Choice is reduced by binary drivers and software. They impose limits on choice of hardware. Additionally, they often restrict the user in terms of platforms and pertinent packages that are supported. Once again, the rigid nature of such drivers (or software) means that the choice of what's supported and what's not is up to the hardware (or software) maker. Decisions get tied to considerations such as a budget, business relationships, business objectives, and neglect of legacy. These factors are not customer-centric, so rights can be abused.
On the other hand, with free drivers, whose acceptance is facilitated by projects like Gobuntu, all code is indpendent from the iron fist of its original creator and maintainer. The benefits are many. They include more control over cost, reuse of old PCs, improved digital preservation, and diversity, which can be important in a plethora of situations. For instance, diversity is sometimes vital for security. It is through obscurity and inconsistency, which are separate from but not opposites of transparency.
The issue of choice can be broken down further to discuss cost separately. Many of us have faced the unfortunate phenomenon ( usually a deliberate business strategy) known as "forced upgrades." The argument which underlines the danger of this -- from the user's point of view -- is that vendors are able to control the way that drivers, much like the hardware that they operate, evolve over the years. Hardware can become unsupported at any stage, and assuming bug fixes are needed, the only choice is to purchase new hardware. This is where the high price of upgrades comes into play.
If an entire operating system is maintained and controlled by a group of so-called 'benevolent dictators' who have full access to all the code, then responsible action will be taken to ensure legacy hardware is supported and bug fixes are delivered without the conflicting interests of hardware makers (profits versus obligations towards the customer). Even if the code is not maintained by this group, which could, for example, be core BSD or Linux kernel developers, a company large enough can hire a professional -- if none is already available in house -- in order to mend driver code, which is both openly available and free to modify.
To sum up, betting one's business on a Linux distribution that is truly Free is a case of controlling one's own destination, direction- and expense-wise. The inconveniences encountered initially, while getting accustomed to a simplified and stripped-down version of Linux, are short-lived. That is because when correct hardware configurations and combinations are chosen (e.g. in the next hardware refresh cycle), there is no trouble ahead. Au Contraire -- trouble is only caused when hardware is picked with long-lasting dependency on the company from which it was bought.
Next time someone enthusiastically says "you should try Gobuntu," ponder this: rather than dismiss this as 'religious' madness, as some people do, you ought to understand that a larger proportion of the industry that surrounds us finally takes a step in the right direction. AMD, for example, proved that the impossible can become a reality. They took the Free route with their highly-valued ATI drivers.
Never shall we say never. If you demand open source drivers and dismiss those which shift control towards the vendor (i.e. themselves) rather than yourself, then change will follow. Remind yourself that the customer is always in charge, and that demand drives sales distribution. A good start would be to attempt installing a distribution of Linux which is free in every [gNew]sense of the word. It might prove to be a nuisance at first, but if you do not stand up for a needed change, who will?