Can't see this going anywhere - it can't easily be integrated into any
existing cellular system, which shift a lot of responsibility into the BSS
(base station) side of the network. For a start you won't achieve anywhere
near the range you can get from MS <-> BTS with MS <-> MS, due to relatively
'simple' antenna used on each with much lower transmit power. Then you have
the issues of latency as each phone relays the signal from hop to hop.
Further complicate that by a 'relay' node having to maintain two
connections - with a TDMA style system this would either mean forcing
single-frequency operation or using more complex hardware (which means that
adding this kind of functionality to existing technology through firmware
upgrades is out of the question). Add the likely unreliability of this
system on top of this. Then add the issue of my battery life draining
quickly because some guy near me wants to have a conversation routed through
my handset.. the list goes on.

This isn't the first time I've heard this idea entertained, but each time I
have it's been shot down, and rightly so.

Aside from a raft of technical issues, this will only work commercially if
there's a dollar in it for the big phone networks.  Otherwise, there will be
no major investment in getting the whole thing to work properly and it will
be, at best, a geek-driven and geek-isolated system.