Alexandra Lange at Design Observer does not like the New Apple HQ
Wouldn't it have been more radical for Apple to double down on an actual town? To act more like J. Irwin Miller in Columbus than CG chairman Frazar Wilde in Bloomfield. Miller hired Alexander Girard to spiff up Main Street, and masterminded adaptive reuse of the old storefronts to provide his employees and his neighbors what they needed in town. Cupertino leaders fell all over themselves in their desire to keep Apple's taxes in town, but wouldn't it be better to benefit from some of its knowledge and physical assets?
This is a classic trade-off. Interaction between people is of course good for generating ideas, productivity, etc. But which interactions, between which people, do you want to maximize: Interaction within the firm (a la Apple's HQ) or interaction between the firm and the outside world (Lange's proposed solution)? Given finite time budgets, more of one means less of the other. For Apple, with its secretiveness part of the formula of its success, the answer is obvious. For many universities, the issue is the same (town vs. gown), even without the profit motive. The most productive work-related random interactions I have are on-campus, not between me and some random town-folk (sorry Minneapolitans I meet on the street).
This is the same argument as Eric Raymond puts forth in The Cathedral and the Bazaar, which argues in favor of the Bazaar model of software development (which gives us Open Source products like Linux) rather than the Cathedral (Microsoft or Apple, e.g.). Open source and open content are good things (my Open source and open content projects include the Metropolitan Travel Survey Archive, Simulating Transportation for Realistic Education and Training, some wikibooks, and the Journal of Transport and Land Use), and interaction with the community is a good thing, for the community. As an employee of a not-for-profit University, I am enabled to do these community-benefitting works.
But if open dissipates the quality of internal interactions, or costs the profits necessary to justify a high fixed cost investment, the reasons for resistance are quite clear. Apple's business model is such that it cannot make large multi-year investments if someone else can get their ideas and come to market at the same time without the investment. In some areas, those with high fixed costs but low excludability (ideas are easy to steal and hard to protect, hence patents) and low rivalry (my possessing an idea does not prevent you from possessing it) require secrecy for development. If the fixed costs of development were low relative to the variable costs of production, the isolation and secrecy would be less critical, since the cost of production would be proportional to units made, that does not describe software, where all of the investment is up front.
The Linux strategy of openness works quite well at producing low-level operating systems (which underlie Android phones, servers, cable boxes, and airplane entertainment systems (see figure) among the countless other systems out there), but not well at User Interface, for that someone needs to come along and put something proprietary on top, or hardware/software integration. Google did a UI stack on top of Linux with Android and essentially gave it away in exchange for advertising revenue (if they are not selling to you, they are selling you). This is a different business model.