Whatever the status quo is, changing it gives you the opportunity to be remarkable.

— Seth Godin

The concepts of cost and value should share a relationship of symmetry. It seemsrolex that more often than not, such symmetry is missing. Quite often the value doesn’t seem to be aligned with the cost (that watch cost how much!?). The key is to look deeper, the costs and the value are aligned, but the truth is often hidden by shame. The shame is often the explicit admission that we value things that are so inherently superficial and ultimately damaging to our future.

Anything that just costs money is cheap.

― John Steinbeck

imagesMuch of the actual value we take in items is hidden. It also matters greatly who does the evaluation. This is clearest with luxury items where people are willing pay a great deal for a brand name because of its cachet. This is a how top designer and brands make their money; people are willing to pay significant premium for having the product with a name. Examples can be found with shoes, cars, watches, and handbags among many items. It becomes a sign of distinction just to own certain iStock_000000286836XSmallitems.

The same dynamic is playing out with supercomputers. In this field there is a great value placed in being #1. For this reason we are quite willing to build virtually unusable computers created solely for the purpose of being the fastest using the standard. The standard is the Linpack benchmark, basically the LU decomposition of a dense system of linear equations. As a measure of scientific computing performance this benchmark should have lost all relevance decades ago, yet it lives on shaping supercomputers and driving decisions.

It all revolves around having a supercomputer that is defined as fast being a luxury item of great value. The damage done by this state of affairs is tremendous although it remains unremunerated because most of the damage is tied up in the alternatives not explored because of the pursuit of this luxury item. The irony is that computer speed was valued because of a computers utility as a problem-solving tool. Ironically we have taken these marvelous tools and rendered them less useful to make them faster. We have taken a tool and created a fetish. It isn’t so amazing or awful because people do this so often. A gun or car or any of the luxury items mentioned about is a similar fetish. The awful part is the lack of recognition that this is what we’ve done!titan-supercomputer

No human endeavour can ever be wholly good… it must always have a cost.

― William Golding

The other side of the dynamic is also in evidence when the value of something is gained without paying the true cost. A chief example is the acquisition and use of natural resources. The long-term costs of mining are often not carried by those who gained profit through the acquisition of a mineral, but pay nothing for the environmental carnage unleashed. These costs are seemingly hidden because they only manifest themselves downstream although a rational examination of the situation would have yielded a determination of the costs up front.

But nothing’s really free, is it? People always make you pay one way or another.

― Jane Lotter

Opportunity cost is similar. The cost of doing one thing over another is not readily evident. The reality is that we don’t know the outcome of roads not taken, but we can reasonably assess it. In the scientific world, the balance between basic and applied research can define this. Applied research can be viewed as short-term profit taking whereas basic research can be viewed as the long-term, high-risk investments.

6-856jf02In the area of computing other opportunity costs are in evidence. We have made significant effort to make high performance computing about big computers solving meaningless benchmarks. We have defined weak scaling as a success. These have come at significant costs such as the diminishment of the efficiency of doing most of our computational work because we have failed to focus on performance at the basic node level. Our pursuit of mistaken goals in high performance computing has also driven a divestment from many important areas of algorithmic research. Perhaps the most powerful tool for effectively using computers, the algorithm, has been stripped of its vitality in the process.IBM_704_mainframe
This has been merged with the wholesale change in the computing market, which as transitioned from mainframes to personal computers to hand-held. In doing so the entire focus of commercial computing has changed as well as grown to a scale unimaginable a generation ago. With this change the opportunity costs are snowballing into something that should start getting attention.ibm-pc-250
Consumptive spending compared to investment is a keen societal example where near term value is taken by using resources now. Investments such as infrastructure or research yield less immediate benefits, but usually provide greater returns over the long run. We have made
serious marketing choices by masking our short-term business profiteering and general lack of investment by merely relabeling such uses of resources as “investment” when it is nothing of the sort. The true cost is implicit and defines much of the economic trouble we find ourselves in today.PhoneComputer

The riskiest thing we can do is just maintain the status quo.

― Bob Iger

Advertisements