I don't think it's about complaining, it's about waking up people to the fact they're putting blind faith into an imperfect program, not an imperfect code, if you can understand the difference. And it's not just about some 125-bit self-developer-imposed limitation. You may very well have some great written code which doesn't work (i.e it seems to work since no one bothered to QA it and no tests exist), and spaghetti code that correctly controls a rocket landing on Mars. It's obvious by now the 130 solver (which might very well be an organization or some huge zombie botnet) did not use JLP's program, and likely not even for 120 or 125. Just because something is not revealed does not make it non-existent. In the same way, just because something happened, it does not mean it happened because of a specific reason. We don't even know if there is yet some unpublished weakness in ECDSA that was exploited. Also why would one gather 1.7 million $ in a single script address and not touch it at all after more than a year and a half, assuming there was a high cost of solving, that needs to be paid...?
I think Kangaroo has already fulfilled its purpose, it is a purely probabilistic algorithm and for future puzzles it is not profitable to use it, it is better to mine blocks, I think it is time for new things that do not depend only on computing power, I am sure there are some things cooking out there: logic + strength.
but I do not think there will be a back door, new techniques yes, after all the keys to the puzzles are insecure it is not surprising that they are deciphered and if the prizes are large, even faster.