes true….Just a joke , however amazing post just to think of such things amazes me and when it really happens (no ideas)… but as a human i would love to see something that does exaclty what my mind thinks and does…Awesome…
The devil, as everyone says, is really in the details. It has been over three decades when the true AI purported to be around the corner. What AI researchers and enthusiasts have realized in these decades is that any fixed approaches solve specific problems rather than the problem of simulating intelligence.
You take the example of creating operatic music using GAs. Yet in a project like this think of the human intervention required: the coding of traits and music in a format that can then acted upon by the program, setting up a set of evaluation criteria that any output will be measured on and finally given any piece of music decide the context it will fit best. Saying this emulates a Mozart or a Bach is equivalent to saying that their talent was primarily a superior search technique in the solution space consisting of a series of musical notes.
GAs, Neural networks, expert systems are all good for certain things but each of these tasks can also be done by a factory of workers in China taught to carry out the basic operations. In the end evolution is, at it’s very basic level, a very powerful search algorithm. It can run on any substrate and with any actors, but cannot be confused with intelligence. A note on Turing tests being the ultimate goal for any program seeking to emulate the mind. John Searle in his landmark paper Minds, Brains and Programs (http://web.archive.org/web/20071210043312/http://members.aol.com/NeoNoetics/MindsBrainsPrograms.html) demonstrated that Turing test can be passed by mere manipulation of symbols rather than true understanding.
I do believe that 2 specific trends will move us towards true (or strong AI). One, our technological ability to create increasingly denser networks. And second, the introduction of true randomness (via quantum computing and ilk) into our computing framework.
Now that’s one most qualified and awesome comment that I have ever read on pluggd.in….[:)] And you bringing in the quantum computer in the subject too, hmmmmm. But for the need of ultimate processing power. That’s mesmerizing. Good one!
Thanks!
Pingback: What does Gov20 mean afterall? Surely not shovelware! – BubbleIdeas
the concept of the singularity has been around for YEARS, dumbass. you aren’t the first fuckface to come up with the idea.