The principle is not necessarily completely unsound. I bought a pair of Quadral Chromium 50 speakers a while ago. I replaced the foil tweeter element with copper leaf, which is approximately 3.5 zillion times thinner than the original material. I made a tool to make a jig so I could form a strip of copper leaf into the /\/\/\/\/\/\/\/\ shape which would allow it to flex as it is suspended between the two magnets in the driver. (With me so far? Current through the strip of leaf interacted with the fixed vertical magnets to make it flex.) Long story short: about a month after installation, the speakers opened up and the 'sound stage' exploded. The quality of sound produced was then about 1.8E7 zillion times better than factory original speakers.

I don't know what the technical term is for what happens when metal becomes 'accustomed' to limited flexing, but that is what seemed to occur. (It is also possible that the new foil was lightly rubbing against one or the other magnets, but anyhow.) After a time, the sound improved more than just considerably.

Depending on the drivers involved in the manufacture of specific ear buds and the materials involved, it may be that so many thousand hours of use works the materials such that they require less energy to move, which would allow more accurate reproduction of the signal. With crappy MP3 files, the loss of detail resulting from compression is not noticeable under normal circumstances, but when accurate reproduction is possible the difference in quality may astound. Whether or not "burn in" will improve any given headset will depend on materials and construction, so a crap shoot at best. But I would not discount the notion completely out of hand as per my experience with full-size speakers.

(P.S. Do not buy Quadral products from the Canadian distributor or Mount Pleasant Stereo. They both suck. Otherwise, the kit is really good.)