Just a friendly comment from someone who's actually written bug-free code: I find your assertion that it's "the very laws of physics" that makes bug-free code impossible to be incorrect.
There are disciplines, starting with "Software Engineering," in 1957 (if you can still find the orignal NATO report, it's a goldmine) and extending through many methods today that allow one to write bug-free code.
Unfortunately, it requires discipline, which most programmers lack. It means designing before you code, testing the design before you code, and then scrupulously coding with precision. Testing can only prove the existance of bugs; it cannot tell you there are no bugs left.
If bridges in the U.S. were built the same way we build software today, we'd see a lot more bridges collapse (like the recent one in Minneapolis-St. Paul, which failed for lack of maintenance). Because there's a rigor to the design, testing and construction processes, there are few such "bugs" in bridges. And, bridges are designed by professionals, who are licensed and/or bonded because they meet certain competency requirements.
Finally, your assertion that it take massively longer to find all bugs before release is not correct, unless you're talking about testing to find bugs. If you design from a bug-free perspective and intention, the total project time from conception to delivery has been demonstrated to take about the same time and person-hours to produce a superior result.
Think about it: For systems you can't ever fully test (defense systems, medical instruments, nuclear plant safety controls), bug-free programming is the only option to ensure the system will work the one time you actually need it.
