This is actually slower than previous art, but with proven complexity. Namely the algorithm in http://eprint.iacr.org/2011/474.pdf runs in time 2^0.72n and constant memory, but is heuristic only.

Apples, oranges. We have heuristic algorithms for knapsack running in polynomial time. The trick is that because they are heuristic, they do not necessarily solve the problem.

Frontiers of Europe from -500 to 2013.

https://itunes.apple.com/app/id543207391

see Table 3 in http://eprint.iacr.org/2011/440.pdf

This is essentially Montgomery reduction, widely used in cryptographic implementations.

http://en.wikipedia.org/wiki/Montgomery_reduction

Actually Jean-Francois Revel did not have a PhD (he explains why in his autobiography "Le voleur dans la maison vide").

A generalization of the attack (and many more attacks) is given in: http://crypto.stanford.edu/~dabo/papers/RSA-survey.pdf

Great essay.

I work at university (as assistant prof. in CS, not in US), and it feels a lot like working in a big corporation.

The difference is that we are free to do what we want. But this is because nobody cares about what we do, so it might be even worse in the long term.