Peter Cochrane's Blog: When will the net become intelligent?

COMMENT

Written in a garden at Lower Ufford, Suffolk, surrounded by wildlife on a beautiful sunny day, and dispatched to silicon.com via an open wi-fi signal provided by some kind and sensible person

At a modest estimate the internet has the aggregate memory and processing power of about one human brain. This growing capacity seems to be doubling at about six-year intervals. So there is only one question to ask: why is it so dumb?

Before you reach for you keyboard to point out that we can't accurately assess the ability of the human brain, and I may therefore be a factor of 1,000 or more out, just think on exponential growth:

2006 internet ~ one human brain
2012 internet ~ 1,000 human brains
2018 internet ~ 1,000,000 human brains
2024 internet ~ 1,000,000,000 human brains
2034 internet ~ 1,000,000,000,000, human brains

It will happen, one way or another.

Unfortunately intelligence is one of those words for which we can assign no description, definition, quantification or indeed any degree of real understanding. It is something we all recognise and talk about but don't really understand. Here is my take on the 'intelligent' internet ranked in order of 'vital' quality:

1. A sensory system
2. Processing power
3. Memory
4. An ability to output and influence the environment of occupation

So here I am surrounded by countless carbon life forms spanning the simple to the complex, whilst my laptop has to make do with my fingers as the input or sensory system. Of course this may be expanded to include the built-in camera and microphone but without the deserving impact we might expect of such sophistication. And the output? Just a screen and loudspeaker, plus of course a connection to the net. No physical actuators, no 'touchy-feely'.

As for processing and memory power, I have as much as money can buy but along with the software it is bolted down, static, unable to adapt to circumstance and input. It's not surprising then that this lump of hardware, fundamentally more capable than any insect and most rodents, is just a passive dumb lump of silicon, copper, gold, aluminium and plastic.

No wonder the internet is also dumb as it is made from an assemblage of the same technology. But will this situation go on much longer? Perhaps not! Sensors and other devices are being coupled to the net in their millions day-on-day. And adaptive software is being produced and let loose. In essence the internet is becoming a rudimentary ecosystem in silicon.

So when will the internet become intelligent and when will intelligence, and even life itself, spontaneously erupt?

My guess is it is only a question of time. Even if we do nothing more than continue to build at the current rate, and keep bolting on more and more sophisticated sensors, plus adaptive software, then something will happen. The snag is we might not spot it or, most likely, we won't be smart enough to spot it - and we most likely will not recognise the potential value.

Personally I look forward to an extra intelligence to compliment our own. To date we use teams of male, female and every colour, creed and educational discipline to power up our organisations and problem-solving teams. But it is a model that is already exhausted and struggling to deal with many of our more complex problem sets, and we will certainly not cope with what is coming down the pike.

Fundamentally we need a different view and capability beyond our own to deal with the complexity we now face.

Weather, trading, pharma, medical, energy, physics and network systems have already demonstrated our shortcomings, and the realisation that conventional computing and computer models will not do the job is just dawning. Intelligent machines look to be our only hope!