Very much a highly technical work in progress still, but promising and inspiring:
Set your expectations appropriately: Don’t think: “when can I download this and start running signal”. Don’t think: “oh, pinephone not librem”. Do think: “microkernel capability-secure OS on real mobile hardware, OMG!!” 
https://genode.org/news/road-map-for-2022
https://genode.org/news/index
Impressive to see their sustained work over the years on such a big and ambitious project.
Capabilities, you say? Minimising to a superficial slogan with some truth to it, you could think of capabilities as “hardware switches in software”. This is desperately needed, because there is no actual hardware switch between one program and another or one internet service and another to isolate them by default and only connect them in the ways we need as users. Don’t mix up capability security with the assorted tech systems that just happen to use the same word but are otherwise unrelated: for example “linux capabilities”: completely different thing that happens to use the same name. Good introduction to “real” capability security - sorry about the lack of HTTPS.
Some context here is that Google has been quietly working on its own capability OS, Fuchsia, for years also. Good to see capabilities finally heading towards the mainstream in some places (see also agoric.com), but I’m very glad somebody is working on a non-Google alternative. Fuchsia seems likely to replace linux on android and a lot of IoT devices (caveat: I’ve not been following news about it much at all). At some level that’s a good thing. However, one problem is it is friendlier to closed-source binary code (hardware drivers in particular), which many have commented is likely a bad thing for free software, hence society. And of course, Google hasn’t exactly set itself up to be a trustworthy institution to make systems on which to build the critical foundation of our society’s communication.