Linux phones are still behind android and iPhone, but the gap shrank a surprising amount while I wasn’t looking. These are damn near usable day to day phones now! But there are still a few things that need done and I was wondering what everyone’s thoughts on these were:
1 - tap to pay. I don’t see how this can practically be done. Like, at all.
2 - android auto/apple CarPlay emulation. A Linux phones could theoretically emulate one of these protocols and display a separate session on the head unit of a car. But I dont see any kind of project out there that already does this in an open-source kind of way. The closest I can find are some shady dongles on amazon that give wireless CarPlay to head units that normally require USB cables. It can be done, but I don't see it being done in our community.
3 - voice assistants. wether done on device or phoning into our home servers and having requests processed there, this should be doable and integrated with convenient shortcuts. Home assistant has some things like this, and there’s good-old Mycroft blowing around out there still. Siri is used every day by plenty of people and she sucks. If that’s the benchmark I think our community can easily meet that.
I started looking at Linux phones again because I loathe what apple is doing to this UI now and android has some interesting foldables but now that google is forcing Gemini into everything and you can’t turn it off, killing third party ROMS, and getting somehow even MORE invasive, that whole ecosystem seems like it’s about to march right off a cliff so its not an option anymore for me.
The same way it was done with Google, Samsung and Apple. Just has to become more popular until banks and credit card companies will have to work with developers to make it happen.
Again, it will have to require the compliance of OEMs. However I see the entirety of these systems disappearing soon as more OEMs want to lock users into paid subscriptions for such features.
I'm not convinced this will ever be useful. Several of the largest tech companies on the planet have tried and all have failed miserably to produce anything useful for decades at this point.
For Linux it could be tied in with terminal commands using an LLM:
"Install Firefox" ->
apt install firefox"Open Firefox" - >
firefox & disownaichat --executealready turns natural language into terminal commands through any OpenAI-compatible API (and OpenRouter provides free Deepseek R1/Kimi K2 access), so there just needs to be speech-to-text.