I just finished listening to The Santiago Boys podcast about the attempts to use cybernetics to run the Chilean economy under Allende in the early 70s. It was very interesting and got me wondering what kinds of technology (hardware or software) that you have come across that you think would be good for managing resources/efforts and democratising society.
What tools do you think could empower people?
The tools are already readily available:
Relational Databases
SAT solvers
The missing bit is social action, which no amount of software can solve.
Yeah, it’s not about automating the decision making, but making the tools available for when the will is present.
The tools are already readily available under FSF approved licenses. https://www.gnu.org/licenses/licenses.en.html
Support the FSF if that is a legitimate concern to you
Well, I think tools that allow spontaneous organization are fantastic. The ability to encode an organizational structure would be great. You can formalize institutions and relationships and describe them completely, enable enforcement of any formal agreements automatically, I would love to see that.
There’s a problem with using technology to distribute resources though. It’s mostly a network topology problem, but also an issue of processing power and lossy information channels. Think of every person as a node on a network doing computation constantly about the optimal distribution of resources. This is a self organizing distributed system, one in which no node knows more than it needs to about it’s needs and resource availability, and makes no computation about the needs of other nodes. This is a system where the outcome, optimal distribution of resources, is an emergent property. An artificial system with a more centralized structure cannot be as efficient because 1) the route between nodes is inefficient because it must traverse the central node, 2) the processing power the central node needs to decide everything is at a minimum as high as all other nodes combined, and 3) the information available to it about needs and availability is noisy and missing important pieces; it is lossy. As a result of these problems, a system in which resource distribution is centrally controlled cannot outcompete a system in which the computation is distributed; essentially, described from a cybernetics viewpoint, a centrally planned economy cannot be more efficient than a market economy. So, designing systems to run an economy is infeasible, this set of problems cannot be overcome with any technology, the only way to do it would be to take humans out of the equation, and humans are the whole point of distributing resources in the first place.
Yeah, I was thinking tools for a more local level that enable collaboration, knowledge sharing, and tracking (rather than allocating) resource use. For higher-level coordination, something that could flag emerging conflicts over resources would be useful to spot problems on the horizon and then enable groups to work things out rather than impose a solution.
I worked on a project earlier this year to build an AI system that would share knowledge about performance and best practices to match communities up with solutions to their problems, as well as modules to help them implement, manage, and track them, all of which could then be fed back into the system. Enterprise tools for grassroots groups, so to speak. Unfortunately, it didn’t get off the ground.
More a comment than an answer, but an old coworker and I used to take walks and discuss how ML could be applied to managing an economy. We called ourselves the Open Source Socialists.
Machine Learning has huge potential. I am more interested in seeing it used to inform decisions and highlight things that may get overlooked than deciding actual allocation.