When we load non-extracted files from the cache, there is a chance that our token is stale and we will 401 (unauthorized) on them.
We will check for 401 when we do something like an ls, but for direct links we let the browser handle it and it doesn't know how to handle out 401s.