A naïve prediction on how the consumer market for computational resources will look in 5 years.

In five years you won’t be able to buy laptops or desktop retail. Or at least it will be a niche market. How is this possible? With the rise of distributed computing and a shift from “computation at home”, to “computation somewhere else”, it is likely this all be outmoded soon. And this makes sense, why write code that is going to be very painful to compile when you can just throw it at a cluster? There are an increasing number of cluster operating systems and things of this nature already. So how will the shift look?

Well some people have already figured it out, namely Amazon, Google and other companies that offer computation on the web. So companies like this offer cheap resources on the web (for now) and you pay for what you use. Right now this is pretty cheap, but I doubt it will stay this way. Of course as computers continue to improve and servers continue to get “better” this will keep costs down (the faster the computer the cheaper it is to use because the less time it takes to compile a program), but eventually the market will reach an equilibrium. It is my hypothesis is that depending on the path towards equilibrium that we take, that equilibrium could have higher prices. Eventually computational resources could be more expensive.

This will lead to an increase in cost in server resources. But that is of course, if we under use our available resources. The Internet itself is just really one big computer. Some people have already started to recognize how this could be used to keep costs down. A great example is the Seattle project run by Justin Cappos. Its a system that allows you to anonymously distribute computation among a bunch of other computers around the world! (It’s free but you need to contribute resources as well) You can find more information about this here: https://seattle.cs.washington.edu/html/ . Otherwise companies will end up in an arms race to offer cheaper resources, at first, then the weak companies will die and eventually we will get an oligarchy of computational resource barons, charging for every computation.

Examples of what the future of computers will look like are already out there, in case none of this jargon really makes sense. A great example is the Chrome book by Google. It’s a computer that only lets you connect to the Internet. I believe eventually this will simply be the norm. Eventually we won’t have individual operating systems, we will just have user accounts that follow us all for all our lives from one device to the next. This will allow users to not have to transfer files. Instead your life will simply live in the cloud or on some server somewhere, virtually. Some people take this pretty far. For instance a collective of entrepreneurs are trying to create machines that we will download our consciousnesses into, thus allowing us to escape death. Here’s the idea: http://en.wikipedia.org/wiki/Mind_uploading . I think this may be taking things too far. But the work is certainly very interesting. I think that it will lead to a whole host of new technologies, however finding the “fountain of youth” is probably ridiculous.

The point of the Internet is about throwing off the chains of individual computation as much as it is about exploring personal curiosity. If we limit the amount of resources available, we could lose something important. Of course I don’t know what that is, but think about it this way: What if Google charged you to use its search engine? Let’s say you are a poor entrepreneur who wants to start a project that would give homeless youth shelter, but you didn’t have enough money to search for the right resources, so scrap the project. Let’s say that one of the homeless youth could have grown up to invent an operating system for quantum computers. Well then everyone would lose out, right?

But you would never get that if you don’t let the poor entrepreneur use the resources for free! (I know there are a lot of if’s in the above hypothesis, but just bare with me). In general this is known as positive externalities. It generally occurs in networks, where the actions you perform positively benefit others, at no cost to themselves.

Now of course companies like Google think about this right now (Google drive and Gmail are examples of free computational resources you could pay for), but for how long will this be the case?

Leave a Reply