> What does it mean to "create programs that are designed to keep growing"? If they keep going by executing a finite program (if modified by a non-static program, what created that program?) on their code and resuming, that would not be more powerful (Turing machines can do that).
To me, your question is like asking this: Do all the programs that we humans ask computers to compute come from a finite program? Is life itself powered by a finite program?
A program which keeps growing can make use of all the code it has seen before, all the results it has seen from running code previously. Although any of its inputs are finite strings, there are no theoretical limitations regarding the source of the inputs it can receive. They come from outside.
The core idea of the undecidability of the halting problem is that we can construct a machine M that can ask the halting machine for a bit, flip the bit, give it back to the same static halting machine code, thus forcing it to produce the wrong bit.
Any human in that situation would change their understanding of the situation, and we would tend to afford them a chance to change their answer based on their new understanding. Being able to adjust execution strategy after seeing new information is core to human life. Why do we neglect to provide computers opportunities which are available to all humans?
A somewhat trivial reply to Turing's original negator machine M would be to emit a new version of the halting code which checks for the exact code given by M and then returns the opposite of what was returned when M calls out to H. Then, in the argument, we would need to address the possibility that the H machine called out in M to may provide a different answer when we pass it all back through H again. The argument wouldn't work as-is if we afforded H the ability to respond to the situation.
What would the statement be? "It is not possible, given the information the halting code just received, it could change its answer response to the contradiction of its previous answer." A child can figure out this simple trick - we all understand why it's a trick - why do we not let a computer the same opportunity in our theory?
> So long as you have enough memory, ...
True enough. My point was in fact that if the inputs our machine receives are all needed, we would run out of memory eventually.