I watched this video where a lightbulb is suspended in a solution of HF. The result is that the HF dissolves the glass, but it seems like the part of glass on the surface reacted faster than fully submerged parts. Here is an image of the lightbulb after the experiment:
The glass breaks at approximately the 2:50 mark, and it looks like at that point the bottom portion was cut from a clean line. Perhaps the larger portion dissolved afterwards or was damaged during cleanup, I'm not sure.
Is there a reason why the glass closer to the surface would react faster than more submerged parts? Why would the very tip not dissolve first?
My best guess is that the specific line happened to be the thinnest glass. I find it unlikely that the specific line is the thinnest part. Wouldn't it be much more likely for there to be random holes in the submerged portion from the thinnest part of the glass?