As long as you never intend to reproject/transform the data to another coordinate system, it doesn't technically matter what srid you use. However assuming you don't want to throw away that important metadata, and you do want to transform it, you will want to ensure your assigned srid matches the data, so postgis knows what to do when the time comes.
So why would you want to reproject from epsg:4269? The answer is because certain types of queries (such as distance) make no sense in this 'unprojected' world. Your units are in decimal degrees, and a straight measurement of x decimal degrees is a different real distance depending where in the planet you are.
In your example above, someone is using epsg:32661 as they believe it will give them better accuracy for the are they're working in. If your data is in a specific area of the globe, you can select a projection that's accurate for that area. If it spans the entire globe, you have to choose a projection that does 'ok' for your needs.
Now fortunately PostGIS has a few ways of making all this easier. For approx distances you can just use the st_distance_sphere function which, as you might guess, assumes the earth is a sphere. Or the more accurate st_distance_spheroid. Using these, you don't need to reproject and you will probably be fine for your distance queries except in edge cases. Newer versions of PostGIS also let you use geography columns
tl;dr - use st_distance_spheroid for your distance queries, store your data in geography columns, or transform it to a local projection (when storing, or on the fly, depending on your needs).