More actions
Main article: Ratings system
The subjective and community-based ratings system
Let's begin by stressing that first we are building a subjective ratings system in the context of a peer-to-peer network. Everyone would have their own network of contacts, choose their own categories for ratings, algorithms for aggregation, weights for aggregation equations, etc. Individuals would identify themselves to their peers of choice with a public key or similar methodology. Their opinions would be encrypted and signed with their private key and decipherable by only their direct peers, those given the public key. With this, individuals would have a tool for tracking and judging the opinions of their immediate peers. They would also have the ability to access someone they are not directly connected to through someone who is. If A wanted an opinion from C but C is not in A’s direct peer network, then A would have to find someone, say B, who was known to both C and A.
Each shared public key would be different, by default, for everyone that received it. Tom can give his public key to Mike and Alice but they would each receive a different key. Tom would know that he gave one public key to Mike and another to Alice. Then when he communicates with Mike only Mike can decipher what he said. Alice wouldn’t be able to decipher it unless Mike gave her the key intended for him. Thus all communication is, by default, completely private. Tom can, of course, choose to use the same key for both Mike and Alice but such a choice is one step removed from the envisioned default. Such a step would be one way to begin moving toward a public node.
In a system like this, individuals remain as anonymous as they like, aside from any minimal identification needed for communication (eg public key). Those who already know each other can reveal their public keys and then communicate through the ratings system. The system could build, over time, mechanisms whereby people can communicate directly but the base protocol would not include this layer.
As an extension of this idea, we can envision a public version of the ratings system where people maintain an identity that anyone can get access to. If a writer wanted to publish for the whole group of users he could do so by identifying himself publicly through his IP address (or other addressable mechanism). The only change here is that the public individual has an address known to everyone. Otherwise, the system is based on the same private peer-to-peer network envisioned earlier.
A further extension of this idea involves a public rating system (or community-based ratings system). This is where the entire community agrees to use a standard ratings system with pre-agreed rules for privacy, aggregation, etc. One way this could be done is to agree to use the ratings system of a public person on the network because the community happens to have a high regard for that individual. The public ratings system is then simply a publicly addressable node on the network that everyone agrees is the community’s rating system. But individuals, barring any community rule against such, would be free to use their own private ratings system as well.
Both the private (subjective) and public (community-based) systems are likely to be useful. The community-based system will grow out from the subjective system depending on the needs of members. As communities become more cohesive and do more projects together, they will probably gravitate toward a common ratings system. This will be especially the case for developing a sophisticated economic system where production standards, for instance, are judged in common.
But it is unclear how this will unfold. Perhaps the subjective system will be able to handle a wider range of interaction and make a formal community-based system unnecessary. This will all no doubt depend heavily on the community and its goals.
Privacy and the evolution of a subjective ratings system
The essential property of the subjective ratings system is the ability to control one's level of anonymity, or privacy. Anonymity is a continuum ranging from the hermit who does not interact to the public persona who maintains a livestream of themselves 24/7. The first and most basic level of interactive capability would seem to be public key sharing with a known person, as described above. Now, at least, statements made can be attributed to some unique identity. Even here we can imagine members making trade-offs between privacy and any benefits they may obtain by revealing some aspect of themselves. It is no accident that the most influential people tend to be publicly facing. Doing so, however, results in maintaining an addressable public identity and would form the public extension of our private network. We can then progress to a community-based system where any number of rules might govern identity. This would again depend on a tradeoff between privacy and benefits.
But let’s for a moment imagine a system that never went beyond the subjective ratings system, at least in any formally stipulated sense. How might such a system evolve? If folks wanted to gain more influence by reaching more people directly, they would reach out to their immediate network and ask them to forward information about them. If this public person succeeded in becoming known to many people he would connect directly with them and effectively be a public address that everyone knew about. This person would now have the power to lead his “members” in community building. He would have direct access to everyone, so everyone would have direct access to each other, through him. He could now, for instance, ask members to reveal their public key (or other minimal identifier) through him so they could all get to know each other directly. We have thus taken our first step into a central ID scheme which could be supplemented by more intrusive forays into private information gathering (eg picture ID). The leader could ask members to work together on projects of economic interest and establish a system of tracking production and reward. A centralized system, presumably, since it would be important to associate who produced what. Inevitably though, this is another step in privacy reduction.
It is easy to imagine a system evolving from a subjective base in this way and perhaps never formally compromising its subjective roots. Here, it is only individual choice that turns it, effectively, into something else. However, it is difficult to imagine that without the choice to become a community, the subjective system can accomplish significant work or efficiently advance the truth. Again, individual and community goals will dictate the results.
CRS vs. SRS for Information U99
We also discussed differences between the SRS and CRS. Dan mentioned that the network architecture of the CRS was likely to be a hub and spoke system, whereas the SRS is a tree. Brainstorming_21 did not mention network architecture but did assume that the architecture of the CRS would be similar to the SRS. When used as a voting system, where presumably everyone in the community votes, the CRS is certainly a hub and spoke system. This fit Dan’s conception of the CRS as primarily a voting system. I assumed that the CRS was both a voting and information system (more on this below). As an information system, it would have a tree-like structure similar to the SRS. The community might ask a panel of experts who would be the primary contacts but those experts would then have their own sources that informed them. But what would this network of sources be? Probably the SRS. If so, we could view the CRS as a hub and spoke system which then implicitly relies on each individual’s SRS.
More fundamentally, we discussed whether the CRS should function as an information service at all. I thought it should but Dan did not. We agreed to discuss it offline and the result seems to be that my characterization of the CRS, as an information system, was poorly explained and misunderstood. So here I’ll try to supply more background, detail, and clarity.
First, the CRS information feature is not a hard requirement about how the system should be designed so much as a prediction about what it will become. As we’ve discussed, the SRS leads to the notion of communities of like-minded people. For the purpose of decision making, they will probably modify the SRS to handle group voting. We anticipate a direct democracy where everyone in the community can vote on matters of public policy. We will presumably already have public nodes on the SRS who choose to be known to everyone and can be linked to by anyone. So the voting system can be viewed as another public node that handles collective decision making.
Needless to say, its architecture will be a hub and spoke system (as noted above) since everyone in the community will link directly to it. To understand the issues they are voting on, the community can certainly continue to use their own SRS as they would have already been doing. But my prediction is that the experts they trust will form organizations and these will also be (or become) public nodes. Difficult decisions by the community will be informed by these public organizations or experts. Eventually these organizations would take on a quasi-official role as the community-trusted source of information on whatever subject is being decided.
Let’s pause here and understand that the community, in this scenario, is the government. As such, its decision-making infrastructure is the CRS voting system. But the people themselves are making the rules, setting the policies, “funding” infrastructure projects, raising armies, etc. The community does not have a legislature to do this for them because it is a direct democracy. It’s “executive”, such as it is, would simply be administrators to carry out the policies decided on by the community. In every way, the governing structure is subject to the CRS and the people of the community.
Given this, it is hard to imagine the CRS having these capabilities without also having an information system to accompany it. Every system of governance needs a system of fact-finding, research, and intelligence. Our congress has its own research service as well as the executive agencies who provide it with information regularly. To expect a community, in its capacity as policymaker, to use only an informal network of contacts for this purpose seems unrealistic. It would be like expecting the president to use his friends and maybe the news channel he watches, rather than the CIA, for information about a foreign adversary.
In any case, keep in mind that the community will need an administrative system with the appropriate staff to implement the policies that are decided. These can be thought of as our current government agencies, most of which are under the executive branch. Such agencies will naturally have information gathering and intelligence services, just as our government does now. They won’t be the only source of information but they will be an important part of it. It seems natural that agencies with the informational ability to carry out policy will also have information pertinent to making policy decisions in the first place. The community members, for their part, will have these organizations available to them as part of their SRS.
Thus the CRS as an information system is not a monolith. It is composed of many trusted organizations, one presumably for each subject area. In practice it is an extension of the SRS and used by individuals who choose to include it as part of their own network. The process by which the CRS information system occurs is natural, an expression of community will, and is not forced on any single person. That the CRS exists at all is a result of trust decided on by community members.
Another thing CRS organizations will provide is research and commissioned studies. Communities will need to investigate emerging but little-known problems (eg pandemics), do scientific research, obtain intelligence about adversaries, solve crimes, etc. Government’s commission studies all the time and, indeed, have entire sub-organizations devoted to doing research on their behalf. Our communities will be no different in this respect.
We can think of the CRS information system in current terms as the websites of federal agencies, (eg https://www.epa.gov/, https://www.cdc.gov/, https://www.state.gov/) Or the code of federal regulations. These are certainly valid sources of information on public policy but they are not what most Americans use. We generally rely on the press to distill important policy details for us since they have reporters who cover the government continuously and write in a more accessible style for average people. This function is obviously one we can anticipate as well. There will certainly be public nodes of the SRS that function in a reporting capacity and these will likely be part of each member’s SRS.
Furthermore, we might conceive of a “government-owned” news organization like the BBC in Britain. Again, this would function like any other media organization but would “belong” to the community in the sense that the community specifically voted on its creation and funding. In a direct democracy, such an action would probably only be taken in response to some defect in the “naturally occurring” news organizations.
The community might also want to use the voting system to rate the CRS in a public fashion. The CRS, after all, works for the community. It would want feedback from the community about how well it was doing and the community would want to hold it accountable. Individuals will do this already for the SRS, if they’ve made it part of their network, but this information may or may not be public. A public rating of CRS organizations seems like something a direct democracy would naturally want.
Given this step, it is possible that the community will use this ratings system to establish public ratings for individuals on non-public individuals. This is a collective choice that obviously borders on privacy concerns. As we’ve discussed previously, perhaps this would be limited to individuals that have broken the law or sunk very low on trust and require more vigilant monitoring.
Preventing Misinformation Bubbles in the CRS and SRS U99
One of our main goals for the ratings system is to prevent misinformation. Both the CRS and SRS are subject to it. The SRS easily permits misinformation bubbles to form because if someone doesn’t know a controversial subject they can easily be lured into the “wrong” opinion. This is analogous to what happens on social media. The ratings system should check this somewhat but if a subjective network is motivated by bias, it is hard to control. Fortunately, the SRS tends to place some natural limitations on the damage that can be done. The CRS, meanwhile, is more difficult to subvert in this way but, if infected, would cause far more damage due to its public-facing nature.
We might speculate that the CRS and SRS would check each the other in order to prevent misinformation. But we already have checks like this and they don’t seem to work. If the CDC can’t convince a significant portion of the population to take a free vaccine against a deadly pandemic, then we know that subjective networks are winning the day. Indeed, the subjective networks succeed by dismissing the official networks as fake.
We’ve discussed before how the ratings system itself might be enough of a barrier to control misinformation. If a source of misinformation exists on someone’s subjective network, then it and its followers can be downrated. But if the network exists within a self-reinforcing bubble, we might have trouble penetrating it. One somewhat intrusive method might be for the community to expose misinformation nodes by “outing” them publicly and subjecting them to the CRS. Then the entire community would rate them as a public matter. The misinformation node doesn’t necessarily have to stop, but everyone would know where it stood in terms of trust.
This may or may not be enough and brings up questions of privacy (can a private node be forcibly made public?). In the end, everyone has a right to misinformation, to a point. To echo Rawls, that point is when it threatens the basic liberties of everyone.