The AI that seeks to optimize blockchain configurations, or simply navigate the environment, is aware of the off-chain "social fabric" because it fills the role of communication: you can improve by introspection if you communicate.
To manage complexity several approaches are possible, including using Lawyer's expected force  (a centrality measure) \cite{Lawyer_2015}  to greatly simplify the problem. When node power is low, influence is a function of neighbor degree. As power increases, a node's own degree becomes more important. The strength of this relationship is modulated by network structure, so it is expected that it will be more pronounced in narrow and dense networks such as social networking (e.g. Twitter channels of bitcoin whales). The network effect, however, has two levels: one is the on-chain\off- chain interplay (symbolic regression model of active accounts/addresses driven by social network activity), and the other is the off- chain\off- chain interplay (for instance, of the dynamics between Twitter channels and Telegram groups).

Vector fields as temporal streams

An activity must be decoded sequentially over time. The intelligent agent may do this by using a combination of genetic programming techniques (after all, somewhat static DNA and its transcription pattern over time creates biologically essential temporal patterns) and signal processing (e.g. Kalman filter, for short-term streams), but a complementary approach for fast evolving systems that are always in flux is to use actual fields. In the case of economic systems such as blockchains, standard signal processing analysis techniques and information theoretical measurements help visualize the historic correlations, but a sound investment strategy should also consider the correlation migration –how the correlation changes (or not) over time. While it is possible to plot a correlation graph for each point in time, we find that using vector fields \cite{_i_insk__2016} allows for mappings with a higher information density, especially for portfolios of a large size –for instance, consider that when tokenized Dapps are also viewed as assets, we are facing prohibitory large portfolio sizes. To implement the method we begin by defining the convention for the vector components. From the possible traffic sources for a new project, we found that referrals and social networks are the more prevalent, especially in the early stages of a proposal listing when word of mouth in social networks such as Reddit and the ability of the founding team to generate buzz in media and news sites plays a role. The resulting vector field gives rise to a flow. A fluid flow provides an effective way to summarize the dynamics of a portfolio to include an arbitrarily large number of entities, rather than simply scaling up the number and size of correlation graphs (not to mention that for communication purposes, a vector field is also a more intuitive representation of cashflows equivalents). In one hand, the (total) vector magnitudes are a measure of strength, in the other, the interaction between the different assets (as revealed by singularities in the flow) present a portrait of the system (the portfolio attention correlations). Figure 9 shows a vector field for a 32-asset mapping in Ethereum during March, April, and May 2016, rendered using 4 techniques to highlight different aspects of the flow.