King Ludd #14: Supernational Business Automation
Or why your accountant needs to be a programmer
One of the most noticeable changes in the last decade of computing is the almost extinction of the pure system administrator, stereotypically the large and bearded obscure branch of UNIX aficionado, in favour of a hybrid developer and sysadmin that leverages cloud based compute resources to achieve a lot more, for a lot less. While Amazon, Google, Microsoft and others will sell you elastic compute capacity in the cloud, the next step is going to be business administration and government regulation in the cloud. The resistance to this is going to be enormous, but once it happens either humanity will find itself with a lot more free time or our role as slaves to the machinations of then robotically administered governments will simply be made explicit.
What really precipitated this change was Amazon launching the Elastic Compute Cloud (aka EC2) back in 2006. These days Google will sell you the Google Compute Engine or Microsoft their Azure platform, and while the details differ the general idea is the same; that being that physical server maintenance is tedious, expensive, and can effectively be outsourced to those companies operating large data centres (locations with thousands of servers) which can be rented by the customers of those organisations on an as needed basis. The technical details of this are quite subtle, and while programmers will focus on issues such as how to deal with your computer suddenly disappearing, the main point for the subsequent argument is that while to a customer it may appear that they are renting space on a computer to themselves that isn’t the reality of the situation at all. The software running in these data centres can virtualise the computers inside it to appear like there are a lot more smaller self contained computers that can be rented out as subunits if that is what people want. These virtual computers can be created and started up in a matter of minutes, and that creation can be triggered by programs running on other computers, meaning the entire process of detecting a need for a more computing resources, running the required tasks, and releasing them afterwards for use by others can be automated, with you only being charged for what you used.
While many people have taken to using these services for all their hosting out of convenience that isn’t really cost effective if you want your own dedicated system continuously. Special pricing for that situation is available, or you could use a more conventional hosting arrangement. Where EC2 and similar shine is when the demand for computer resources changes dramatically over time. Perhaps your iOS game is getting a big surge in users, and you need to be able to rent more computers until you find a more sustainable situation or to just now and then to see you through busy periods of the day or week. More classic examples from my experience tend to be jobs run on about a daily to weekly basis, or those requiring steps that need to remain somewhat secret, which previously would have been executed on machines provisioned either by the corporate IT department or ad hoc by someone on a tower under their desk. Moving those operations into the cloud dramatically reduces friction in corporate bureaucracies, often at the long term expense of the IT department.
That’s the world of computing though, but the real fun will begin when we can do this to companies as a whole. One of the better reasons to be excited about Bitcoin is it presents the hooks for writing completely automated software agents that sit somewhere in the cloud computing infrastructure doing some task which makes the owner profit with minimal oversight. This is every hacker’s dream, but can you imagine the North American style tax arrangements for someone that has an income from a couple of hundred software agents doing their bidding on systems around the world in return for Bitcoin? Terrifying. My contention here is that is closer to the reality of work in the future than going to the office 9 to 5, so how will we resolve this?
The answer is there is a simply enormous opportunity out there: corporate virtualisation. In the same way that Amazon will host your computing for you and get you to share with all sorts of people you have no idea about, in the future someone will make a killing by having a global network of corporations you can effectively rent space in. Getting limited liability by proxy is not going to happen, but to give a useful example, a small Canadian business might find it easier to get a virtual Luxembourg based corporation for a few months of the year to deal with some suppliers and customers in Europe reducing their tax obligations and administration needs in Canada. The trick will be reducing the cost of starting that virtual corporation to being in the tens of dollars by having it merely being a firewalled part of a single actual corporation shared by all the customers of the larger company and using software to completely automate every aspect of that separation.
The details of achieving that firewalling are likely to be really hairy, and whoever works it out could already be enjoying a lucrative career in international tax finance. If for some reason it’s escaped you what a big deal this is I suggest you check out the surreal proceedings going on in the UK with US drug company Pfizer attempting to buy UK pharma outfit AstraZeneca for £63bn, with a large part of the justification being to relocate Pfizer’s corporate HQ from the US to the UK to benefit from their lower corporation tax as some perverse reverse semi-takeover. (The governmental rumblings over this I would uncharitably compare with France over Alstom, which led to the brilliant description of a French government/Siemens meeting as an “open, trustful and amicable exchange” – words which if you have to say them mean it may superficially resemble it, but in reality is anything but).
If you attach a programmer friendly mechanism to creating virtual companies, along with providing the means to process payroll and automatically deal with the tax authorities in the relevant countries, suddenly all the options open to wealthy individuals or large corporations become available to almost anyone. This is basically inevitable. The reason was the big If there, as the tax authorities in each country are going to find themselves dealing with ever more insane situations which can only realistically be audited by, you guessed it, yet more automated agents operating in real time. To deal with those situations they are going to have to expose hooks similar to Bitcoin to enable the reporting of transactions as they occur, along with processing payments or refunds with no human involvement whatsoever. The first country where the appropriate regulator recognises this and embraces it head on is going to be the tech hub of the next hundred years as it will enable a level of individual entrepreneurship automatically co-operating with the government to keep costs down in a way that has simply never existed before, while simultaneously dramatically reducing the opportunity for rent seeking businesses to emerge ensuring people stay focused on solving tasks of actual value.
The problem here, at least as far as North America is concerned, is the tax preparation industry will balk at the idea that end users should be given these kinds of capabilities en masse, through pure protectionism. This is merely going to defer the inevitable, because while it may be hard to get the appropriate expertise together to build such a system today when computers do finally cross that line of interpreting natural language successfully it’s game over: they will be able to be programmed to create the infrastructure themselves, only they’ll be able to deal with a complexity no human can comprehend, and thus humans will some day literally be unable to understand their tax returns.
While companies crossed this line a while back, I think individuals are increasingly becoming what I consider to be “supernational” entities meaning they stop adhering to the notion that nation states are really much more than local oddities and accidents of history. For anyone but the super rich the idea of maintaining a corporation on another continent either out of convenience, tax avoidance or as a safety net has been a rarity, but given the ability to maintain virtual sub-corporations for any purpose that is going to change. We already have the problem that programmers spinning up servers can be blissfully ignorant of the jurisdiction they choose to do it in, and the US are going so far as to claim that the Irish subsidiary of Microsoft is beholden to US search warrants in an attempt to expand their remit into this new realm. (Typical for a country that makes all its citizens living abroad still file with the IRS every year). This same situation will occur for whole organisations. The sad fact for the US government, and really any other nation state, is they just aren’t that relevant, unless you happen to find yourself in the eastern Ukraine. Quebec’s own national argument is losing primarily because it is based on a now anachronistic conception of what a country is even for.
Assuming some basic level of security and freedom of movement, the country you’re in is basically incidental to where your income is coming from, and while the tax authorities of the US in particular seem to struggle with this idea, I remain optimistic that many of the others are much more enlightened. If you start from such a baseline and enable the people where you are to operate such complex supernational automated corporate structures as individuals the money of the world is going to start flowing very quickly into your area as the opportunities for individual wealth creation will explode. Nigel Birkenshaw runs atomirex. He can be reached at [email protected]