5 ms·

There's a great amount of truth to this; anyway, it's essentially what Michael Kalecki argued in 1943 in his "Political Aspects of Full Employment". Anyway, I

There's a great amount of truth to this; anyway, it's essentially what Michael Kalecki argued in 1943 in his "Political Aspects of Full Employment". Anyway, I find it unfortunate that the most frightening reform in the face of the rentier class is Marxism-Leninism, a system that simply doesn't work. It honestly seems strange that the radical proposals of Henry George (and his Land Value Tax[1] that would confiscate all rents) aren't used anymore to challenge this class. Last time Georgism infused economic reform, we had the age of trust-busting and worker protections. I hope I'm not dreaming when I think this could return. [0] http://delong.typepad.com/kalecki43.pdf http://delong.typepad.com/kalecki43.pdf [1] https://en.wikipedia.org/wiki/Land_value_tax https://en.wikipedia.org/wiki/Land_value_tax

danieltillett 9y ago
Marxism-Leninism might not work as a mechanism for efficient production, but it is mighty effective in exapropriating the rich. One thing I have thought about recently is that with better ML if it would be possible today to model the economy well enough to now make Marxism work in practice. The soviets tried in the 1960s using linear algebra, but their models were not practicable. At some point it should be possible to model the economy well enough that capitalism is no longer the only viable alternative.
- bufordsharkley 9y ago
  Hell, it's easy to imagine that machine learning could do it. But like any man-made system, how do we guarantee that the system isn't trained on inputs that generate bias? For in a large economy, bias in allocation is a form of human suffering, large or small.
  - danieltillett 9y ago
    I am sure there would be bias - the question is would it be any worse than the serious bias inherent now. Anyway it is just a thought I found interesting :)
    - bufordsharkley 9y ago
      It's very interesting. I'd worry most about one thing: that the free enterprise system is inherently flexible, and will allow for new inputs its "designers" never considered. It'd be a very poor outcome for a neural network that it refuses to consider a new industry that it wasn't trained on... imagine a network trained in the 1950s introduced to the technology of today. (Though perhaps a different way of framing things would let this fall away.) In general, a market of prices is a useful system for measuring how much people want things. Without a price system, we'd need some new way to measure this...
- vidarh 9y ago
  > but it is mighty effective in exapropriating the rich. It is much more effective in moving wealth from one elite to a new elite. The problem with Leninism is that it basically sets up a "us vs them" system where those best at exploiting that mentality ends up on top. Hence people like Stalin. It was not lack of supporting technology that made the Soviet Union fail, but that they started with a coup d'etat against a liberal socialist government that actually had a support base, and from then on whether or not a theory was pleasing to the party elite mattered more than whether or not something was productive or scientifically valid. See for example Lysenko. You can not achieve productivity in a system that is geared towards punishing you if your ideas displease a leadership that is there largely because they are the ones most willing to grab power and with least compunction to use it.

danieltillett 9y ago

Marxism-Leninism might not work as a mechanism for efficient production, but it is mighty effective in exapropriating the rich. One thing I have thought about recently is that with better ML if it would be possible today to model the economy well enough to now make Marxism work in practice. The soviets tried in the 1960s using linear algebra, but their models were not practicable. At some point it should be possible to model the economy well enough that capitalism is no longer the only viable alternative.

bufordsharkley 9y ago
Hell, it's easy to imagine that machine learning could do it. But like any man-made system, how do we guarantee that the system isn't trained on inputs that generate bias? For in a large economy, bias in allocation is a form of human suffering, large or small.
- danieltillett 9y ago
  I am sure there would be bias - the question is would it be any worse than the serious bias inherent now. Anyway it is just a thought I found interesting :)
  - bufordsharkley 9y ago
    It's very interesting. I'd worry most about one thing: that the free enterprise system is inherently flexible, and will allow for new inputs its "designers" never considered. It'd be a very poor outcome for a neural network that it refuses to consider a new industry that it wasn't trained on... imagine a network trained in the 1950s introduced to the technology of today. (Though perhaps a different way of framing things would let this fall away.) In general, a market of prices is a useful system for measuring how much people want things. Without a price system, we'd need some new way to measure this...
vidarh 9y ago
> but it is mighty effective in exapropriating the rich. It is much more effective in moving wealth from one elite to a new elite. The problem with Leninism is that it basically sets up a "us vs them" system where those best at exploiting that mentality ends up on top. Hence people like Stalin. It was not lack of supporting technology that made the Soviet Union fail, but that they started with a coup d'etat against a liberal socialist government that actually had a support base, and from then on whether or not a theory was pleasing to the party elite mattered more than whether or not something was productive or scientifically valid. See for example Lysenko. You can not achieve productivity in a system that is geared towards punishing you if your ideas displease a leadership that is there largely because they are the ones most willing to grab power and with least compunction to use it.