A couple of days ago Google’s parent Alphabet (actually its DeepMind division) announced that their AlphaGo deep-learned software system has mastered the ancient Chinese board game Go. This is a very big deal during these pre-Singularity years and marks a milestone toward intellectual peerage with humans. AlphaGo is not just a souped-up version of Deep Blue that defeated chess champion Kasparov.
The number of possible moves in Go is trans-astronomically bigger than those in chess for which the next good move can be discovered by generating an expanding search tree of possible future move sequences. In Go such search trees are literally not possible to either construct or ‘grow’ in any useful manner. The alternative approach is to mimic the human brain in its ability to assess the current situation through abstracting board patterns, and then derive the best move by evolving more possible future patterns from alternate courses of action. That in essence is what we do when we play Go, and more importantly, attack myriads of other problems in our daily round. (more here and here)
Alphabet intends to productize and promote various versions of AlphaGo into a number of markets which require high levels of performance within large data and complex image environments. For example, most experts foresee that skilled and experienced physicians and medical practitioners who interpret diagnostic imagery (MRI, CAT, X-ray, sonograms, …) will quickly be surpassed by the new machines. Similar results are expected in fields such as finance, genomics, athletics, complex design, decision making and planning, and so on. The portents here are that large cadres of very highly educated and experienced workers will become increasingly redundant as customers demand state-of-art performance at lowest cost, and companies compete to satisfy such demands.
Given the recent turn of our socio-political expectations, it will not be long until some politicians will argue that it is a human right for everyone to obtain such ‘best practices’ goods and services. But before we go sailing off into that future fraught with unknowns, there are other voices rising to be heard. The 28jan16 WSJ reports that an emerging cadre of ‘experts’ is now considering how to put the brakes on AI developments.
The advance, published in the journal Nature, impressed artificial-intelligence experts and revived a debate about whether limits should be placed on computer programs that gain humanlike abilities to learn by themselves.
In any event, DeepMind CEO Demis Hannabis is “thinking very carefully about how to ethically use and responsibly deploy” the technology. When DeepMind was acquired in 2014, its founders required Google to agree not to use the technology for military purposes. Google has created an ethics committee to ensure such benefits “accrue to the many, not the few,” he added.
I urge you to evaluate these sentiments in light of the countless enterprises worldwide which are developing, advancing, and racing to market the latest concepts in AI. Putting on the brakes?! - talk about pissing into the wind.
[30jan16 update] Robots will devastate third world jobs; robots will replace most white collar jobs; robots will do all future manufacturing; robotic vehicles will replace all transport jobs; … . The Financial Times reports that “The Rise of the Robots Threatens the Poor”; and studies from prestigious universities are now coming in telling us that the robotic “revolution could leave up to 35% of all workers in the UK, and 47% of those in the US, at risk of being displaced by technology over the next 20 years, according to Oxford University research”. (also google ‘robots, job loss’)
RR has been a little alarm bell for the last decade relating news and prognostications in these pre-Singularity years about the growth of systemic unemployment due to accelerating technology. Now finally there emerge some louder voices repeating this message, but still not getting it quite right. Meanwhile, the socialists among us are proposing policies which will guarantee a future world of tyranny much as foreseen by Orwell in the 1940s – recall that tyranny is the most stable form of governance against which ‘democracy by the dumb’ doesn’t stand a chance. And technology-augmented tyranny is a rock stable slam dunk.
I believe the probability is close to one (certainty) that today we are irredeemably beyond the tipping point. There remain few approaches – e.g. the Non-Profit Service Corporation schema - to implement the required redistribution of wealth which may preserve the Bastiat Triangle of Rights. However, all of them require an overwhelmingly educated and enlightened citizenry to pull off. Instead, we are overrun with collectivists and central planners whose interests lie in fomenting class warfare between various manufactured cohorts of ignorant yet compliant voters with scant intellectual assets to figure out where they are being led. (Today we remain in a government induced Depression2 with a hokey unemployment numbers and GDP growth rates last seen during Depression1. See also PriceWaterhouse study on millennials.)
A delayed advent of Singularity that could give rise to a transhuman world is looking less like the saving grace that it once did.