Nature is not a Machine
Nature is not a machine. I’ve noticed a growing hubris in the way in which people are talking about the BP catastrophe, not least Mr. Obama, who although he’s a lawyer, should know better. We are listening to these people talking about ‘putting it right’ in the same way that one would fix a broken watch. Replace broken parts and all is well. This drive for the metaphorical arises from the way our cognitive systems seek pattern, which is mostly an evolutionary good, but it has its limits. The wrong metaphor can lead us to desperately wrong analysis.
Nature is not a machine. There are no spare parts. There is no fix. Rather, it’s a cohesive biological system of unimaginable complexity. What has happened in the Gulf of Mexico has changed the environment – forever. The system state has been radically altered and the expectation that it can be put back the way it was is scientifically naïve. I’m sure there will be remediation efforts, but that will not, ever, put that coast back the way it was. This conceit of nature as our pet machine is clearly of machine age origin, but now that we have so much better metaphors for the way the world works, why do we insist on continuing with the same old nonsense? In the last 30 years, with the advent of genetics, the language of modern biology provides us with far more useful metaphors – and modes of analysis. We have nudged a complex ecosystem out of equilibrium. It will eventually find a new equilibrium, but that may not include the existence of blue fin tuna. You were tired of sushi anyway, right?
This is a map of what the Gulf looks like. Notice that it looks a lot more like a nervous system than a machine. It is. It is the nervous system of the beast we’ve made.
(Image thanks to Matthew Baker at ESRI Educational Services)
If we look at this as a biological system whose health we have compromised we can start thinking in terms of the way in which biological systems recover – they heal. And that healing process leaves scars. Life continues, but not with the same vigour as before. And sometimes recovery is impossible and life ceases. The Southeast coast of America may partially recover, but some of the more fragile populations will not. The complex web of life that makes up that ecosystem has been diminished, and pretending that it’s all just going to get taken to the mechanics and put right is offensive. Pretending that it’s just a matter of the BP executive writing impressive cheques makes it worse.
While we’re at it, why is the Gulf of Mexico a catastrophe and Alberta is a business opportunity? Every single peak oil geologist and ecologist I know has been on about what the downside of the oil production curve will look like since Colin Campbell started the Association for the Study of Peak Oil. No mystery. No magic. They have been saying for at least a decade that it will make parts of earth look like bad science fiction. Irrevocably damaged/destroyed ecosystems – which are not easily isolated from the larger planet wide ecology – will inevitably leak their toxins. What cancerous horrors await the good people of Alberta we can only imagine. Massive bird casualties are routine. But no one seems to think of it as a catastrophe.
Let’s put it in context. The largest dam in the world is China’s three gorges project. The second largest is in Alberta. It’s called the Syncrude Tailings dam. It contains 540,000,000 cubic metres of toxic sludge. That is one tailing pond. It total Alberta has 840,000,000 cubic metres in tailing ponds, covering 170 sq. km. This is an Edward Burtynsky photo of what that looks like…
And you thought Mordor was scary? So while I’m not in any way trying to diminish the scale of what is happening in the Gulf, I think it’s good to remember that it’s business as usual in Alberta. Maybe we’re at a turning point. Maybe the BP spill is the thing that finally makes us think about what we’re up to. Oil is killing us, because we are part of the same biology. We are not a machine.
Links:
Wintermute
“Welcome my son, welcome…to the machine.”
Pink Floyd, 1975
Twice a year since 1993 the TOP500 project has released a list of sites operating the 500 most powerful supercomputers. This June, in the number one spot is the Cray Jaguar at the Oak Ridge National Laboratory, with a benchmark speed of 1.759 Petaflops. A Petaflop is 1,000,000,000,000,000 floating point operations a second. In second place is a Chinese system at the National Supercomputing Centre in Shenzhen, called Nebulae. Nebulae has a benchmark speed of 1.27 Petaflops. The benchmark is used to standardize results, but gives a slower speed than the machine’s theoretical maximum. The theoretical maximum for Jaguar is 2.33 Petaflops, and for Nebulae it’s 2.98. So there are now supercomputers running at 2 Petaflops.
What makes this interesting is that futurist Ray Kurzweil estimates the computing speed necessary to emulate human intelligence is 10 Petaflops, or 1016 operations per second. A check of the Top 500 lists from the last few years confirms that Moore’s Law is alive and well in supercomputing. In June 2008 the 1Petaflop barrier was breached for the first time by an IBM cluster called Roadrunner at the US Department of Energy at Los Alamos. So it’s reasonable to assume that by June 2012 a machine will reach 8 Petaflops – within striking distance of Kurzweil’s estimate. Two years after that a machine will exist that is nearly twice as fast as brainspeed.
These speeds are opening up entirely new kinds of science. Simulation of complex systems like climate, astrophysics, fusion, and genetics. Researchers will be able to run simulations faster and cheaper for a huge variety of scientific and technological problems which have until now been impossible or prohibitively expensive. The most challenging of these problems is how to simulate operation of the brain’s 100 trillion neurons – the most complex object in the universe.
The Transhuman event horizon is nearly upon us. The merger of our biological intelligence with non biological intelligence is now within one decade. We are in the process of radically accelerating our own evolution. We began doing it inside computers in the 80s, molecular biology labs in the 90s and now we’re about to do it to ourselves. We are entering an era in which we will no longer be the smartest one in the room. By 2020 supercomputers wil be running at 64 Petaflops, more than 6 times brainspeed.
This represents the most fundamental paradigm shift in human evolution – a radical discontinuity in human history.
The TOP500 is a list only includes the systems that are volunteered to them. As Dr. Lockley, of the Oxford Supercomputing Centre, pointed out: “The spooks have got some pretty big machines.” So it appears that within five years either the US or Chinese intelligence agencies will have machines that are capable of simulating human intelligence.
Do we really want the NSA to be the first one on the block to have a functioning AI?
Good morning Wintermute, how are you today?
leave a comment