tines

29 ms·

481.
by tines 2y ago
Can’t you always ask why something wasn’t done sooner, no matter how soon it was done? I understand that longstanding issues can cover up real causes, but just because an issue is longstanding doesn’t mean it isn’t the real cause.
482.
by tines 2y ago
Man, China going straight for the 1984 playbook. Scary times.
483.
by tines 2y ago
What an interesting thought.
484.
by tines 2y ago
What heuristic distance function is being used? If it were using Euclidean distance, wouldn’t it be a lot more efficient and not pursue paths that are going in literally the opposite direction as the destination?
485.
by tines 2y ago
Is there an overview of the rules of the game somewhere?
486.
by tines 2y ago
I think the OP was trying to say that type theory is difficult, not that the kind of "no mutable state" idea is difficult.
487.
by tines 3y ago
Wow, cool to see Dr. Burtscher's name here! I was a student in his undergraduate parallel computing course and was invited to work in his lab. Very smart guy.
488.
by tines 3y ago
This is exactly what I've always wanted, hope it grows!
489.
by tines 3y ago
I am totally ignorant, can you explain more about Samsung's negative externalities?
490.
by tines 3y ago
Sabine Hossenfelder's explanatory video: https://www.youtube.com/watch?v=M0Xh-bGDVJ0
491.
by tines 3y ago
Yes, but that's more likely to be the LED emitting frequencies outside the range it's supposed to.
492.
by tines 3y ago
> To figure out whether they’re real or BSing you, the best way is to get them to tell you a lot of details about the situation—the more you get them to tell you, the harder it will be to BS all the details. This is exactly the feature t
493.
by tines 3y ago
This isn't what the paper gives evidence for. They found that if you use a different function in the training phase, the networks learn formal languages perfectly well. It's about the capabilities of a particular training process,
494.
by tines 3y ago
> However, replacing standard targets with the Minimum Description Length objective (MDL) results in the correct solution being an optimum. The paper is about how the "default" training doesn't work, not how neural network
495.
by tines 3y ago
No, the faculty-activists are the ones _being pushed back against_ by the faculty seeking the "primacy of teaching."
496.
by tines 3y ago
I mean, given the fact that the Government isn't a business with a goal of making money, that seems pretty good, no?
497.
by tines 3y ago
I thought this line near the end was interesting: > These rights and freedoms may in no case be exercised contrary to the purposes and principles of the United Nations. As if the purposes and principles of the United Nations will always
498.
The Universal Declaration of Human Rights(un.org)
29 points by tines 3y ago|39 comments
499.
by tines 3y ago
It's so fascinating to me how, in the minds of certain educated people, the fact that the author of a work did abhorrent things makes that work unethical or pernicious or taboo to consume and appreciate. Hasn't scholarly society a
500.
by tines 3y ago
Sounds like a computer. Why would getting access to one to play around encourage people to buy one? In other words, it sounds like a quantitative difference rather than a qualitative one.
501.
by tines 3y ago
What even is a mainframe, exactly?
502.
by tines 3y ago
Exactly - management culture and engineering culture are opposite ends of a spectrum.
503.
by tines 3y ago
I think I see what you mean now, and I understand that that would be possible. Wouldn't you need some extra parameters if it were a binomial distribution though? For a uniform distribution every item is equiprobable (i.e. there's
504.
by tines 3y ago
Sure, I guess I'm just arguing that even if you overlook that detail, then your intuition should tell you that the balls having a binomial distribution is suspect, i.e. you have to overlook a small thing and a big thing to misunderstan
505.
by tines 3y ago
I think the thing that clarifies it for me is that if the balls are chosen in such a way that they have a binomial distribution, then, intuitively, the question would be implying retrocausality by saying that N was also "chosen":
506.
by tines 3y ago
I thought that it was pretty clear from the part where it describes how the urn is filled: > n of them are red, and 100-n are green, where n is chosen uniformly at random in [0, 100]. N is chosen first, and that many red balls are put in
507.
by tines 3y ago
Not really; he's hypothesizing a potential draw order, but what I was listing were the possible contents of the urn before any are drawn. Since the contents of the urn aren't ordered, they form a set, not a list, and shuffling ele
508.
by tines 3y ago
The balls in the container are not ordered.
509.
by tines 3y ago
I think this is easiest to see if you imagine the urn is filled with three balls. You draw one, and it's red. The possibilities are that the urn originally contained: 1. Red, Red, Red 2. Red, Red, Green 3. Red, Green, Green (The Green,
510.
by tines 3y ago
Putting class methods in separate impl blocks is actually objectively better; it allows you to define new methods on types that you didn't write, like typeclasses in Haskell.

More ›