Thursday, May 8, 2014

Further thoughts on climate change and singularity

In my circles, climate change is a serious concern, and it was for me too until recently. Now, while I still believe in climate responsibility, I am beginning to see it as a low priority issue compared to addressing the impacts of AI. I touched on this in a previous entry, but I want to elaborate on how I came to this perspective, and invite those with more knowledge on the subject to confirm or critique my thinking.  I've been looking around for something written by a climate expert that addresses these ideas, but I haven't found anything yet.  If I'm wrong, I want to know.

Consider these predictions: In the next 10-20 years, solar or fusion energy will become so inexpensive that fossil fuels and nuclear will become prohibitively expensive ways to generate energy. As solar or fusion start to rapidly replace fossil fuels, energy consumption will increase, while energy prices will drop, eventually reaching the point where reversing carbon emissions (i.e. extracting carbon dioxide from the atmosphere) is inexpensive enough to become politically feasible.

In my experience, a common reaction of climate change activists to these sorts of predictions has been something along the lines of "that's just speculative nonsense. It's merely wishful thinking to assume technology will magically solve the world's problems when we are facing real and imminent climate catastrophe."

However, it seems to me that the reverse is true.  Any predictions for our climate that do not incorporate long term trends in alternative energies are based on unsupported speculation.   In particular, it is not reasonable to assume that long term exponential trends in solar and fusion that have been going on consistently for nearly half a century will suddenly come to an end right before bearing fruit, for totally unexplained reasons. Without a strong argument for why these longstanding trends will suddenly halt, any model that does not incorporate rapidly diminishing fossil fuel usage in 20 years, and possibly much earlier, should be suspect.

Consider these graphs about the trends that I am referring to. The price of solar has been rapidly and steadily dropping since the 70s:

Screen Shot 2013-10-07 at 1.14.47 PM

Meanwhile the efficiency of fusion has doubling every couple years:


If these trends continue, in the next decade or so, solar and/or fusion will be lower cost than even the least expensive fossil fuels.

An interesting quality of these longstanding trends, is that they will not appear to be making much difference at all in energy prices or carbon emissions until they make solar or fusion price competitive with fossil fuels.  That is only starting to happen now, and only for the most expensive types of fossil fuels.  However, over the next decade, if the trend continues, solar should be nearing parity with the least expensive fossil fuels.  At that point use of fossil fuels will plummet as will associated carbon emissions, and the price of energy. Before that point, however, nearly everyone will opt for the less expensive choice, and it will appear that advances in these fields are making little difference. Because that price parity point is in the future, and is something people have been waiting for for a long time, it is easy to dismiss predictions based on these trends as something that would be nice, but not something we can rely on. I think that would be a reasonable position if we were dealing with short term trends going back only a few years, or if experts in the solar and fusion fields thought they were hitting a wall in how far they could push these technologies. However neither seems to be the case. Thus, the less speculative inference seems to be that we will soon be entering an era of inexpensive and clean energy, as these trends will soon bring solar and fusion to price competitive levels, and shortly thereafter make fossil fuels prohibitively expensive in comparison.

If that turns out to be the case, the long term impacts of climate change--say anything further out that two or three decades--may never come to pass.  If energy is inexpensive enough, then proposals to re-capture and store atmospheric carbon will become inexpensive enough to be politically viable.  So while I believe in climate responsibility (as insurance against an unexpected disruption of trends in solar and fusion, and also to mitigate the effects of climate change over the next few decades), I am optimistic about climate stabilization, and tend to think that climate change concern is a distraction from more pressing issues.

Over the next 30 years technology will bring such radical changes that our economy will hardly be recognizable. There will possibly no longer be human labor in any field. Energy consumption may skyrocket while climate stabilizes. Our political and economic systems will need to be re-imagined and transformed or we will witness the most extreme inequality imaginable. We need to consider all the myriad ways that the future will be different from the present, if we want to make our work meaningful. So, for instance, it makes no sense to predict climate change trends out fifty years, without extrapolating trends in solar and fusion as well. The world is not going to stand still while climate change, and other ecological damage, smoothly follow today's trends.  Cheap clean energy, nanotechnology and AI will alter the equation so drastically over the next 30 years that any plan for the future that doesn't take them into account will be radically outdated.

So that is how I think about climate change these days, and I think it is not an uncommon perspective among singularity thinkers.  Is it wrong?  If so, where can I find a clear, thoughtful analysis explaining the flaws?