A future direction: civitai as the dockerhub of SD models

You guys seem to really be looking for the BIG uses of SD, and have already invested in AI farms, so I figured maybe you would want to take things even further.

Based on https://www.reddit.com/r/StableDiffusion/comments/17vxaob/comment/k9f35zw/?utm_source=share&utm_medium=web2x&context=3

What about getting involved in a new layer to SD generation, where instead of manually downloading HUGE models, that you only actually need 1/10 of.... your client dynamically grabs fragments of models, that are specifically related to your prompt?

cutnpasting fromn that post:

For computer-aware people, this could be considered somewhat akin to "dockerhub", but for model images instead of compute images.

With docker, you can tell it, "build me a container with apache, mysql, and redis, and then do (this)" It will go download apaache,mysql, and redis .. unless it had previously downloaded any of them. And it will then continue on to build the final combined container and run it.

So what if civitAI became like dockerhub, but for downloading more narrowly tailored models?