This animation shows the distribution of monthly global temperatures (from https://data.giss.nasa.gov/gistemp/tabledata_v4/) for rolling 30-year periods. Each bar represents a temperature, and the higher the bar, the more commonly that global temperature occurred. In any given 30-year period, the temperatures cluster around a typical value, and very hot and very cold temperatures relative to the mean are equally unlikely. But over time, the "typical" value changes. The centre of the plot is the average temperature for 1951-1980, and a curve fitted to the 1951-1980 temperature data is shown. You can see that what used to be unusually hot weather is now normal. Temperatures abnormally far below today's average correspond to what was normal in 1951-1980.

People sometimes talk about the climatic effect of fossil fuel burning as being like "loading the dice". If you took a fair pair of dice, then the most common score you'd get would be 7, and about one roll in every 36 you'd score 2, and one roll in 36 you'd score 12. If you had unfair dice that rolled 6 more often, then you'd score 12 more often and 2 less often, but you'd still never score 13. The effect of climate change on temperatures is not to make heatwaves like the Earth has experienced before more likely, but to enable heatwaves of a magnitude humankind has never before faced.