It has been nearly two decades that the networking research community is striving to solve internetworking at global scale with a new clean slate architecture that would better fit contemporary Internet business requirements, and workloads, largely dominated by video streaming traffic flows; rather than incrementally improving it, with sophisticating patches. In fact, the current Internet evolvable architecture over-relies on the CDN global infrastructure of hundreds of thousands of servers. ICN (Information-centric networking) is a promising breakthrough alternative to contemporary CDN-backed large-scale IP internetworking. It promises built-in effective and efficient capabilities to handle content distribution traffic at Internet scale, backed by enhanced routers packaged with additional well-provisioned memory and behaving natively as content caches. To its full potential, the almighty lord enabling, it will support a smart GPT-like enhanced service operating over an Internet-scale distributed cache of arbitrary topology, able to locate strategically, copies of information binaries, matching information consumers interest predicates. This paper reports key conclusions from the ICN reported research haystack, to consider in future steps towards the next Internet architecture and protocol stack, integrating in-networking caching along with forwarding based on binary content names instead of contemporary address-based IP forwarding. We conclude by sketching directions not to overlook in future research on ICN, from our african glasses.