4 ms·

Politics are interesting. Perhaps politicking wouldn't be necessary if we could determine with accuracy the results of policies before implementing them. For ex

Politics are interesting. Perhaps politicking wouldn't be necessary if we could determine with accuracy the results of policies before implementing them. For example, maybe this could be done by complex computer simulation in the future. Furthermore, policy might be synthesized by the same processes if it can be simulated. Which then raises another interesting point: would democracy be less efficient and necessary if we could determine optimal policy for society before implementing it? Would we just vote on the desires of society rather than policies to get us there?

tvanantwerp 9y ago
I work at a think tank. I don't think any of the models used by any policy experts are at a level where such simulation is possible. We can't predict the weather more than about a week out, and that's just a physical system--policies depend on quirky humans with lots of weird incentives and preferences that are difficult to predict. Modeling can let you make educated guesses at best. Of course, plenty of partisans have an interest in producing models that spit out results that just so happen to show what they already believed (e.g., Tobacco Institute), and politicians will cite those results with the same regularity as more rigorous work. Without expertise in the matter, it can be hard to find any given model's weak points; and they all have weak points. As for voting for desired outcomes rather than desired policies: I'm pretty sure this is what most people are doing already. Lots of political pressure all around have your cake and eat it too. Sadly, I've never seen any policy that could solve a problem like that--too many trade-offs in real life.
tonyedgecombe 9y ago
A previous government in the UK spoke a lot about evidence based policy, in reality it was evidence based policy that also happened to match their ideological outlook. Perhaps our future AI overlords will have a different approach.

tvanantwerp 9y ago

I work at a think tank. I don't think any of the models used by any policy experts are at a level where such simulation is possible. We can't predict the weather more than about a week out, and that's just a physical system--policies depend on quirky humans with lots of weird incentives and preferences that are difficult to predict. Modeling can let you make educated guesses at best. Of course, plenty of partisans have an interest in producing models that spit out results that just so happen to show what they already believed (e.g., Tobacco Institute), and politicians will cite those results with the same regularity as more rigorous work. Without expertise in the matter, it can be hard to find any given model's weak points; and they all have weak points. As for voting for desired outcomes rather than desired policies: I'm pretty sure this is what most people are doing already. Lots of political pressure all around have your cake and eat it too. Sadly, I've never seen any policy that could solve a problem like that--too many trade-offs in real life.

tonyedgecombe 9y ago

A previous government in the UK spoke a lot about evidence based policy, in reality it was evidence based policy that also happened to match their ideological outlook. Perhaps our future AI overlords will have a different approach.