wondering aloud ...

… if the fluctuation-dissipation theorem holds for what we are seeing in the economic state. Basically this theorem describes the power spectrum of the Fourier transform of a particular state variable in an system at or near equilibrium subject to external driving forces. That is, it helps you figure out where most of the driving force behind observed changes in that state variable are, while measuring the dissipation or irretrievable loss of energy in the system.

It seems like there should be a pretty simple mapping between energy and value or valuation, and a number of other of the variables. Maybe this is well studied, or better models exist. I don’t know. What is driving me to think about this is looking at the large swings on the DJIA (Yahoo has the coolest graphs!), and thinking of an equilibrium system (the goes-in-ta’s and goes-out-of’s in balance with the internal machination of the system) undergoing an external force which causes dissipation of value. That and I read this article. For the fluctuations, look at the 3 month graph. I have been worrying that by intervening, the US government may prolong our economic recovery from a strong correction to a longer term malaise. I have heard economists describe the 14 or so year recovery in Japan after the issues in the early 1990’s in these terms, that the government was unwilling to make hard and unpopular choices, and thus managed to prolong misery, out of a desire to do good. Here in the US no politician who wants to keep their job wants to be accused of “doing nothing while Rome burned”, ignoring the role that the market itself has in handling corrections. That is, JP Morgan has been snapping up assets left and right, at bargin basement prices. This helps cement a valuation for these assets, and deal correctly with some of the market concerns. Investors have been coming to Michigan and other areas, and buying up houses that have been foreclosed. Yeah, it is terrible to be in that position to lose a house. The market has found a buyer for the distressed assets. This sets a valuation on the assets. So what is the net effect on the market if we intervene to prevent the market forces from setting the valuations, which, not so curiously, is precisely the role of the market? What we are seeing in the day job is that some prices have been rising (products that use lots of imported parts, or petrochemicals), while others are relatively stable. With a weaker dollar, the day jobs' products and offerings are much more attractive to a global market. We export. This is a good thing for the US economy which has sadly, allowed trade imbalances to exist for far too long. If the government intervenes to help a larger competitor that is distressed, how does this help the market? It doesn’t. It removes leveling from the playing field. Unfortunately, the hand of the free market has been also seemingly stayed in the automotive companies. The local big three have had their desire not to be held accountable for their bad decisions codified into low interest long term loans this past week. So the free-market, which would have pretty much guaranteed to have taken at least one of these companies down, has been prevented from doing so. And there is no guarantee that these loans will be repaid in full, or that the money will be used wisely. Given the protectionism we have seen run rampant in the past, I really don’t expect much significant change, other than getting the tax payers to effectively subsidize product development for an industry in decline in the US. A fundamental question that needs to be asked is, was this wise to do? I am of the opinion that the answer is no. Time will tell, as they will have our money. Lets see if they spend it any more wisely than in the past. Again, histories verdict on them adapting well to market forces has not been generally in their favor. The market is Darwinian. You adapt or you die. They have failed to adapt. In HPC, growth continues. Customers need more storage capacity and computing power, and they need it for less money. This is driving lots of people to look at our product offerings, and speak to us. I told you I was busy, I wasn’t kidding. Current market economics favor the more cost effective producers. You might think these are the larger producers who can leverage economies of scale, though the way we see it, it also favors those who can live on smaller margins due to much smaller OPEX cost structures. If your OPEX is out of whack, you are going to be massacred in this financial state. On the pull side of this, customers need more simulation, not less. HPC is continuing its relentless drive, to the desktop. This is where the accelerators come in. But accelerators, to be meaningful vehicles for future supercomputing, need to deliver an order of magnitude additional performance per node. Not per core. So if I have an accelerator that gets me 10x performance relative to a single core, but only 1.2x relative to a node with 8 cores, why on earth would I even consider trying to sell it to a customer who might be better served by waiting the 2 years for the ~2x “free” from Moore’s law changes? This is why, IMO, some of the economics of the accelerator market, specifically those with cards/SDKs/chips that cost more than the price of a node, are fundamentally doomed to failure. The need is there, but sensible strategies and business models need to be in place. Just some thoughts, Sunday morning, between cups 1 and 2 of coffee.