I hope you are joking... You seriously understand physics this little? A little guiding hint: Every process generates heat, that has to be transferred - whether in a self replicator, or in bacteria. Bacteria solve the problem by a huge amount of redundancy in structures so local heat flow's negative effect on overall performance is lowered. However, this is bad design, as creating redundancy has its own, higher, energy cost. Self-replicators solve the problem by - ignoring the question (They do not transmit information beyond the basic entropic generation limit of the system, so they are cooled by an external substrate, and the researchers pretend the carefully designed substrate just "exists" naturally )Starglider wrote: That makes no sense, as both machines and organisms can and do have an indefinite amount of external energy and resources supplied. Thermodynamic arguments are really only relevant to 'grey goo', where it is indeed necessary to make a thermodynamic case for why the proposed self-replicators will substantially outperform bacteria.
As for your "indefinite external energy": I am not a biologist, but seriously - you do have some inkling of extinction events, famines, natural selection, etc?
As for machines - you do have some inkling of energy crises ?