Many of you may know FeFe "Felix von Leitner" Extreme-Coder/CCC-Member with his infamous but german blog "https://blog.fefe.de"
His statement/no citation but sense of words:
"REAL crpytologists will take
1.) a long time,
2.) many attack tests and
3.) mathematical proofs
before they dare to call a crypto safe ENOUGH"
And this statement remained valid till now, just think about the eliptic curve that was shaped to comfort the NSA.
So if you accept fefes prediction you can really deduce that the contrary to the researchers claims will be the case, because of many reasons.
1.) narrow sight - if you're doing research your biggest enemy is you, because you are in danger of being so full of yourself or your idea that you won't see the invariants.
Just remeber how often you have written code you thought must work 100%, and got supprised because you didn't catch an "invariant" that was actually in plane sight.
2.) hostile thinking - and well this is much worse we can suspect one thing especially after the "Rescola" Gambit
The agencies gotten too smart to only taint the sources, because that's to obvious you need a social drive like a group leader of a standardization group, or the official statement of people with an unscathed background (social engineering people into a certain behaviour).
Be paranoid, don't trust people analyse their arguments!