Now that I have basically digested the Synereo white paper (not all the math formalisms are understood in minute detail which will require some more time for study, but I get the overall concept), I want to write down some of my thoughts about Synereo's design and in the context of how I am thinking about priorities for any major paradigmatic shift in social networks.
1. Per the post upthread where I elaborated, I do not think the attention model of a social network can be tied to Synereo's crypto currency AMPs, nor should it, nor will that be compatible with allowing each user to choose their own attention model. There may be a use for crypto currency in social networking but it is not to hard code an attention model for what is supposed to be an individual user choice paradigm for a social network. I suppose instead Synereo could offer their attention model as one of the variants users can choose, but I think it will be a failure for the reasons I stated upthread. I think we will need free market competition between attention models in order to find out what works well and what doesn't. Also I think this means we can't assume nodes obey some global process calculus.
2. Users will not run a social network hosted from their own decentralized home computers ephemerally connected by asymmetric bandwidth ISPs
as is the initial planned focus on Synereo. Fugetaboutit. They will signup quickly (and/or
download the mobile app) just like Facebook or any other mainstream social network. Thus they will have their data and nodes hosted and thus just fugetaboutit this nonsense about
censorship resistance in the white paper, except perhaps in terms of encrypted sharing (but later the
government will demand a global decryption key as they have for streaming encrypted voice and video communications in the USA, Uk and rest of the 5 Eyes countries). See my prior post upthread for where I think this political-economic battle over private rights is headed.
3. The advantages that matter most to users which can only be gained from a decentralized structure (i.e. not the way existing social networks are structured) are the ability to control their own data and to choose their GUIs, apps, configuations, and even attention models. In other words, each user should be able to independently choose their own social network design, store the user's data independently of other users' data, and these user nodes should interact to exchange data, which each user node filters and stores locally the relevant bits to that user. Ideally, the entire platform should be programmable on top of some maximally generalized decentralized protocol layer.
4. The details about how to implement a maximally generalized decentralized protocol layer and define the potential motivation (economics) for all participants (users, service providers, cloud storage providers, developers) such that there isn't chaos and lack of focus so as to insure the end product converges to easy-to-use and compelling for all participants, seems to be at least on first contemplation mindbogglingly complex to distill, because there are competing/conflicting priorities in such a design. For example, a streaming music download provider must invest development resources to develop all the stop/start/pause/forward/rewind interactive functionality between the server and the client player, but then needs to somehow monetize both the bandwidth costs and the development investment, as well as attain an ROI that compensates for opportunity cost (not just a return of investment). Typically the streaming provider (e.g. SoundCloud, Spotify, etc.) would want to create a walled garden around the musicians who are uploading, the app developers who are calling their API, and the fans who are listening (e.g. through the provider's player and even embedded into provider's mobile app), so that it can extract some revenues on this ecosystem such as through a combination of advertising and/or music sales. One might propose microtransaction payments to the service provider for each song played, but this requires standardized APIs for all integration with the rest of the social network so that service providers become fungible (i.e. substitutable with each other) and in which case the service providers are reduced to competing on costs and thus can't attain a ROI to compensate opportunity costs for investing in innovation. Even if we say that users can use different service providers with their proprietary players (all fungibility embeddable in the social network Timeline), this doesn't resolve the issue of needing a standardized API for these providers to communicate fungible attributes (between all service providers) that users need in order to make an attention model more fine grained, e.g. for the music the genre and sub-genre of each track played/shared.
Thus it seems that any decentralized protocol will need to be too general and that interoption standards will emerge initially adhoc and then probably go through standardization via organizations such as the W3C.org. Similar to the browser wars of yore, we would repeat another wild west of adhoc experimentation and competing corporate offerings with the community settling eventually on the defacto standards that work best for all.
In other words, we would be creating a new decentralized protocol for the internet which would foster competition and participation instead of the walled gardens of Facebook, Twitter, Apple, etc...
Sounds exciting! Where do I sign up to create this?
Note
I was listed as a contributor on the W3C standard for CSS2.1. I contributed to the either the multi-column or table specification (I contributed on both but I was only recognized on one afaik), given my past history at the dawn of deskstop publishing in the 1980s.
i can't even find how to store Synereo in my hard drive. 
