You could easily sort the conversations by the parties involved. This should solve any conversational confusion. As for programming knowledge, I would encourage JavaScript. It's what I'm currently learning and it would be very capable to execute a project such as this.
You are envisioning the option to add a pictorial element, almost like achieving a vid podcast quality to the vox?
Now that I think about it, this would complicate things. I think I would ultimately settle for plain color backgrounds for the tiles. The array of tiles (I'll throw together a sketch of what I'm envisioning soon) would simply allow for a colorful, more unique browsing experience compared to an ugly wall of text.
if we were to use imgr/soundcloud. But free usage of these has severe data limitations.
I could pay for a VPS server if needed. I would much rather control my own data than a third party controlling it.
Yeah, the simple tiling would definitely reduce data needed and make it seem less like it was trying to compete with face time or skype VM, or what have you. Preserving the beautiful simplicity is pretty important to me. I may have the server space necessary for this. Here's the crucial question: let's say I put this together and a million people start using it (good kinda problem). How would I make the project sustainable in terms of cost of data storage/transfer in a task/service which can and should be free? This is the conundrum I can't quite wrap my head around. Perhaps I should start building, and see if the kinks can be worked out as I go.
The way you could do this is putting non-intrusive ads on the site and/or setting up an account for donations. These are the two ways that would keep the simplicity and create revenue. Your web host may allow upgrading your plan by buying more disk space and/or bandwidth is needed. Who do you use?