It seems that the main issue is that people are only now discovering that PhDs and perhaps sometimes undergrad degrees in humanities are not vocational. This isn't to say they aren't worthwhile or that students shouldn't pursue them. There are plenty of degrees or experiences that are unlikely to be marketable in the future economy. The main thing students must learn is how to learn, and one can accomplish this by studying a variety of fields.

It does, however, mean that subsidizing higher education is a risky proposition. Universities go to great lengths to create humanities requirements perhaps in part motivated by a genuine belief that core requirements are necessary for a proper education, but likely also to beef up departments that have been isolated from market pressures. These departments are often political (almost always liberal), ideological, and rarely subject to any kind of quality control (there was the famous case of Larry Summers questioning what exactly Cornell West was doing as a tenured Harvard professor). Subsidizing learning with no direct market rewards can be a very smart long term investment in human capital. We need history and literature to have the cognitive tools to innovate and to articulate ideas necessary for the Western world to exist. But on the other hand, it's a dangerous proposition to create a mini-industry where you pay people without a clear understanding of what they are doing and why.

One of the reasons the status quo is very hard to change is that academia is built on reputation and prestige, and there is really no other measure of success. That means that if we are at some stable steady state going outside the system and doing something like opening up your data to everyone versus trying to publish in a brand name journal will be a disadvantage to you since the number of publications in these types of journals are they way that you are judged. The issue isn't about whether the journals charge for content or not. Journal subscriptions are cheap compared to labor and reagents and, as always with third-party payer systems, the incentives aren't really aligned to skimp on them. The real question are there better ways of giving people credit for their work in a way that enhances their career in a proportional way to their achievement? Are alternative systems better for rewarding the right people faster, and thus moving research faster? The answer may be yes, but there is a significant energy of activation barrier to making any kind of switch from the publishing-as-a-measure-of-achievement model.

The variance you are exploiting in this arbitrage setup is a result of the inherent variance in information about the chairs, i.e. uncertainty. Since they are used, there is lemon's principle at work. You can argue that Aeron's don't really depreciate in quality that much though over time so that variance in unjustified and in fact there should be some exact market price (perhaps as a function of time used, or more simply just dependent whether its used or new). But the thing is, more likely than not, that variance IS justified. Some of the chairs may be less broken than others. Some of the sellers may be less dependable than others. You can of course condition on this stuff - pick the low price chairs and mark them up by being the most dependable supplier, vouch for the chairs, become an expert in the chairs and be better at out the lemon's, but those things all add cost, so NOT doing all that work is already priced into the market. But cool experiment nonetheless! I have also found that I can buy a used Aeron, use it for a few years, and sell it back for like the same price. Not really a quickly depreciating asset .... maybe better than the stock market the last few years ...