3 ms·

So an elementary model which cannot possibly hope to model all the real world variables is gonna predict the future? No no, this is too much but it's a cute toy

• actually_a_dog 5y ago
  Unfortunately, real world data since the 1970s matches up well with this "elementary model". See this comment: https://news.ycombinator.com/item?id=29535554 https://news.ycombinator.com/item?id=29535554 Do you have a more specific or substantive criticism of this work? Otherwise, I'd say it's essentially proven itself.
  • bart_spoon 5y ago
    The predictions from the model all match up in line with the parts of the model predicting an increase in various metrics, but the crux of the model, the inflection point and collapse, all have yet to occur yet. In other words, there would be nothing to prove that this model is any better or worse than a model that predicts these metrics continue to increase indefinitely.
    • actually_a_dog 5y ago
      The fact that there is good agreement between the model and real world data extending 40 years past its creation should, at least, raise your prior probability that the model is correct enough to be useful. If not, you're just being irrational.