Some perspective from my own remake experience (I run 4 remake projects currently, 3 of them stagnant because i'm alone in that 3)...
OpenArena only got off the ground because i'm my own lead artist, animator and skinner, because no one else would (well, Enki and kit89 helped 0.7.X with skins and more models, but that was it). When you go the "GPL game" route, there is a cursed stigma already around it that immediately associates it with low quality/scavenged work (BZFlag, Blob Wars, Open Quartz) and artists that could help won't understand what that's about and just simply turn your back to you, or their practice opportunities for their portfolio will be foiled since the license allows derivatives. There are a lot of snobby ones about that. Also typically associated is license philosophy zealousism, i.e. "doing it for the Stallman, for GNU/Linux and Free Software, and Freedom". That goes a long way towards a true motive for a project to the point it just
scares people that you're not doing it for the game but for promoting the license. That also builds the low quality stigma. It also seems to depend highly on the role of the leader. If you're a coder, be prepared to have patience with a lengthy put-up of placeholder programmer art. If you're a creative artist, if you build it, they (coders) will come. Eventually. OA used to live on stock unmodified fresh q3 VMs until Sago007 came along some years into the project (and he pretty much saved it with new gametypes

). Also,
there's no such role as "leader". I am not sure if I ever saw a game get off the ground for the leader, just being a 'leader'. You have to do something, or nothing will happen. Writing a design doc won't fabricate a team with the game right in front of your eyes. Even a role of "web master" won't do anything good these days.
And by the way, the "GPL game" scope is a LOT broader than the "Remake a Quake TC" scope you appear to be discussing. Quake TCs are relatively obscure, even the really good ones like Malice. The only one the majority remembers is the horrible XMEN one. Going this way just limiting production opportunity. Also since you're deriving from community work, you'll also need individual permissions as there isn't a singular company to hold the rights wholly to the produced work. This will essentially put your TC remake ideas in a potential legal development hell. For Malice's case though,
it is a company, the developers Ratloop are still very much around today. Can't say the same for ATF though. I liked ATF too and I also wish it had a bit more polish than it had, especially considering that in 1997 when
no one could ever model a convincing low poly female for Quake. "biker bitch" still haunts me to this day.
And yes, the audience is small. 6 people will download your hard-worked mod, and even less will play it through, unless it is directly derivative of a mainstream game because then you'll just hijack an established fanbase. That's the ideal motive behind all this PSP Quake modding, and is why you hardly see the originals shown about (Bulletz).
Today, Darkplaces has IQM formats, skeletal animation blending, csqc physics,... i don't think there's an artist's excuse in the world now.
BUT REMAKES ARE HARD EVEN WHEN THE DESIGN DOC IS ALREADY DONE FOR YOU!
Start small!! Make a game where you get one gun one level and shoot things just to familiarize with a pipeline first. Playing 'replace null assets' won't make you learn fast, it's more trial and error and you also have less creative control. Even Navy Seals Quake only started off small as "M-16 replacement with unskinned terminator hands", and remember... that is the earliest beginning of the enormous Counter-Strike empire.