3 ms·
These types of things always make me think about solutions, as I'm sure many of you do, but I'm not sure I know an adequate one. Common to suggest to just bar
by godelski 3y ago
These types of things always make me think about solutions, as I'm sure many of you do, but I'm not sure I know an adequate one.
Common to suggest to just bar trading, which would make political jobs a major disadvantage. So others suggest only allowing for a third party managed fund that would be index focused. This sounds good at surface value, but when you go look at indices like VOO (S&P 500 based) you find, unsurprisingly, that Apple, Microsoft, Amazon, Nvidia, Google, Meta, Tesla, B&H, JPMC dominate. Any index fund will tend to be like this, unless we created some very special fund to track the whole economy (I think there could be something to this one fwiw, but construction would need to be done with extreme care). The reason this can be problematic is that they could still grow their wealth by targeting policies at the top companies. All you'd need is to know what the index composition is. I mean should be obvious since Nvidia is in VOO.
The other common suggestions are salary based.
Some say to move it up to disincentivize them. Can make sense, as most people do not care much when their earnings reach a certain point. I'm sure that includes many here, especially given that the a large part of the idea behind Google/Big Tech perks is because increasing salary is less attractive for people making $150k+ cash (prior to equity). But politics naturally attracts people who are power seeking, and well... we still see Elon giving a shit about how wealthy he is when reality is there's very little his net worth (~$200bn, current #2 richest) gets him that Peter Thiel can't also get (~$6bn, rank #451) (nuance needed about liquidity and other aspects as money doesn't work like it does for normal people in this wealth range, but claim is generally true). I mean Elon could create multiple CERNs that could run indefinitely through returns on investments. So there's good reason to believe that the wealth incentive model that works on the average person doesn't work well on politicians.
So there's also the argument to decrease salaries, do disincentivize power seekers from looking to politics as a way to obtain power and wealth. Idea here is to only attract people who are passionate about the job, like teachers (before teachers get burnout). But the nature of the job will always include power and Goodhart's Law is always in effect, so it'll be unsurprising that people can exploit the rulesets (or in this case, they make their own) and can channel money (to combine with their power) through third parties, which sets the stage for market capture.
So others suggest term limits. There's also pros and cons to each decision here. Some people also suggest changing the number of politicians: fewer to centralize power, which can streamline many processes but that makes us vulnerable to autocratic takeover as takeover is dependent on the collection of keys to power; others say to increase the number, in an effort to further decentralize the system and make it difficult to capture enough keys, but this can lead to very slow moving and ineffective governments as well as reduce dependence on experts as the point of representative powers is to act as smoothing functions to wisdom of the crowds. Incidentally the highly distributed case doesn't prevent capture, but rather towards Plutocratic settings instead of Autocratic as it is unlikely that a single individual can more or less perform a 51% attack (to put it in less politic, more tech nerd speak). So one can argue that it is easier to replace a few keys when things go wrong, in favor of centralization, or increased robustness, in favor of distribution.
There's plenty more suggestions and I still find a lot of them lacking. Most of this is even independent of such things like Capitalism and Socialism that people often conflate as solutions. The matter of the fact is that this is an incredibly complex solution space and I feel pretty confident saying that there is no global optima. And if we're going to get a little math nerdy, we can really think of voter preference as people considering "the optimization problem" (like we're solving a linear program) but each individual places different values for the various constraints (I believe lots of fighting revolves around this, as in any given subject we may be optimizing for different objectives but we assume others have the same conditioning on constraints as us so we think each other dumb). So we have this solution space that almost certainly has more than 100 independent dimensions, if not thousands, with magnitudes more interconnections as the systems are highly coupled, not to mention how the environment is dynamic and there are many outside forces (natural and human driven) that one can not vary.
Fuck this is a hard problem and I really wish we could just talk about how it is hard. I'm absolutely tired of hearing about easy solutions. If it was easy, we would have solved it thousands of years ago. Honestly, I think our resistance to even admitting to a little bit of complexity bars us from even making good steps in the direction of optimization. I know it is human nature to desire simplicity and waste energy with complexity, but we're 21st century beings and all the easy problems are already solved. So we need to rely on what makes humans unique, which is in big part about being able to understand multivariate systems and how long-term optimization does not necessarily align with short-term optimization (which these days we seem hyper-focused on short-term). That we can reject rules, momentum, and guides in recognition of the distinction of the alignment between desired outcomes and how we measure outcomes or set guides to achieve them (we are also very good at hacking all this, but machines are good at reward hacking, but not so good at alignment adjusting).
(Does anyone know if there have been any serious efforts to place a mathematical description on this optimization problem? Like a giant linear program? Seems like something someone would do, more for fun than utility)