The Game of Life is a 2-dimensional cellular automata (CA), so given the 1-dimensional rule 110 has been proven to be universal / Turing-complete [1], it becomes less mysterious. Albeit the complexity of the system required to set it up to do anything "useful" would be prohibitive.
I'm currently finishing up my OU MSc and the project I picked was specifically around cellular automata - only in this case relating to them calculating any arbitary automatic sequence - which are sequences you can create from finite state machines - that really opened my eyes to the fact these sorts of very, very simple machines can, with the right (and rather complex) setup, be made to do pretty much whatever you want from a computational PoV. In that paper by Rowland and Yassawi they give a constructive proof to calculate the required update rules for a CA that outputs any particular automatic sequence. That itself gives some hints at some ways of deriving the input and rules for these systems to do some particular job. [2]
I know Wolfram often gets dunked on for ego/hubris but in Chapter 11 of a New Kind Of Science he goes into how the Rule 110 CA can be setup to "calculate" (output) other CAs. From there it starts to become a little less mysterious that these systems can generate behaviour you could imagine running on a CPU of some sort.[3]
I'm currently finishing up my OU MSc and the project I picked was specifically around cellular automata - only in this case relating to them calculating any arbitary automatic sequence - which are sequences you can create from finite state machines - that really opened my eyes to the fact these sorts of very, very simple machines can, with the right (and rather complex) setup, be made to do pretty much whatever you want from a computational PoV. In that paper by Rowland and Yassawi they give a constructive proof to calculate the required update rules for a CA that outputs any particular automatic sequence. That itself gives some hints at some ways of deriving the input and rules for these systems to do some particular job. [2]
I know Wolfram often gets dunked on for ego/hubris but in Chapter 11 of a New Kind Of Science he goes into how the Rule 110 CA can be setup to "calculate" (output) other CAs. From there it starts to become a little less mysterious that these systems can generate behaviour you could imagine running on a CPU of some sort.[3]
[1] https://mirror.explodie.org/universality_in_elementary_cellu...
[2] https://arxiv.org/abs/1209.6008
[3] https://www.wolframscience.com/nks/chap-11--the-notion-of-co...