Weather/room temp wise we probably never will. I'd rather think of my environment in terms of 0 to 100 than in terms of -18 to 38. For science and engineering, Celsius is ideal, and I can convert between the two in the very rare occasion I need to because I'm not an idiot who can't do basic math.
That's entirely a matter of habit. There is nothing special about 0°F (random point in the cold range?) or 100°F points (random point in the hot range?), you've been lied to.
We don't think -18°C to 38°C, we think -50°C to +50°C (regular Celsius weather thermometer, covers almost any temperature observed on Earth), with 0°C differentiating between snow/ice, "wintery" weather, and rain/mud, "non-wintery" one. That's how we know whether to take umbrella (no point if it snows, hat is your best friend), what kind of shoes are the best fit - cold-resistant or highly waterproof - or which kind of jacket is gonna fit the situation. Melting point of water is actually incredibly important weather-wise and entirely ignored by Fahrenheit scale.
When it's not winter, normal range is 0-40°C, with 20°C designating comfort temperature.
For science and engineering, Celsius is ideal,
The SI base unit for temperature is Kelvin with 0 K being the coldest possible temperature. 273.15 K is the melting point of ice. But it’s a lot better suited for temperature differences. Celsius is only a derived unit.
And well, all units and measurement systems had a lot of changes over time because some things turned out to be impractical or inaccurate.
Initially Celsius had 100° as the freezing point of water, 0° as the boiling point of water. Fahrenheit had 0° as the coldest temperature he could produce and the (wrong) average human body temperature at 90°. Kelvin was initially defined via Celsius, that got reversed, they have the same scale. There is also Rankine, which starts at 0 like Kelvin, but uses the Fahrenheit scale.
And the US partially uses SI units anyways, all units are derived from them to use their superior base unit definitions. This system came into existence to have unit definitions that are better reproducible and change less over time. Since everything was redefined and all numbers changed anyways, they also tried to make use of the "new" decimal representation of numbers. And new unit names were nice to create some general units, in contrast to foot and pound, which were always different from place to place, at times even from city to city.
I don’t expect the US to ever switch. The US switched to international yard and pound instead of switching to a decimal system. After US, UK, Canada, Australia, New Zealand and South Africa agreed on that one, all countries who remained using these units had a uniform definition for them. Since then you don’t need to know any longer which yard or pound it was. Though not all units got standardized by that.
And some countries didn’t drop all old units and metricized some instead. Even SI kept the ton(ne). You can’t know what 1t exactly means without knowing the context, it can be 2240lb, 2000lb or 1000kg (~2204.6226lb).
Aviation is already backwards; aviators give distance to travel in nautical miles, visibility in statute miles, altitude and runway length in feet, speed in knots, weight in pounds, volume in gallons, and temperature in celsius. My favorite is the standard adiabatic lapse rate is given as 2°C/1000 feet.
