The overarching theme of my seemingly divergent interests is this (so far):
I would like to contribute building tools (specifically, decentralized computational tools) and the appropriate culture to manage them, so that they can help us:
- Rapidly come up with solutions to complex design/optimization problems
- Gather wisdom about ourselves and society, based on which we can deliberate and decide collectively at a large scale
Those computational tools would lift a large weight from our shoulders (perform intensive calculations rapidly and transparently, the thing which humans are not good at) and would allow us to concentrate on the tasks all humans can (and should) do:
- Communicating our preferences to secure decentralized information systems as well as to other humans, with total control and privacy. This is roughly what we now call voting.
- Accessing and understading the implications of the solutions/designs that those computational systems have formulated based on the multitude of our (often) mutually conflicting preferences.
- Based on the previous steps, deliberate, negotiate, concede, accommodate, harmonize our mutually conflicting preferences. Also, do the same on the models themselves (those on which the AI systems are built).
- After deliberation, submit once again our preferences to the system (voting).
- Repeat this process frequently, until a satisfactory solution and a form of consensus are reached.
We struggle to live together, but we cannot live without each other. Conflict is inseparable from the human condition. The dynamics of power and preferences is what politics deal with. However our current political systems oversimplify and corrupt Step 1 (voting), perform Step 3 (deliberation and consensus) with only a handful of representatives tied in a network of perverse incentives, and repeat the cycle every 4 years holding the citizens hostage in the meantime. Most importantly they don’t even implement Step 2: the political system and most citizens are oblivious to facts, evidence and scientific models; they end up shooting themselves in the foot either directly (through their personal behaviour) or indirectly (by electing their leaders).
We can do better. Technology and an appropriate surrounding culture can help us to make this process better, more fluid, transparent, frequent and science-based.
For this we need a multidisciplinary effort without precedents. Also we need robust mechanisms to determine human identity, to communicate and confront mental models tied to evidence, and to securely submit our preferences.
This may sound impossible, but at least it’s technically feasible. It may sound like a utopia, but it’s not. I’m sure it has lots of flaws. The path to progress is never ending. I just think it would be much better than the way we operate society today.
As I see the world right now those are the features of a social-computational infrastructure that could limit disinformation, prevent the concentration of power, preserve life and consciousness in our beautiful planet and elevate the human condition. I believe this is the way forward.