What does the social biology of ants have to do with online reputation? What kinds of organizations will dominate the future? Heather Moore caught up with Matan Field, CEO of Backfeed, to discuss.
Backfeed is one of a number of new technology groups building a platform around “decentralized reputation,” a concept that has been developed for the blockchain. Decentralized reputation might sound innocuous, but it has real implications for how we navigate online data, especially social ranking data that is used by consumers on sites like Yelp or Reddit or even Amazon. All of the current reputation systems are centralized, and the way in which we explore the data is under the central authority’s control. But the idea of a decentralized reputation system makes other possibilities, well … possible. Matan discusses his vision for how these types of tools will enable new forms social organization on a massive scale, potentially replacing the corporate organization model that has been so dominant for the past 100 years or so.
Tell us about yourself and Backfeed.
My name is Matan Field, I am a theoretical physicist. I studied mathematics and physics. I hold a masters and PhD in theoretical physics and string theory. Backfeed is a distributed governance system for blockchain-based applications allowing for the collaborative creation and distribution of value in spontaneously emerging networks of peers.
How does a theoretical physicist become the leader of all of this?
After my studies concluded, I continued with physics for a couple of years. During all of those years, I was always interested in alternative economies, alternative organizational structures, cooperatives, etc. So, I played around with those models. Somewhere during the past decade I kind of played around with an idea to make a real-time, ride-sharing application. I started working on it with some friends and through the progress of that project, we saw an enormous problem in raising the critical mass of users. Some systems, the most powerful ones, don’t work without a large number of users. The solution we came up with was to look at our users as members of a single organization. For example, 100,000 people, employees, partners, all within a single organization. That would be your critical mass. It raised the question of: how would such an organization behave? How would hundreds of thousands of people agree about stakes and share the value created, so on? That was the same time where these questions popped up around the world with the launch of the Ethereum blockchain.
When was this?
We started October, 2013. Ethereum published their whitepaper in December, 2013. Everything kind of came together. We started getting into this space – decentralization/blockchain. The more I went along, the more I realized there are theoretical and practical layers which are still missing from the program, in the larger aspect. I decided to quit at the ride sharing company and to found Backfeed to build or fill that gap. That gap of systematic and collaborative decision making, allocation of resources, governance, and voting distribution and, in particular, reputation system. We had a vision for how these new tools would look (and it’s very visual). It’s all so very different from what people are used to, because every known notion of an efficient system heavily relies on the centralized nature of the project it sits on. Decentralized and distributed organizations have not beenas efficient or powerful as centralized, top-down organizations.
The first challenge was to design a new kind of protocol which would give reputation on a decentralized network, which leads to a completely new notion of reputation and decision making.
How so? Is decentralized reputation similar to how a Reddit community votes things up or down finding a fluid reputation score?
So that’s the main (social) ranking system that exists today. You have tons of ranking systems from Amazon to Reddit or Ebay. Even Google’s PageRank search engine is a type of ranking system which determines which websites are more important than others. But all of them, from Reddit to Amazon, etc., if you try to take the algorithms as they are and implement them on a decentralized network, they will immediately fail. Right now, for example, Google’s PageRank algorithm relies on a centralized node — Google’s home page — that it trusts 100%. The question is: who makes the monitoring authority on a decentralized network? That’s why people thought it was impossible to come up with an algorithm to do this. The funny thing is that it’s actually easier.
One way to look at it is to say: there is no monitoring authority (which gets you stuck). Another way to look at it is to say: everyone is a monitoring authority (which makes it easier). Once you identify the right structure, it is less complicated.
Yelp has been in the press recently for people being able to completely decimate a company by writing devastating reviews of it, whether those reviews are warranted or not. Many people are moving away from online reputation, in general. Are there similar pitfalls in a decentralized reputation system?
Yelp is a great example to look at and imagine decentralized. And there are several ways to look at it. One example: if you are making very bad reviews on a restaurant while everyone else is making good ones, then your reputation can decrease. Secondly, if you don’t agree with the judgement of the people around you, you can always ‘fork’ the network and have a reputation system of your own. In that way, you can organically, naturally, dynamically organize the network into a network of networks, each having their own set of values. The pitfall of Yelp’s bad user reviews is it’s centrality. It’s too rigid.
So a particular community could explicitly say what their values are and organize around that?
The beautiful thing is, they don’t need to explicitly state their values. You can rely on a mechanism called stigmergy.
It’s from biology, the term that describes how ants self-organize. There is no direct coordination. They don’t agree on anything. They just feedback; or, Backfeed. Out of that feedback, the whole takes shape.
Communities are organically divided into subgroups and so-on that are more in line with one another. Networks naturally obtain micro-alignment and macro-diversity. This works out organically without a coordinating authority.
You’re using some of your knowledge of physics and string theory in this?
Yeah, I guess. I haven’t yet found a direct relation to String Theory (still looking!). But physics tools are available and helpful. It’s on the verge of mathematics, physics, computer science, complex systems, social physics – between all of those branches.
What do you see as the main advantages to decentralized reputation?
There are several. Each of these is quite dramatic. For example, one is scalability. Centralized networks are simply not scalable. The problem is, the more you scale, the more overhead costs or transaction costs you have per-person. There is a growing tension between scalability and efficiency. Even then, the only way to pull the tide is to make the system more and more rigid, less and less freedom and creativity — and even that is limited. Whereas a decentralized network is fractal. It works the same for millions as for hundreds. It becomes more and more resilient with scale. In nature, decentralized networks are more resilient, they encourage freedom, creativity, flexibility. There is no one saying no. So scalability, resilience and freedom are the major advantages.
Online reputation involves a kind of social collaboration. Do you see Backfeed promoting this collaboration?
In the free market parlance, competition is incentivized. With this kind of mechanism here, it makes a new kind of framework where collaboration is the winning strategy. If I want to be more precise, it’s something we call coopetition. If you look at a diversity of opinions, nobody is forced to put anything. Not even encouraged to put anything, but in a way, different groups or agents will cooperate on the things they agree and compete on the things they disagree about. Everything is equally valid.
So it’s more fluid, adaptive way of working?
Absolutely. I believe it will lead to a completely new notion of organization and scale. What will it look like when millions of people organize and work together?
I’m willing to bet that in the next decade, we’ll find out that these kinds of decentralized networks will replace corporations. Once the infrastructure is well-established, it will simply become more efficient. No centralized network will be able to compete with them.
You were talking about resiliency earlier and there’s always a balance between being efficient and homogenous and on the other end being flexible, inclusive, adaptive. In the industrial age we moved so far towards efficiency that corporations have become unable to handle the current atmosphere.
Which means that being efficient in one dimension is not efficient in another.
And a new kind of organization will evolve past our current corporate model?
These kinds of old-school organizations tend to be more efficient in small time-scale, less efficient in long time-scales. More efficient in technical stuff, less efficient in creative. That’s why startups are so small. For big companies it’s hard to innovate and create, rather than maintain.
What kinds of services have been implemented, using this protocol? What do we have to look forward to?
We have built several proof of concepts and are collaborating in partnership with many groups to build tools together more efficiently. One proof of concept is built as a chrome extension of Slack, so that teams that already use Slack can start using reputation tokens. There is a proof of concept for a decentralized game. For a decentralized curation of a library of content for a search engine. Another group is doing decentralized journalism.
Do you see applications in something like Reddit?
Yes, for sure. We are working on something not quite like Reddit but maybe more like Medium. It involves decentralized curation. Reddit is about posting links and commenting on them. Once we establish posting links then we can establish commenting on them. So we’ve started with this Medium-like project.
How do you see this sort of system being used by large amounts of people. Right now it’s early adopters. Do you see a path towards mass adoption?
Yes, it’s like any technology. In the early days of the internet, hardly 10% of the population really understood what it was about and they were the pioneers. But after the pioneers, there were enough to make such a difference and they had such a big advantage that everyone else had to catch up. The same thing happens to any technology. Even Facebook. Twitter. Any new technology that changes the mindset.
I’m completely sure that in two or four years these technologies will be in the mainstream.
At the moment it seems very technical. There seems to be a huge opportunity for creating a user experience that is translatable to less technical people.
Exactly. We are working on that, reducing the level of difficulty and improving the user experience. The number of people moving into that field (UI/UX), improving these experiences, continues to grow. People are moving in.
We often hear that this infrastructure is similar to the internet, is that how you feel?
We really feel we are building a new infrastructure for a new layer of the internet that is more collaborative and more “trustless.” Meaning, you don’t need trust; you will have more distribution of value and information and responsibility …everything, which has much more chance to survive.