Well, it depends on your definition of "sounds the same". As said, if you listen to it on cheap equipment or in a noisy environment, chances are that you cannot spot the difference.
To give you some anecdotal evidence: When I was still very new to tracking, I had a very crappy pair of speakers, and in general couldn't spot the difference between 8-bit and 16-bit samples, or let's even say 11 KHz and 22 KHz samples in the worst case. So I downsampled lots of samples and converted them to 8-bit resolution. Today I can easily spot the difference.
If you cannot spot the difference between the two samples I attached in my previous post, you should first get some proper speakers/headphones and stop converting samples, because that would mean that you indeed are unable to hear what you're doing.
In general, the difference between 8-bit and 16-bit is definitely audible, and you'll hate yourself in a few years for converting the samples just to save some bytes. We're not in the floppy disk age anymore after all, but not that it would make a difference anyway: even when modules were still spread and stored on floppies, people already used 16-bit samples.
So whether it makes sense to convert the sample to 8-bit should be decided on a sample-by-sample basis.