Offer Network Primer

Francis Heylighen

This text is an introduction and summary of the new offer networks paradigm, a very promising architecture for designing a global sharing economy. This architecture would use the capabilities of AI and the Internet to efficiently and democratically coordinate activities across the world. It would achieve this by specifying a universal protocol for the sharing and exchange of goods, services and information between all individuals, organizations and technological systems.

Such a protocol would interconnect and unify the myriad commercial and community platforms that have arisen over the past two decades. It would play a role similar to the TCP/IP protocol that unified computer networks into the encompassing Internet, and to the HTML/HTTP/URL protocol that unified sundry methods of information retrieval into the World-Wide Web.

An offer network is an ICT platform that for all connected agents (individuals, groups, software agents, machines, robots…) would at least contain:

  1. their offers, i.e. which goods, services or information they are ready to provide to others that may need them. E.g. John may be willing to provide eggs, to repair bicycles, and to answer questions about Egyptian history.

  2. their requests or demands, i.e. the goods, services or offers that they would like to get from some other agent(s). For example, John may need apples, to have his pine tree pruned, and to get an introduction to elementary geometry.

  3. the conditions (hard constraints) that need to be satisfied before the offer can be realized. These can be expressed as condition-action rules, e.g. John: apple → egg (to be read as “IF apple, THEN egg”), meaning that John is willing to offer an egg on the condition of getting an apple in return, or Marc: saw → prune tree, meaning that Marc is willing to prune a tree, but needs a saw to do this.

  4. a collection of intelligent algorithms and heuristics for matching offers and demands, taking into account the condition-action rules. These may include backward chaining (starting from a demand searching backwards for a sequence of conditions necessary to satisfy that demand until an offer is found that fulfills the conditions) and forward chaining (starting from an offer in order to find a demand that it directly or indirectly satisfies).

  5. general optimization criteria (soft constraints) that select the best or most efficient way to satisfy a demand or exploit an offer. For example, from several offers for apples, the system may select the ones that are available nearest to John’s place, and that are most similar in taste to the type that John prefers.

Offer networks were first proposed by Ben Goertzel as a system to exchange goods and services without the intermediate of money—thus precluding many of the problematic side effects of using money as the basic source of value 1. His idea was to build a more efficient form of a barter economy by greatly expanding the pool of offers and demands, and by using intelligent algorithms to find matches across this pool without need for one-to-one exchanges between individuals.

With input from several other people, this idea was extended by Francis Heylighen 2. He envisaged it as a strategy for transiting from our present consumerist society with its many problems of inequality and unsustainability to a more utopian vision of a world coordinated by a distributed intelligence (“the Global Brain”), where scarcity and friction would have been essentially abolished 3. Inspired by gift economies rather than barter economies, Heylighen added two features to Goertzel’s design:

  1. the mobilization of unconditional offers, which take the form of gifts, voluntary work, or open-access information. For example, if Ann’s garden has produced more apples than she can eat, she may be willing to offer them to John without demanding anything in return. She may similarly lend out her saw to Marc, so that he can unconditionally prune John’s tree.

  2. a reputation system for agents that would reward them for unconditional offers and for the general quality of their contributions, and that would penalize them for consuming more than their fair share of unconditional offers and for others forms of abuse or poor quality work. Thus, Ann and Marc will gain reputation because of their altruist deeds, while John may lose reputation if sooner or later he does not compensate their good deeds by unconditionally helping them or other people in the network.

Heylighen argued that with such features an offer network would be much more flexible and efficient. Moreover, it may be able to not only complement but gradually replace our present market economy, by making the accumulation and exchange of money increasingly less central to how society functions. Such a slow “eclipse of capitalism” was also envisaged by the economist Jeremy Rifkin 4 as a side effect of the emerging collaborative commons, an idea similar to a global offer network but less concretely formulated. The advantages of such a socio-technical organization include much greater efficiency, less waste, more recycling, reduction of poverty and inequality, greater resilience and sustainability, and a renewed focus on non-monetary values such as community, cooperation, well-being and wisdom.

The challenge now is to formulate the offer networks model in a sufficiently precise and concrete way so that it could be implemented—as least as a proof-of-concept, preferably as a testing ground for experiments with actual communities of people using such a platform to share a variety of goods and services.

Implementation will require first an ontology that formally specifies the basic components of an offer network, such as agent, offer, demand, condition, optimization criteria, and the links between them. Independently of the offer network proposal, Florian Kleedorfer has developed a similar framework called the “web of needs” (where needs are equivalent to requests or demands). The basic components here are called “user proxies”, which are virtual equivalents of a user’s offers and demands 5.

The next step is to develop a set of algorithms for efficiently searching this abstract representation for good matches. Goertzel with his collaborators have started to experiment with constraint satisfaction methods, using standard AI algorithms. Kleedorfer proposed to implement a variety of software agents, using different heuristics, that would to some degree compete in finding the best matches. At this stage, optimal matching is probably not yet a priority, as there exist a great number of techniques that have proven their usefulness in these kinds of problems, and as their performance will only increase with further progress in hardware and software.

The most important issue at the moment seems to be to clarify the conceptual framework to a sufficient degree so that the offer net community can either build a proof-of-concept themselves, or inspire others to apply these ideas in their own platforms. A universal sharing platform such as offer networks is an idea that is very much “in the air”, in the sense that many people are independently developing similar ideas, including the already mentioned “web of needs”, “collaborative commons”, “pervasive social computing” 6 and “social web of things” 7. Moreover, there is a tremendous and growing variety of sharing platforms limited to specific domains, such as carsharing or couchsurfing (Last, 2015).

Yet, from the various reactions we received until now, it seems that the offer network paradigm is conceptually more elegant, broader in its implications, and more concrete in its applications than these related ideas. Therefore, it seems like the right starting point for developing a socio-technical infrastructure with the potential to solve many of the most pressing problems of present-day society.

We hope that this document and the papers it refers to will convince other researchers to join this project, further elaborating the as yet unclear issues, and experimenting with implementations. Together, we may thus generate the critical mass for initiating a fundamental transition towards a better world—the way Tim Berners-Lee and his collaborators changed the world by designing the World-Wide Web protocol.


  1. Goertzel, B. (2015). Beyond Money: Offer networks, a potential infrastructure for a post-money economy. In B. Goertzel & T. Goertzel (Eds.), The End of the Beginning: Life, Society and Economy on the Brink of the Singularity (pp. 522–549). Humanity+ Press. http://globalbraininstitute.github.io/onet/files/offer_networks.pdf

  2. Kleedorfer, F., Busch, C. M., Pichler, C., & Huemer, C. (2014). The Case for the Web of Needs. 2014 IEEE 16th Conference on Business Informatics (CBI) (Vol. 1, pp. 94–101). Presented at the 2014 IEEE 16th Conference on Business Informatics (CBI). http://publik.tuwien.ac.at/files/PubDat_237340.pdf

  3. Heylighen, F. (2015). Return to Eden? Promises and Perils on the Road to a Global Superintelligence. In B. Goertzel & T. Goertzel (Eds.), The End of the Beginning: Life, Society and Economy on the Brink of the Singularity. Humanity+ Press. http://pespmc1.vub.ac.be/Papers/BrinkofSingularity.pdf

  4. Rifkin, J. (2014). The Zero Marginal Cost Society: The Internet of Things, the Collaborative Commons, and the Eclipse of Capitalism. Palgrave Macmillan. http://www.bookdepository.com/Zero-Marginal-Cost-Society-Jeremy-Rifkin/9781137278463

  5. Kleedorfer, F., Busch, C. M., Pichler, C., & Huemer, C. (2014). The Case for the Web of Needs. 2014 IEEE 16th Conference on Business Informatics (CBI) (Vol. 1, pp. 94–101). Presented at the 2014 IEEE 16th Conference on Business Informatics (CBI). http://publik.tuwien.ac.at/files/PubDat_237340.pdf

  6. De Florio, V. (2015). Fractal social organization as a foundation to pervasive social computing services. arXiv:1601.01222 [cs]. http://arxiv.org/abs/1601.01222

  7. Ciortea, A., Boissier, O., Zimmermann, A., & Florea, A. M. (2013). Reconsidering the social web of things: position paper. Proceedings of the 2013 ACM conference on Pervasive and ubiquitous computing adjunct publication (pp. 1535–1544). http://dx.doi.org/10.1145/2494091.2497587