In this article you say that you can learn a lot from a single data point, and then give the example of an alien learning a lot about humans by finding a Ford Focus. I don’t think this example illustrates your point very well. The reason you can’t learn much from a single data point is because you don’t know if that data point is average or a big outlier. For example, an alien can learn a lot from a Ford, but can also be majorly misled by a playground space ship. The alien doesn’t know whether the artifact is pretend, fake, unusual, outdated, and so can’t tell much for certain about humanity, beyond just ‘one human made this at one point’.
In this article you say that you can learn a lot from a single data point, and then give the example of an alien learning a lot about humans by finding a Ford Focus. I don’t think this example illustrates your point very well. The reason you can’t learn much from a single data point is because you don’t know if that data point is average or a big outlier. For example, an alien can learn a lot from a Ford, but can also be majorly misled by a playground space ship. The alien doesn’t know whether the artifact is pretend, fake, unusual, outdated, and so can’t tell much for certain about humanity, beyond just ‘one human made this at one point’.