I agree. This was his first post, and the tile reads: Buyer beware...EPIC GPS DOESN'T work. Why wouldn't you first ask if others are having the same issue, or at least check out the other threads on GPS performance, where most are reporting positive results?
Well the OP did report that he had contacted Sprint tech support, which reportedly acknowledged a "known issue." That is the news here.
As bit of background, be aware from the frustrated experience of those who have wrestled with the GPS bug(s) on other Galaxy S platforms for the past couple of months:
1) One property of the symptoms reported is that they do seem to occur on some phone but not others, and even phones that appear to work later are reported to work incorrectly. No one had ever isolated what, if anything, distinguished these cases. That does not mean there is no bug. It does mean the bug is very had to find and fix.
2) Other carriers at first did exchange phones over this issue. I exchanged one Vibrant with T-Mobile, then later returned its replacement. Curiously, a third unit I tried seem to work at least for a few weeks. But after T-Mobile became convinced that there was some generalized problem with Samsung, the carier stopped accepting exchanges over the GPS issue because a replacement unit was just as likely to be bad.
3) Samsung and the other carriers have not actually acknowledged a defect (pure weaselly corporate ass-covering), but they have announced that new firmware in September will "optimize" the GPS. It is well known that both Samsung and the supplier of its GPS chip, Broadcom, have been working on new drivers and firmware. So this report does not come out of the blue.
4) Samsung said the GPS on the Epic had already been "tested and validated," so the presumption was that the fix was already present on the Epic. But no one actually said the fix made it into the release.
5) The other threads on this forum clearly demonstrate that there is at least one new GPS bug unique to the Epic: Multiple users with different test scenarios all report that their test utiilites show the imputed accuracy is always reported as 98 feet, as if that volatile variable's value had been hard-coded into the firmware. So something, at least, is objectively wrong.