Yes, computers work in the real world.
But if you want to see what a computer is doing in the real world, open it up and look.
What you're doing -- as always -- is confusing the system which includes a computer and a knowledgable interpreter of its action with a system including the computer alone.
If I build a tornado box, I create a real tornado which behaves like a tornado in the real world. The "world of the tornado box" is real.
If I simulate a tornado, there is no tornado. To understand that a simulation is going on, it requires an interpreter whose brain responds correctly to the virtual output.
As has been established, consciousness is a real process in spacetime. It has location and extension in space and time. The hologram-like-thing being produced inside my skull by my body isn't being produced a hundred years from now in Paris, but rather inside my skull right now.
If I want to build a machine that produces a tornado or a hologram, or any real spacetime event, I can't get away with pure programming; instead, I have to build a machine whose actions in spacetime produce an actual tornado or hologram in spacetime.
The hologram-like-thing which is the mind, which is consciousness, which is the phenomenology, is no different. It requires direct and real physical causes, not virtual ones.
A virtual simulation of a tornado isn't a real tornado in spacetime. You can inspect the actual material and energy apparatus of the computer and you will find no real tornado.
And if you want a hologram, you can't program a general purpose computer and dispense with all other hardware except what's needed to run a program. If you want a real hologram which acts in real space and time, you need to connect that computer to some sort of physical apparatus to make the real spacetime phenomenon occur.
The laws of physics demand this, without exception.
That includes all bodily functions, without exception.
If you want real digestion to occur in the real world, if you want a real pulse in the real world, and if you want real consciousness in the real world, a programming-only solution is simply impossible.
What you're doing is changing horses in mid-stream.
You say that the computer can produce a real-world effect, but when I say, OK, examine the real-world computer and show me, you then pull a switcheroo and move your frame of reference away from the physical world and into a virtual world which has no independent reality but which depends on an interpretation by a 3rd party observer.
Obviously, neither your phenomenology nor mine is dependent on someone else observing our brains.
That's why you continue to be 100% wrong about this point.