Man in the Middle

Our astrobiology graduate seminar has started, and last week it got interesting as we pondered "weird life".

I'm a co-PI of the Penn State Astrobiology Research Center, and as part of our program we run an Astrobiology graduate seminar, which is generally topic based.
This year it is more student oriented, but occasionally the faculty are invited to kibitz.

Last week the theme was "Weird Life" - we fling this at all groups of students we can gather in astrobiology, in the hope that someone will eventually have a flash of insight.
The essential topic is: "if there is life out there that is different, ie not RNA/DNA with sugars and proteins, then could we know, and what would the signature be"?

The discussion gets more sophisticated as we iterate, and the field progresses.
A fair amount of time is spent on alternative carbon life - possible genetic material, metabolic paths and excretion products.
Then we get to spend time on non-Carbon options, and it gets fairly speculative.
I like to think Greg Egan would feel at home, though I sometimes feel the young'uns need to get out and read a bit more - embarrassing when they start to rehash old tropes from '70s science fiction.

We got quite far out there this time, but one topic that came up went down an intriguing path.

If we grant strong AI is possible, which is a possible starting hypothesis, then it is likely that in the not too distant future, we will have intelligent life based at some level on silicon chemistry - artificial life.
Note that it doesn't matter much if it is GaAs, or photonic, or some other descendant technology. The final product will most likely have some lineal descent from doped silicon devices with metal and silicon oxide paths.
Coupled with von Neumann machines and mechanical technology capable of directed reproduction, this would set up a machine silicon ecology.

Given finite resources, and either random errors in reproduction or directed variations, there would be some sort of selection process - ie there'd be variation, reproduction and selection through some fitness metric.
ie artificial life would evolve.

So, the question was, could silicon life like that arise naturally?
The immediate answer is "no, most likely not", and that is due to lack of selection.
Nature will provide variations in blobs of silicon with oxide and metal veins, and there is some sense of reproduction - rocks to beget rocks - but no obvious selection, except maybe for some trivial tensile and compressive strength, and solubility.

But... are we not natural?
Would strong AI with von Neumann machines be less natural once they entered the ecology merely because at one point they had symbiotic, or parasitic relations with carbon based intelligent lifeforms?
Well, that becomes a philosophical question...

However - we can step back a bit.

Another, plausible hypothesis, is that zeolites or something like them, may have played a critical role in biogenesis.
Nanopores on clay surfaces - aluminium silicates - may have provided an important role in pre-RNA chemistry, possibly providing catalysis for some of the pre-biochemistry.

Then we let go of our disreputable, dumb-as-rock, progenitors, and entered the carbon phase.

But, maybe the whole RNA/DNA carbon chemistry thing is just a phase.
Maybe all of life is just a way for zeolites to jump the awkward adolescence and go straight to full intelligent reproduction (yes, I know, we're getting teleological here...).

So life starts in silicates, goes through a brief few gigayear carbon intermediate phase, and then continues back into silicon chemistry, with a rather different ecology.

Be amusing if true.