I've said it before, and I'll say it again: one of the only few bad things about Tolkien's legendarium is that it makes 90% of all other fantasy worlds look either completely or somewhat mediocre in comparison.
Like, what do you mean you don't have a fictional language for your fantasy world? WEAKLINGS
“On the use of the phrase ‘on the’ and the colon-delimited subtitle in academic papers: A meta-analysis”
When my mind hands me a small thought, I can ask “Is this true?” and my mind will answer “Good question, let’s go over three thousand different rationality techniques and see what answer each of them gives.
When my mind hands me a big thought, I ask "Is this true?” and my mind just answers “It better be, m——–r, cause this is how you’re gonna be thinking about everything from now on.”
Zero is a funny number. The opposite of one is negative one. But the opposite of zero is still zero.
imagine being a billionaire. like, you put in work, nearly fail, but push onward anyways. you develop something novel. you make it scale. you push the envelop to do something that's never been done before. but then, part of the thing you win in the end game is a bunch of communists saying you should die or that you should be forced to hand over what you've earned smh. what an infantile ideology
One drawback is that, despite having other work-related things to do, I am essentially drowning in the vast amount of material that is continuously recommended to me—books and information that are both good and very difficult to put down.
Both my biologist and math friends are constantly giving me new ways of looking at the world. I have a rich inner life because of this.
Xylem and phloem be like 💀
"If we keep on optimizing the proxy objective, even after our goal stops improving, something more worrying happens. The goal often starts getting worse, even as our proxy objective continues to improve. Not just a little bit worse either — often the goal will diverge towards infinity. This is an extremely general phenomenon in machine learning. It mostly doesn't matter what our goal and proxy are, or what model architecture we use. If we are very efficient at optimizing a proxy, then we make the thing it is a proxy for grow worse."
Too much efficiency makes everything worse: overfitting and the strong version of Goodhart's law
I remember Toonami. And Cowboy Bebop and Gundam Wing and Dragonball Z and Yu Yu Hakusho and Rurouni Kenshin at 3am. Anime reruns on the screen.
people should learn more history so that when it comes time to make comparisons it’s not Everything Is Transatlantic Slavery Or The Holocaust