If you get as bored as I do with the bad math and biological ignorance of the creationists who constantly claim some mutation or another is too improbable, check out what happens to Cornelius Hunter when a commenter named Duke goes to the trouble of doing the calculations for the evolution of a mitochondrial protein for him. Hunter just yanks a "million million" figure out of his ass, and then gets confused when Duke actually takes the time to put together a lower bound for the number of corn mitochondria:
Just to give an idea of the scale of numbers we're talking about here. Let' say that every corn mitochondrion has the minimum number of genomes, 2 (although that number will probably vary, since having more or fewer genomes is a common genetic abnormality, even in humans), and let's say that every corn cell has 100 mitochondria in it (Google "How many mitochondria in a plant cell" and take the first result), which is the low estimate for plants. Let's also assume that each corn plant has a millions cells in it, which is ridiculously low, but I'm making a point here. That means we have
2 genomes/mitochon. * 100 Mitochon/cell * 1e6 cell/plant
That means 2e8 mitochondrial genomes per plant, and that's a ridiculously low estimate. Even so, that takes you to one five-thousandth of your legendary "million million" number.
But wait! There are 4e12 corn plants grown every year. That means there are 8e20 (that's 800,000,000,000,000,000,000) separate mitochondria, every year, each one ready to randomly stumble upon a simple protein in that tiny 1e12 haystack.
Hunter's only reply after a lot of obtuseness by design is to declare the comparison "apples and oranges and not making sense".
Hunter also reveals that he doesn't understand that genomes do not copy themselves perfectly:
Hunter: You can have as many copies of the mitochondria as you like in that line of corn, they share the same genome.
Duke: Um... Reference? You're saying that mitochondria copy themselves perfectly? With no mutation at all? You'll have to show your work on that, I'm afraid.
Nor does he understand how common gene duplication is:
Hunter: And finally, the vast majority of the de novo gene we're talking about has high similarity to two existing segments in the genome. You're saying that unguided mutations just happened to create a new gene that mimics two existing sequences?
Duke: Um... Yes? This sort of thing is common and observed. Part of the genome is copied and added onto the genome twice. It's like some unguided copy-editor added two versions of Chapter Four to a book.
And for the best moment, one of Hunter's fans named Natschuster tries to come in and save him:
Natschuster: I'm not convinced that it could turn the corn into a new species. If I make small random changes to my car, I might get lucky, and one of the changes will improve my cars performance, but I don't think that it will ever turn my car into a truck.
Duke:Really? You can't see turning your car into a truck in small steps? Your car into a truck? Really?
Are you serious?
Of course, your thought experiment of turning a car into a truck is wildly wrong, betraying a gross misunderstanding of what your opponents believe. What you'd actually be doing is copying your car, repeatedly, thousands upon thousands of time, with minor variations. Those copies in the second generation that cease to function are destroyed; those that are better at performing whatever you need them to do are kept (in the case of the car-to-truck transition, this might be hauling cargo rather than carrying passengers) to breed the third generation. Repeat as necessary. This sort of thing does happen and has been demonstrated happening.
Ideological denial, meet reality, courtesy of Duke. Check out the entire exchange, it's a howler.