4 ms·

I don't really agree. I think you can know what ingredients are more likely to make you happy, and you can direct your life towards them. With your outlook wha

I don't really agree. I think you can know what ingredients are more likely to make you happy, and you can direct your life towards them. With your outlook what does one do? Just sit there and wait for happiness to appear?

csdvrx 4y ago
> I think you can know what ingredients are more likely to make you happy, How do you know you know for sure? Are you positive you know everything that can make you happy? It's a knowledge/discovery problem. > With your outlook what does one do? Just sit there and wait for happiness to appear? No, we can use your approximation indeed, but after the facts (ex-post) instead of before (ex-ante): you need take sample of the multidimensional space then try to do gradient descent. The sampling part has to be random - so you randomly try things. Some of things will make you happier, and sometimes much more than they should, so you tweak one parameter, try again and see if you are less happy or more happy. That's just using algorithms on actual life.
- brokenmachine 4y ago
  Pretty disruptive to try out big changes on a whim though such as different jobs or living location, and those kinds of changes are the ones I would think are most likely to have a large effect on happiness level.
  - csdvrx 4y ago
    First, you are assuming the size of the change to be proprotional to the change in hapinness level - that's a questionable first assumption! Maybe it's the opposite? Or maybe it's uncorrelated? By default, to play it safe, keep the randomness and don't preselect for a big change! A random change would be of any size: say, randomly try a different food today? Then go to a aerobic class instead of a boxing class (or vice versa) Or try swaps: if you often to go vacation on the same place you enjoy, have you tried swapping it to your main residence and work remotely? That should be less disruptive, through all change is disruptive!

csdvrx 4y ago

> I think you can know what ingredients are more likely to make you happy, How do you know you know for sure? Are you positive you know everything that can make you happy? It's a knowledge/discovery problem. > With your outlook what does one do? Just sit there and wait for happiness to appear? No, we can use your approximation indeed, but after the facts (ex-post) instead of before (ex-ante): you need take sample of the multidimensional space then try to do gradient descent. The sampling part has to be random - so you randomly try things. Some of things will make you happier, and sometimes much more than they should, so you tweak one parameter, try again and see if you are less happy or more happy. That's just using algorithms on actual life.

brokenmachine 4y ago
Pretty disruptive to try out big changes on a whim though such as different jobs or living location, and those kinds of changes are the ones I would think are most likely to have a large effect on happiness level.
- csdvrx 4y ago
  First, you are assuming the size of the change to be proprotional to the change in hapinness level - that's a questionable first assumption! Maybe it's the opposite? Or maybe it's uncorrelated? By default, to play it safe, keep the randomness and don't preselect for a big change! A random change would be of any size: say, randomly try a different food today? Then go to a aerobic class instead of a boxing class (or vice versa) Or try swaps: if you often to go vacation on the same place you enjoy, have you tried swapping it to your main residence and work remotely? That should be less disruptive, through all change is disruptive!