According with the estimated cost given to that machine (few hundred million dollars) and the problem's nature, what they propose is very similar to TWIRL, an h

According with the estimated cost given to that machine (few hundred million dollars) and the problem's nature, what they propose is very similar to TWIRL, an hypothetical machine that could factor 1024-bits integer to break RSA. That was the reason that made a lot of people consider 1024-bit RSA not secure anymore and change their keys to 2048 bits. The same should happen with DH now.