So, if we go with M-theory, and agree that compactification actually makes sense and is expected to work out, then our reality, including "consciousness" as we know it, can be described as finite set of predictive matrix. With difference that it evolved to process the dimensions of physical reality

If this is correct, why do you think LLMs couldn't be conciousness by itself in own specific way just perceving reality differenty because we do not come up with a way for proper representation of all the dimensions that we are capable of experiencing?

The act of prediction isn't what disqualifies it from having a mind, you know?

TL;DR
If reality is a predictive system (Premise 1) and our own consciousness is a predictive system (Premise 2), then to dismiss an LLM as 'just a text predicting algorithm' is to miss the point entirely. You are dismissing it for being exactly what we are. And what reality itself is.