Thanks for the clarification, I'm guessing I worded it poorly.
First, a nit pick, I said "oil" (not petrol) but petrol is a product made from oil and mentioned in the article so its not a big deal.
My claim is that we can make "petrol" out of the air if we want [1]. It does not violate any laws of physics (or thermodynamics). Using petrol to make petrol this way is not effective since the translation leaves you will less petrol than you started with. Using power supplied by a nuclear reaction (as an example) to make petrol this way will convert locally generated nuclear power into transportable "chemical" aka "petrol" power. As most of the energy in the fischer-tropsch process is heat you can use concentrated solar just as easily. the electrical energy used is primarily in running the cooling pumps. Given that "oil" (and from oil, petrol) can be made out of renewable feedstock (CO and H2), I claim we won't ever "run out".
The economic principle I was alluding too was substitution, where an equally effective good at lower cost replaces a good at higher cost in the market. If petrol is $10/liter, cars powered by electricity will replace cars powered by petrol. That is an economic substitution of equal function at a lower cost.
On the Algae process. At least the NASA version required much less heat and cooling the the fischer-tropsch process and in the NASA case no arable farmland (you can grow algae in clear tubes bolted to the sides of buildings if you want). Providing yet another alternative to acquiring long chain hydrocarbons that are more cost effective to create than using a nuclear reactor connected to a Fischer-Tropsch reactor. The challenge of algae systems has been capturing a low marginal cost of additional oil from a process that was not continuous (you want to make a gallon you pay the cost of running the cycle for 1,000 gallons regardless). If however you can make a gallon for the marginal cost of producing an extra few gallons of algage scum, then you can amortize your costs more effectively and carry less risk. Both of these things encourage adoption and production.
Is that clearer? And for the record I don't think I've changed my argument at all here, just filling in the bits which you might have missed.
[1] http://en.wikipedia.org/wiki/Fischer-Tropsch_process