Lest somebody misunderstand, the very essence of an enterprise (any enterprise) is that it is a bundle of labour and capital whose essential structure and identity is independent of and more persistent than the labour it employs.
That's a horrible oversimplification, but let's take it as read for the purposes of this discussion.
It is for this reason that any contemporary HR policy is aimed at (and this is important) divorcing the work from specific individuals.
What this means is that all and any employees must (and this is essential) be plug-replaceable as a matter of policy.
Unfortunately, if you adopt that policy, you have immediately and severely restricted your ability both to hire and retain the most effective staff and to build the most productive teams.
It is obvious that HR would love for employees to be plug-replaceable in such a way, and it is obvious why. However, the reality is that in a creative industry, and particularly in one related to technology, no two employees offer exactly the same potential contributions. If as a matter of policy you won't hire or depend on anyone with unique contributions to offer, then almost by definition you're only going to have staff with typical combinations of widely available skills and no special experience or unique insights to draw upon.
However, given that the creative component of technology companies is often where much of their value comes from, a business that can't or won't hire the most creative people is always at risk of a competitor disrupting their business model with a new product, service, distribution model...
Moreover, technology can be a dramatic effectiveness multiplier. A single smart, creative person using the best technologies can sometimes outperform an entire team of mediocre people with average technologies. More importantly, in scalable fields like software and on-line services, a relatively small team of smart, creative people with complementary skill sets and the best technologies has the potential compete with a much larger organisation on raw effectiveness, before you even consider the overheads that the larger organisation must bear.
Finally, one of the major factors in maintaining productivity and developing existing technological assets and IP is keeping sufficient knowledge and expertise available within the development team. That can be done through good documentation, tools, processes and so on, but in reality this very rarely happens and word-of-mouth advice is a far more efficient and effective way to pass information around. Of course, if you treat everyone as replaceable you not only forego that most efficient mechanism but also incur the very substantial overheads of trying to use other documentation and tools to compensate. In short, high turnover is an efficiency killer in technical teams where shared knowledge is a vital asset.
If you think all of this is nonsense, you might consider that this is a discussion about Cisco, whose main business model is currently facing an existential threat from modern technologies like SDN. I'm guessing you're a business studies or MBA student, so you might like to consider the commercial relationships between Cisco and the likes of Amazon, Facebook and Google as a case study for what can happen. Cisco's recurring model of spinning out and later buying up side companies to do the R&D for innovative technologies would also make an interesting case study.