I am not averse to using pointers, Ive done a fair bit of C and assembler pogramming over the years. But this seems like a bad idea.

For example, Python assumes strings are immutable. If you use a string as a dictionary key, for example, it will use a hash of the string content as the key, on the perfectly well justified assumption that the string value can never change. That is just one example. If you add code that randomly craps on the Python intepreter's private memory, it is likely to break things very badly.

You are deliberately injecting memory corruption errors into your code, and the resulting bugs can be a nightmare to find and fix. Bugs might pop up in areas that appear to have no connection with the code that is causing the problem. The bug might appear or not appear depending on all sorts of random runtime factors that can make the bug very difficult to reproduce and track down.

If you really need to access actual memory locations from Python, it might be better to look at the buffer protocol (https://docs.python.org/3/c-api/buffer.html), or just write a bit of C code,