Nope - you didn't mention time horizon in your article. Top tip - describing finite things as infinite is bad style.
What seem to have wanted to say is
1) that the number of bugs in a non-trivial piece of software is sufficiently large that they will probably not all be found before the software is obsolete. Which is dull but probably mostly true (given the wriggle room in "non-trivial" and "probably")
2) that offering a bug bounty because of this large latent pool of bugs is pointless.
This second one is just not valid because
1) bug bounties encourage reporting of bugs
2) not all bugs are equal - there are different costs for finding them in a particular product and a bug bounty will encourage people to find and report the easier ones.
3) There are finitely many black-hats. As the easy-to-find bugs in the pool are exhausted then the cost per bug to the black-hat increases in this product.
At this point the black hat has a choice - pursue finding harder bugs in product A (which has a bounty) or go for the easy to find bugs in product B (which doesn't). Blackhats are running a business - they will go for the return on investment in product B.
This neglects the very large positive advantages of reporting which others have covered earlier (discovery of systematic issues, healthy ecosystem of investigators, disincentive to black-hats).
At this point your "bug bounties are useless" falls apart because it neglects the fact that black-hats are running a business - spending $10million to find a bug in Apache will not happen because the blackhats cannot get a return on their investment. They will spend $10k looking for exploits in Flash, or PDF, or other low hanging fruit.