Git was already designed from the outset specifically to be peer to peer. The central server is the addition to that model, not the other way around. I have not looked into this project, but pitching whatever it is as P2P for Git just sounds like the devs learned about Git though GitHub and don't fully understand it.
We're quite aware that Git is a distributed code collaboration tool, even trying out the email based flow in the past :)
You may notice that the title is "peer-to-peer code collaboration stack built on Git", emphasis on "built on". We're using Git as a storage and transport mechanism for code collaboration data. So we're building a local-first tool for the social data, e.g. patches and issues.
Alternatively, the fact that the very first thing basically everyone who isn't Linus has done with vanilla git is introduce some kind of central authority might suggest that what git was "specifically designed for" is more of an outlier than you want to admit.
Everyone knows Linus invented and "specifically designed" git as a drop-in tool for his existing email-patch-based kernel development workflow, which is not how 99.9% of the rest of the world prefers to operate these days.
The difference there is that code can still be pushed (or pulled) between the git repo and the new centralized instance after forking. Anyone coming from pre-git centralized source control (or shudder the NAS of periodically rsync'd folders) recognizes that this represents a significant difference from that earlier world.
On topic, though, I have no idea what Radicle's value-add is, though.
