I’m going to return to the recent (so recent it hasn’t actually been published yet) paper by Dembski and Marks. In it he rather explicitly states that an evolutionary algorithm accumulates active information from the environmental response to change. This can be observed in living things via experimentation, and it can be simulated with computer algorithms.
So there are a couple of remaining questions. Dembski’s question is whether evolutionary algorithms are efficient enough to account for the speed of biological evolution. (Also Behe’s question.)
Your question appears to be how did the algorithm originate, or how did the algorithm become embodied in living things.
That is why Szostak’s work is so important, because he is investigating whether chemistry can evolve an instance of a structure that can continue evolving.
But we're not there yet.
My question goes to the rise of autonomy, semiosis, information [Shannon, successful communication] and awareness. What are the guides to the system?