In June, the SEC gave some of its most concrete guidance to date that cryptoassets can start out as centralized projects, possibly initially sold under securities laws, and eventually become "decentralized" and thus no longer sponsor-controlled, and no longer sold or transferred under securities laws. It makes sense that a decentralized protocol does not fit the definition of a security.  There's not a clear single issue or promoter (for purposes of reporting, etc); tokens are often generated on an ongoing basis (which would constitute a "continuous offering" and related registration requirements); tokens are generated on a fully peer-to-peer basis in the protocol (potentially implicating independent nodes as transfer agents or broker dealers, or requiring those as middlemen); not to mention how all of the above are complicated by new issues like forking. Of course, all of this leaves some open questions on what exactly constitutes decentralization, but I am confident we will work through those to come to a usable definition.  For example, "is the network forkable" is one simple (but incomplete) heuristic.  Another is: "would the network continue operating if the initial sponsor went out of business". This second one is perhaps the most concrete, and I believe the the net effect of the SEC guidance is that we will begin to see protocol development companies (the initial "sponsors" of cryptonetworks) set a course to intentionally self-destruct. How this is done, exactly, will remain to be seen. Already we have seen a company / foundation split as one way of setting the protocol in the hands of a long-term custodian that is not the initial sponsor.  Some projects may bake self-destruction into their initial charter (and any associated coin offerings or distributions); others may make self-destruction part of the protocol software development roadmap. But regardless of mechanism, I believe many projects will begin to contemplate their "path to decentralization" -- that is the takeaway from the SEC guidance, and a self-destructing company (leaving behind only an autonomous, decentralized protocol) is the logical result. This is going to be a messy process. For example, the recent launch of the EOS network demonstrated some of the challenges of handing off a protocol and network to the community.  But luckily we will get to see many more real-world experiments as projects move out of the fundraising phase and into the build->ship->decentralize phase.