Can Gold Ever Not Be Gold?
A post about sine waves, daylight hours, and the origin of the universe
--
There’s nothing like sitting on your front porch and enjoying a cool summer day… while also trying to listen to the sine wave in my own humming voice.
Yeah, I’m a weird dude! But here I am, re-reading mathematician Steven Strogatz’s lovely book, The Joy of x, and thinking about just how funky the universe is. If you don’t know what a sine (sinusoidal) wave is, this is what one looks like:
For a practical example, consider the number of daylight hours in one day. Because days grow longer and shorter relative to our position from the sun, in which we lovingly refer to as seasons, we get more daylight in summer months and less in winter months. When plotted on a graph, this produces a very gentle sine wave. Although much more subtle than the diagram above, the chart below displays sine waves for daylight hours at different altitudes. And humanity has experienced almost alarming consistency in terms of the seasons being the same year after year.
As a machine learning engineer, I’m always thinking about probability. This is because machine learning is all about producing probabilistic inference (or educated guesses) about what will happen in the future. We use fancy terms like standard deviation to express how much a single event could stray away from what that event usually does, called the mean (average).
The reason machine learning can be so dang hard is that lots of things aren’t consistent. Consider my current morning routine. I wake up about 5:00am, drive to Starbucks, then come home to play video games on my front porch until I start work. Based on what I just said, can you predict what my morning routine will be like come December?
The answer is sort of, but mostly no, you can’t.
There are variables that fuel my morning routine. I’m largely able to have my current routine…
