3 ms·
Dissolving the Fermi Paradox[0] is a great and easy to read paper to disabuse anyone of the idea of multiplying point estimates when you could convolve probabil
by the8472 5y ago
Dissolving the Fermi Paradox[0] is a great and easy to read paper to disabuse anyone of the idea of multiplying point estimates when you could convolve probability distributions instead.
[0] https://arxiv.org/abs/1806.02404 https://arxiv.org/abs/1806.02404
- floxy 5y agoI think this paper should be a top level HN discussion thread.
- the8472 5y agohas been https://news.ycombinator.com/item?id=17560462 https://news.ycombinator.com/item?id=17560462
- AceJohnny2 5y agoAlmost 3 years old, it's a candidate for repost.
- robocat 5y agoI was confused by the line in the original article “by multiplying uncertain variables, the level of uncertainty grows exponentially, to the point where no firm conclusions are possible.” Because when I looked for an explanation, it wasn’t exponential: https://foothill.edu/psme/daley/tutorials_files/10.%20Error%20Propagation.pdf https://foothill.edu/psme/daley/tutorials_files/10.%20Error%...