I'm learning about Turing Machines for the first time. I think I get the gist of it, and I'm asking myself the question, "What's the big deal?". Here's my attempt at an answer:

1. Consider the idea of Thingspace. Things are there components/properties. You could plot a point in Thingspace that describes everything about, say John Smith.

2. You could encode that point in thingspace. Ie. you could create a code that says, "001010111010101001...1010101010101" represents point (42343,12312,11,343223423432423,...,123123123123) in Thingspace.

3. A Turing Machine seems like it basically says, "If the state is 0001010101011...10101011, change it to this." It's looking at things at a really really low level - the level of individual bits. These bits are a map that, in theory, seem to actually represent the territory with perfect accuracy (or really, it's capable of doing so).

So a Turing Machine could:

a) Look at a model of reality on the lowest level possible.

b) Manipulate a model of reality on the lowest level possible using a).

So back to the original question - what's the big deal? The big deal seems to be that "When you operate on such a low level, you could 'do anything'. The model could be perfectly accurate, and you aren't limited to making coarse adjustments to the model."

To what extent is my understanding accurate? Can anyone elaborate?

EDIT: It seems somewhat obvious that "if you had such precise control, you'd be able to perfectly model reality and all of that". But to me, the hard parts seem to be:

1) Creating a physical machine that does this. That has enough memory, that computes things quickly enough, and that is wired to do things based on state. (hardware)

2) Giving the machine the right instructions. (software)

I sense that these initial impressions are ignorant of something though - I just don't know what.