The specification ofno-store can be read to allow the response to be reused for another concurrent request, since it's specified in terms of non-volatile memory. Likewise, other cache directives are silent on such reuse.

So-called "request collapsing" has become quite common in proxies, reverse proxies and CDNs. Clarification as to whether such reuse for temporally concurrent requests would be helpful (in the case ofno-store and otherwise).