Another problem is that it's hard to experiment in the browser; the stakes are very high when bugs or poorly-conceived ideas quite literally break the internet for users. As a result, the development cycle is righteously slow (~18 weeks), the cost of landing code in Firefox is righteously high, and it's just a difficult environment to try new things.
On Test Pilot we aren't trying to build the solution, we're trying to find it. It's a different model of product development than we've historically had at Mozilla. In this case, side tabs are the start, but you can expect new features and concepts to start coming down the pipeline as we start gathering feedback and data about how people use them.
It solves the problem for me using a combination of auto indentation and auto collapsing of inactive tab trees.
Ideally, I think I'd like something similar to the spatial system of the Classic Mac OS Finder, which I coincidentally expounded upon recently in another thread[0]. Tab Groups goes part of the way there with its spatial organizational system but it lacks the hierarchy and persistence of those Classic folders and desktop.
[0] https://news.ycombinator.com/item?id=11664984