Wolfram's theory was already very politely but firmly shown to be wrong 20 years ago, by people who took a look into it. Scott Aaronson for example. https://arxiv.org/abs/quant-ph/0206089
>"To prove that 110 is NP-complete, what is needed is to show that Rule 110 allows efficient simulation of Turing machines"
But if the universe is a 'computation', then the efficiency of it doesn't really matter right?
Similar to creating a Blender rendering, it doesn't matter how long each frame takes to render compared to any other frame, as long as the observer sees a smooth sequence of images.
We already see this with existing simulations, though they are of much simpler objects than conscious beings. If I simulate water sloshing in a bucket, it doesn't matter how long the simulation runs, the results are the same.
Now, I may add in short cuts to make it closer to real time to support practical applications which leads to different results, but a slower machine running the same shortcuts will see the same differences.