Word to the wise: Exhaustive search is probably not the way to go, at this point.
I put the Python code from that Math.SE post
here, with my own C++ translation; the latter is not appreciably faster.
Wu's "Introduction to Pipe Theory" (thanks, Jean-Paul) lists some candidate Goldbugs such as the 173-digit number
14210043931054384697281797631385857688949646035466487917230047060794326605647294255025896867407138484058379599443343129460393572580187625386158635949233364384758021881468840
but I myself wouldn't know how to go about proving or disproving that number's Goldbug-ness or even its Goldbach-ness. :) (I'm sure people with more math and more fluency in mathematical programming will find at least one of those tasks easy; I bet that that number is the sum of two primes (easy?) and is not a Goldbug (hard?).)
Craig, does your bounty apply to the finder of any new Goldbug, or only to the provably next Goldbug? Presumably the former, but I suppose that detail might affect some readers' strategies. ;)
(Not me; I've satisfied myself that I've got no chance here.)
–Arthur