Even extrapolating Moore’s Law’s exponential growth of x2 per 1.5 years would still take ~200 more years to reach the limits of computation — Improvements to particle accelerators aren’t anywhere near as rapid — so we have a long way to go before we reach the end of experiment even if we improve as fast as possible and have almost no surprises.
Not that I disagree with your general premise that some massive surprise could be waiting for us; I happen to think there is, and I get the impression that it’s totally unconventional in the physics community to say “at least one of GR and QM is wrong, but we don’t have any experiment that tells us which”.
I hazard that should we find such a surprise, it will be under somewhat extraordinary conditions not commonly found in nature, and quite likely ill-suited for power generation on the nanometer scale.
I’m not a betting type, but if I had to take a bet I would expect it to be ~”and this is how the conditions of the Big Bang can be recreated and exploited to make something that looks like a star but which will exist forever. Also the interior of our fake star is now a new universe 50 billion light years across even though the outside circumference is ten light seconds.”
Not very useful for cell-scale engineering. Might count as “magic”, but the bigger-inside thing is a serious idea for making the Alcubierre drive less unphysical.
In reality there are limits; there's no reason to think technology can improve indefinitely. We may still be far from those limits, as e.g. Drexler's books show, but fundamental rules such as "Effort requires energy input" are unlikely to ever change, and it's equally unlikely that most modern scientists would regard limit-tech as being magic.
As for the average guy on the street, he probably considers computers to be magic...