As I understand it, the ears aren't there to hold the cap in case of an off-spec detonation, they're there to ensure the cap is fully threaded in otherwise the gun won't "close" and you can't fire it. As a bonus, if the threads started stretching and the cap can't go forward enough even when threaded in, the ears will also make sure the rifle isn't operational. Their strength should not be relied on as a safety mechanism because they're made out of the same material as rest of the receiver, which is a non-pressure bearing part, and one that also swings on that pivot under the barrel.
So, instead of trying to beef up the ears, I'd go for a U-shape on the receiver that envelops the bottom half of the threaded cap. That would definitely be more material and it would mean you can't just cut the whole receiver out of a single piece of flat metal, but it should increase structural integrity of that area. Another idea is to "invert" the ears. Have a short step that the threaded cap slides into so the barrel only locks into place if the cap is fully threaded, and then the ears slope upwards away from the cap, about twice the distance they go now. The idea here being to try to deflect the cap upwards and away from the shooter's face without shearing the locking ears, because even after shearing the ears off the cap had enough energy to break half of Scott's face. A third idea would be to somehow build the "won't lock if not threaded in" function to the hammer itself. I'm not sure if that's possible with the mechanism they've got going on there, but it's also an option.
I imagine Serbu thought about most of these things before but decided, for whatever reason, to land on the design he's selling now. I don't know just how hard his destructive testing was.
