The political debate over "working inside the system" is certainly important to have. But the organization that makes those hardware tokens used all over the place could vanish, and it would be a minor systems integration inconvenience.
The reputation hit to a fundamental algorithm is going to be confusing programmers for a long time. I don't even know how to start measuring the cost of that.
Of course that probably won't happen since programmers who don't know what they're doing implementing crypto seems to be as popular as ever.
And then, the degree of "knowing what you're doing" is important too, because I'm pretty sure I have a better background in algebra than some professional cryptographers, but human blind spots can get pretty subtle.
The difference between a PRNG and an asymmetric cipher is easy to understand. The cognitive load of associating RSA the company with RSA the algorithm (and ECDRBG the PRNG with ECC the PKI for that matter) is difficult to overcome even when you're aware of the potential bias.