Skeptical is good, but considering the way the folks at Corning have been dissing sapphire, simply matching GG3 would be quite a feat. So, if it is sapphire, that takes a bit of wind out of Corning's sails.
Now, the HTC One test did not include a knife-point test - it was a knife-edge, and a slicing action, at that. Obviously, we don't know how the HTC One would have fared with knife-point, so no proof, one way or the other.
Knife point vs. knife edge is quite critical, as the force of a knife-point test is focused on one point, while knife-edge is distributed across the area of blade that simultaneously contacts the screen (if the blade doesn't contact squarely, the force is distributed over time as well as space).
In materials of this sort, breakage generally takes place where there is an existing scratch (the traditional method for cutting glass). Depending on thickness (and other properties), glass can be quite flexible (think fiber optics), so flexibility by itself may not prove much, other than that there are no existing flaws. There's no way other components of the phone (LCD display, metal case) could survive being bent to even a fraction of what was shown - knife-point is a far more practical test.
I'm wondering about that bend test, as one of pure sapphire's characteristics is greater rigidity than glass.
However, based on information that's circulating, Apple's screen may be a composite, with an outer surface of sapphire, a sapphire/glass transition zone, and glass backing. As with glass/polycarbonate composite "bulletproof" glass, it's supposed to inherit the best of both materials, with the end result greater than the sum of the parts.
A sand test may not prove much more than the knife test. Quartz (the hardest mineral you'll likely find in quantity in beach sand - Mohs hardness 7) is harder than common glass (5.5) and plain steel (4.5-5), but not as hard as hardened steel (7.5-8), and is likely softer than GG3 (reported variously from 7 to as high as 9) and, of course, sapphire (9). So it's possible the knife could be a more rigorous test than sand - depends on the knife.
A more interesting test would be a sandpaper test, as the force applied can be more effectively controlled, and the hardness of the abrasive material is known - garnet (6.5-7.5) is similar in hardness to quartz, while aluminum oxide and tungsten carbide are equivalent in hardness to sapphire (9) (well, sapphire is aluminum oxide).