Say that current-day RAM modules would cost $100. These are the 100% correct modules, all others are wasted. By selling a decent subset of the wasted ones for a profit of, say, $20 each, a new audience is addressed, namely the one that thinks $100 is a lot (mostly home users). The professional market has the tendency to go for high-quality materials, and they will continue to be willing to pay $100 per module.
Some of the professional users will however turn to the cheaper alternative; in the given example, this has a break-even point where the professional market's switch to the cheap variety makes up a fourth of total sales; not a figure I would expect!
A possibly more important reason why this idea is of commercial interest is one of competitive advantage: Being the first to start a sales line for bad RAM gives a head start which has often been shown to be crucial on a dynamic market. This technology will make it possible to start a new company label, marketing differently priced RAM modules for different guaranteed percentages of correct bytes.
Please note that a similar practice is commonplace in the processor market; if a processor fails testing at a high clock frequence, it is tested at a lower frequency and possibly sold if that succeeds. This means the model proposed here is a proven strategy, but applying it to the RAM business is at the same time innovative because it is currently considered impossible to work with partially defective RAMs.
This weird idea seems refreshing and new [that's hard to establish in the Linux world nowadays --- most thoughts have been thought already!].
It combines my interest in hardware and software.
And, let's not forget that: It may yield me a few additional DIMMs in my PC.
Why do it for Linux?
Linux is Open Source software.
This makes it more suitable than, say, Windows.
Furthermore, it's architecture is quite well-done, thanks to many developers cautiously checking it from their particular perspectives.
Linux is also (to the best of my knowledge) the most flexible OS around at the moment. Furthermore, it is used most on home computers, the kind of platforms this technology (initially) aims at.