Uh yeah. That's called a Personal Computer - "PC" if you haven't heard of it.
The problem is that technology came along and gave us a "PC" in our hand (cell phones) that wasn't a "PC". What we need is an actual "PC" in our hand. That is: a cell phone that is under the control of the user (and no, a "global single-sign on" sounds absolutely worse than "web3" <- whatever that is).
Failing that, I'd like a computing model that I call PAO, Personal Application Omnipresence, where all of my devices are terminals to my own application server [1].
I have a powerful desktop PC in an inconvenient place, and a much weaker laptop which can be anywhere I am. All I want in life is to pause work on my desktop, pick up my laptop, and resume what I was doing.
As an extension, I never want to think about which machine is running my app. I just want it to run wherever there are enough resources, and be displayed on whatever device I'm in front of.
Network bandwidth isn't a limiting factor these days. Between all the devices I own, I have more compute power than a high end modern desktop. Why do I have to physically go sit at one machine for one task and a different machine for another? Every individual part of this problem has a solution, but no one has put them all together yet
https://moronlab.blogspot.com/2010/01/urbit-functional-progr...
It’s a big project to fix the underlying issues to get us there, but it’s not impossible. It does require more than trying yet another DOA federated system that doesn’t fix the underlying incentive problems that lead to centralization.
https://zalberico.com/essay/2022/09/28/tlon-urbit-computing-...
You know, I actually had this back in ~2011 with the now defunct AlwaysOnPC—it was a really weird service which you could pay like $30 once for an then get access to a remote desktop on a random Debian install on some server they had. It was kind of a security nightmare in retrospect (only the kernel was used for isolation, i.e. everyone was on the same machines!) and I'm suprised it didn't die sooner from abuse (bad actors doing illegal shit from their servers) but it was amazing for the few years it was a thing. You could just sign in from your phone, tablet, or even a random computer at the library using a little java applet and get access to all your stuff just as you left it. I eventually moved on and just started remoting into a server back home but it was very cool at the time to have a fully end to end solution for personal cloud computing.
Of course, we've given up our privacy and autonomy for that convenience. The problem is that we seem to keep ending up on the convenience side of the trade-off.
For instance, I didn't want my sharing URLs to enter directly on my home IP. I felt that would expose my house to an enumeration of vectors that I couldn't possibly control. I ended up creating a small server in digital ocean that runs Traefik on a public address with various public DNS records pointing at it. That instance of Traefik then points to my home servers Tailscale addresses. I actually go a step further by running Traefik on each server so that I can host multiple services on port 443 on each server, but this isn't absolutely necessary.
Hopefully that all made sense!
The FCC does not allow this because of their certification requirements for the baseband processor. You can either lobby for this to change (good luck :-( ) or go the hacker way and build something that is under control of the user, but violates the laws.
Formerly, yes. But from 4G phones on, they rather communicate via DMA, i.e. the baseband part of the SoC has full access to the device's memory.
Also see the relevant section in the Wikipedia article: https://en.wikipedia.org/w/index.php?title=Baseband_processo...