I have a GNU/Linux phone I carry in my other pocket. Here are the biggest issues I can see:

1. Driver support for components in the mainline kernel (lets you install any distro and things like camera, Bluetooth just work)
2. Power management; turns out it is a hard technical problem to have your phone suspend to save energy, while being awake enough to know what and when to turn back on to receive chats/calls, playback music, etc
3. Cameras have a lot of stuff beyond drivers happening behind the scenes these days in software that would need to be developed, especially given it is a big reason people choose their phones for
4. Phone certification is tough, this has stopped even companies like Fairphone from shipping their devices worldwide, I imagine even harder for a device like the Purism Librem 5 where you can literally upgrade Wi-Fi, BT, and cellular generations like a gameboy cartridge
5. App ecosystems take a while to build up, it is a chicken/egg scenario. I think things are in a useable state for all the default apps an iPhone has, but if you want Uber, Uber Eats, you either have to draw even more power essentially running Android via Waydroid, or use a typically more janky web app that may be missing some features

I see a lot of people responded with a true clean slate, but really, a fork is a clean slate.

It's not like Graphene, or Lineage, or any others would stop working. More maintainers would be needed for security issues, but way less than to get (non-Android) Linux phones up to speed.

Many graphene users, myself included, use all FOSS software from outside Google's store.

Most of them aren't necessary to most people, but the main concern is features that should reasonably be part of the core Android experience being removed, or features that have no reason to be reliant on Google at all.

For example, GrapheneOS can't support the detection of your phone being quickly ripped away from you to auto-lock the device, even though that should only require onboard sensors and processing, and it can't support the additional custom clocks for lock screen customization, because Google decided those would be built into the Google app then extended to Android after, rather than being built into AOSP.

You can reasonably see a future where other functionality gets put into these proprietary blobs too. Maybe the launcher becomes proprietary and GrapheneOS has to use or develop a separate FOSS one that might not support all the same features. Maybe charging optimization gets locked behind proprietary code because Google claims it uses "special algorithms" to adjust how your phone charges. Maybe Private Space gets turned proprietary because Google claims it needs special security features.

That's why it's particularly concerning, because in the future, Google could just decide that any number of features aren't part of AOSP anymore, and now GrapheneOS either has to give them up entirely, or make/find an alternative.

You said 'When' - does that mean you stopped?

Yes, it is based on Alpine.

Its going to be motorola and the first phone that can support it wont roll out till 2027

I think they were asking you to elaborate

How's letting Google doing this shit end up helping us ?