American Scientist discusses the trend for changes in how well people score on intelligence tests and notes that the Flynn effect, whereby the population has been scoring increasingly well on intelligence tests over time, seems to be slowing down or reversing in some places.
It is well-known is psychology that performance on cognitive tests changes over time and across populations, which is why the most widely used tests (particularly the Wechsler series) have different versions for different countries, and are re-released every few years with new comparison data.
An IQ score is always relative to the average performance of the rest of the population, so an IQ of 100 always means you score the same as the average of the population on a current test.
As new tests are released, this average may shift, so it is difficult to directly compare IQ results from previous versions of a test.
On old tests, however, it was noticed by Flynn that people were scoring better by about 3 points per decade. The American Scientist article notes that this effect is starting to slow down or reverse in some places though.
Does this mean we’re becoming less intelligent? Probably not. It likely reflects the fact that the skill set of population is changing and that we become practiced at different tasks at different rates as modern life develops.
As an aside, IQ tests considered trustworthy by psychologists rarely go above 160, so anyone quoting a 160+ IQ is likely to be talking nonsense.