fabfilter eqs are really good. I'm really missing them on linux
Oh, Tantacrul is very good channel if you are into music production! I like the ones where he breaks down DAWs and other production software, really liked the Sibelius video.
>>1515>It's ridiculously complicated. It's a partition problem: How many ways can you add up numbers to get to 20? ... So I ran a script on Python to calculate some of this. It got to 22 by running over a weekend, and I couldn't get it to go any further.
The trick to calculating this efficiently is to instead think about the expected end-of-turn score given that you have reached some current score. Suppose your strategy is to stop at 20. If you reach a score of 20, you will stop, so your expected end-of-turn score given your current score is 20. Simiarly, if you reach a score of 21, your expected end-of-turn score given your current score is 21, and so on. If you reach a score of 19, we can calculate your expected end-of-turn score given your current score by averaging the expected outcomes of your next die roll. If you roll a 1, you get 0. If you roll a 2, we have to look up the expected end-of-turn score given a current score of 19+2=21, which we already know is 21. And so on for 3 through 6. Using this method, we can calculate the expected end-of-turn score for lower and lower current scores until we finally reach what we want, the expected end-of-turn score given a current score of zero, as you have when you start your turn.
Letting expectation(score, stopping_point) be the expected end-of-turn score given a current score of "score" using a strategy where you stop at a score of "stopping_point", we can calculate as follows:#!/usr/bin/env python3
from fractions import Fraction
from functools import cache
def expectation(score, stopping_point):
if score >= stopping_point:
return Fraction(1, 6) * sum(expectation(score + roll, stopping_point) for roll in range(2, 7))
for n in range(41):
print(n, float(expectation(0, n)), expectation(0, n))
Runs in a fraction of a second.
This trick is very useful for solving otherwise hard probability problems.
Innovation gets passed around. Rarely does a good idea ever stay in the same place
hate how long these videos are
there's a link to an extended cut version in the description :P
Holy fuck that is so COOL. Made me wonder what the fuck I'm doing with my life pausing my openCV learning, a huge boost of motivation.
the strength of robotics is something that hasn't been televised/sensationalized much aside from Boston Dynamics, but it's quite amazing what modern sensors can do.
Guns and missiles are just point and shoot. Aircraft have had targeting for decades, missile launch systems are completely computerized, heck computer-operated flak batteries were even a thing back in WWII on the German side.
I remember watching a military documentary from either the 90s or early 20s on "smart guns". It's been a long while since I watched it, but I think the gist was that the aiming sight was computerized and would give aiming information to the operator, and then the bullets themselves would somehow course correct mid-flight to ensure contact with the target. Ultimately the documentary revealed that the project was canceled, likely due to cost and the end of the Cold War, but also as a pragmatic decision; if you're only going to be fighting farmers with AKs, it scarcely makes sense to invest in a new weapon when the current one works fine. If the new Cold War with China heats up, though, I wouldn't be surprised if that project gets taken off ice and we start seeing crazy auto-targeting rifles in actual combat usage.
Unfortunately, wars are a great impetus for progress.
bleh. "Early 2000s", not early 20s.
currently interested in revealing IRL sad video about homeless people and drugs.
Feel free on this board
It hid my bookmarks bar, is that it?
how many species have gone extinct in the past decade?
A lot, but most of them have been insects rather than mammalian species IIRC.