This is contrary to the proposals made by proponents of Universal Darwinism, which suggest that copying with variation and selection are ubiquitous physical and chemical processes which can be used to model a wide variety of systems.
As readers may or may not be aware, I am both a proponent of Universal Darwinism and a fan of Cairns-Smith's ideas about the origin of life, so which side of this argument to support is a kind of dilemma for me.
The first thing to say is that Cairns-Smith got it wrong in detail, that simple physical and chemical systems do evolve and exhibit adaptations in much the way that he argued against. Cairns-Smith gave too much weight to the idea that simple physical and chemical systems lacked high fidelity copying - and would therefore undergo an error catastrophe, or a "mutational meltdown". We now know that with positional inheritance, high fidelity copying in prebiotic systems is ubiquitous.
Other problems besides lack of high fidelity copying mean that these systems typically do not go on to launch systems capable of open-ended evolution. One such problem is having a genome with a "low information ceiling" - and Cairns-Smith did discuss that problem. Another such problem is "local exhaustion". Dissipative systems destroy the energy gradients that they feed off. Unless they can continually find new energy sources, they will exhaust their own energy supply and die out. This happens with lightning strikes, for example. They typically do not last for very long, and the reason for that is because they run out of fuel.
It seems possible to me that prebiotic "chemical evolution" will involve adaptations that are on the pathway towards the origin of life (as Cairns-Smith argued against). However, I don't think negating Cairns-Smith's point about the relevance of ochemical evolution has all that much effect on his other arguments. Clay mineral crystals still seem like attractive candidates for the first living things, for essentially the reasons that Cairns-Smith gives in his books on the topic.