The old implementation would actually retrieve the entire HTML page
where the README is located. We want to simply make the same request
with a different ACCEPT header so we can retrieve the portion of the
README html that we'd actually want to display.
