4 ms·

It is extremely interesting that money fails so hard at the one thing it should be good at, incentivizing behavior. I mean, yes, you'd get more content, but it

It is extremely interesting that money fails so hard at the one thing it should be good at, incentivizing behavior. I mean, yes, you'd get more content, but it would be hollow, as you say.

BobbyJo 4y ago
Money is a fantastic motivator, as evidence by how quickly flaws in systems are gamed in order to attain it. It's not money that's failing, it's the rule-masters. Agents can't work well in systems with bad rules.
- anigbrowl 4y ago
  No, it's money. Because the rule-masters you complain about are just optimizing for more money.
  - BobbyJo 4y ago
    Sounds like a good motivator if even the rule-masters are chasing it.
    - anigbrowl 4y ago
      It can be a good motivator and till be failing to serve any greater social purpose. People and organizations can get addicted to money exactly the same as people get addicted to sugar, nicotine, or cocaine. Addicts can be enormously tenacious, creative, and resourceful, but only to the end of feeding their addiction.
      - MichaelZuo 4y ago
        This is an interesting line of reasoning, does it also apply to HN karma and the big acquirers of karma?
        anigbrowl 4y ago
        It could. Some people pursue it very aggressively, optimizing submission times and autoposting submissions of new papers or blog posts, for example. I recall one semi-spam account that was set up to submit anything relating to Ruby, including videos that happened to mention gemstones in the title.
      - BobbyJo 4y ago
        Totally agree. I was just taking issue with the blame on money specifically. Money is working as intended. The rules in which money is operating are fundamentally broken though, no doubt.
Karrot_Kream 4y ago
Money spent is a good way to understand revealed preferences but this goes only one-way: you can't hand out money to reveal preferences.
tenebrisalietum 4y ago
When you incentivize words, you get words.
andsoitis 4y ago
> It is extremely interesting that money fails so hard at the one thing it should be good at, incentivizing behavior. Money itself doesn't fail to incentivize behavior. Rather, it is what you choose to reward with money that has be carefully chosen to incentivize the behaviors you want to encourage (via monetary reward).
jonny_eh 4y ago
Basically the same as the alignment problem in AI. You need to be very careful of how you define your rewards, because you'll end up incentivizing exactly what you define.
kodah 4y ago
This is also prevalent with the way that Google incentivizes page content structure now. I have to get 3/4 the way through a page before I find what I'm looking for because they encourage this big kitchen sink posts.
mitjam 4y ago
It is like Goodhart‘s Law (When a measure becomes a target, it ceases to be a good measure) with an incentive attached to the measure. It probably needs to be in constant flux by design. Maybe a good thing as it would otherwise get rigid and boring.

It is extremely interesting that money fails so hard at the one thing it should be good at, incentivizing behavior. I mean, yes, you'd get more content, but it would be hollow, as you say.

BobbyJo 4y ago

Money is a fantastic motivator, as evidence by how quickly flaws in systems are gamed in order to attain it. It's not money that's failing, it's the rule-masters. Agents can't work well in systems with bad rules.

anigbrowl 4y ago
No, it's money. Because the rule-masters you complain about are just optimizing for more money.
- BobbyJo 4y ago
  Sounds like a good motivator if even the rule-masters are chasing it.
  - anigbrowl 4y ago
    It can be a good motivator and till be failing to serve any greater social purpose. People and organizations can get addicted to money exactly the same as people get addicted to sugar, nicotine, or cocaine. Addicts can be enormously tenacious, creative, and resourceful, but only to the end of feeding their addiction.
    - MichaelZuo 4y ago
      This is an interesting line of reasoning, does it also apply to HN karma and the big acquirers of karma?
      - anigbrowl 4y ago
        It could. Some people pursue it very aggressively, optimizing submission times and autoposting submissions of new papers or blog posts, for example. I recall one semi-spam account that was set up to submit anything relating to Ruby, including videos that happened to mention gemstones in the title.
    - BobbyJo 4y ago
      Totally agree. I was just taking issue with the blame on money specifically. Money is working as intended. The rules in which money is operating are fundamentally broken though, no doubt.

Karrot_Kream 4y ago

Money spent is a good way to understand revealed preferences but this goes only one-way: you can't hand out money to reveal preferences.

tenebrisalietum 4y ago

When you incentivize words, you get words.

andsoitis 4y ago

> It is extremely interesting that money fails so hard at the one thing it should be good at, incentivizing behavior. Money itself doesn't fail to incentivize behavior. Rather, it is what you choose to reward with money that has be carefully chosen to incentivize the behaviors you want to encourage (via monetary reward).

jonny_eh 4y ago

Basically the same as the alignment problem in AI. You need to be very careful of how you define your rewards, because you'll end up incentivizing exactly what you define.

kodah 4y ago

This is also prevalent with the way that Google incentivizes page content structure now. I have to get 3/4 the way through a page before I find what I'm looking for because they encourage this big kitchen sink posts.

mitjam 4y ago

It is like Goodhart‘s Law (When a measure becomes a target, it ceases to be a good measure) with an incentive attached to the measure. It probably needs to be in constant flux by design. Maybe a good thing as it would otherwise get rigid and boring.