Those who believe that one day AI will start improving itself exponentially without limit are ignoring Kolmogorov complexity.
For last week’s puzzle, you may wish to take into account the fact that false positives can run in both directions.