projectglint.com
146 posts
124 followers
346 following
Regular Contributor
Active Commenter
comment in response to
post
They do but they go to the human hell which turns out to be pretty good for the bacteria themselves
comment in response to
post
Actually 4
comment in response to
post
Less than 5 months
comment in response to
post
what i really want is to just blend a bunch of different feeds together, without having to write that algorithm.
comment in response to
post
Juxtastat 626 5/5 (mogged (i'm only replying because this is the first time I beat your score))
🟩🟩🟩🟩🟩
juxtastat.org
comment in response to
post
it's absolutely miserable that they can't take any sort of blame. I understand why they won't say "Elon did it", but they won't even take the responsibility as an organization. Fundamentally childish people
comment in response to
post
there's been a trend, even before AI art, for people to think that "the point of art is to be 'good' at it"
Like people are trying to be supportive but I think it's actually weirdly degrading to have your art reduced to the skill required to produce it
comment in response to
post
There could be a point that, even if you have a ton of money, you literally won't be able to buy the goods you need due to shortages. Just like during covid.
comment in response to
post
They think that when you buy something from a store, the store has "stolen" from you with "unfair business practices" despite the fact that YOU GOT SOMETHING FROM THE TRANSACTION.
It would be better if it was just based off of Trump's personal grudges.
comment in response to
post
The Vietnamese and Cambodians don't import many US goods, because US goods are expensive. This leads to a very high trade deficit, which the administration has incorrectly interpreted as a "tariff".
comment in response to
post
I really dislike Jonathan Haidt. I have this sense when listening to him that he wants to impose his morality on everyone else. He seems incurious as to what actual kids are feeling.
comment in response to
post
We should not just think about removing the phones, but solving the problems technology is currently solving. Otherwise, it's just an expansion of the safetyism that permeates our culture already.
comment in response to
post
People use technology primarily because they solve problems. The assertion is that it's a prisoners dilemma, that everyone gets stuck using phones because everyone is defecting, but I think it would be more useful to think of a phone as solving many different issues.
comment in response to
post
A smartphone for me in highschool wasn't just a way for me to talk to friends, it was a lifeline to keep me from the insane boredom in the classroom that taught to the lowest common denominator.
comment in response to
post
He takes a very parent-centric view of any problems with kids, and although I'm not surprised, I am disappointed. He has an 18 year old kid, who probably plays video games, but it seems he hasn't played a video game since he was a child.
comment in response to
post
I really dislike Jonathan Haidt. I have this sense when listening to him that he wants to impose his morality on everyone else. He seems incurious as to what actual kids are feeling.
comment in response to
post
a good mascot
comment in response to
post
I wonder how Iowans specifically are going to react to this. Will this finally be the thing that turns people on Trump?
comment in response to
post
I like large standard libraries, but I don't exactly know why, and also I like rust.
comment in response to
post
Wow, this is incredible
comment in response to
post
comment in response to
post
This makes me wonder if a LLM can, given the description of a small DFA, simulate that DFA? When do they break down, is it just at context limits or earlier?
I'm also curious why LLMs don't already have this ability. If we trained them on related data, would they gain this skill (generalizably?)
comment in response to
post
When you're 6, you connect the blue dots and color the touching bits
When you're 16-30, you see at least 3 possible solutions and you have a long internal debate about which one is simplest
When you're 30+, you connect the blue dots and color the touching bits
comment in response to
post
So I think it would look something like:
1. Define the predicate as a type (for example, NonZero for n != 0)
2. Define a constructor for that predicate that takes a proof that n != 0 and gives you `NonZero n`
3. You can then use `NonZero` to make more predicate-types
comment in response to
post
It feels lazy to me to just say "dependent types" because dependent types are the ur-feature for type systems
comment in response to
post
I should also add that when looking at electron probability density plots, you're almost always looking at hydrogen, which is a single electron. That single electron "takes up" all the space in the sphere so it looks like the spherical harmonics for the entire sphere.
comment in response to
post
Therefore the other part, the kinetic energy part, is described not by the distance to the center of the atom but where it is on a sphere. This is a perfect fit for spherical harmonics
comment in response to
post
This is because the electron doesn't "prefer" to be specifically on top/bottom/left/right/etc to the proton, all it cares about is distance from the proton.
comment in response to
post
r is the distance between the electron and proton, and is a scalar. pi and eps_0 are constant factors. q^2 is a scalar from the electrons charge. This means this entire term is spherically symmetric.
comment in response to
post
Ok here's what I got after reading a bit.
You can read the following equation in this way:
Total Energy = Kinetic Energy + Potential Energy
So E*psi is total energy, h/2u(del^2psi) is kinetic, and q^2/(4 * pi * eps_0 * r)*psi is potential.
comment in response to
post
If there are "parallel universes" it's shocking how little we can access from those parallel universes. If we can only interact with these parallel universes a little bit it seems like they are correspondingly "less real" to me.
comment in response to
post
I find myself drifting towards the latter two because of Holevo's theorem en.wikipedia.org/wiki/Holevo'... which states that you can only retrieve n classical bits of information from n qubits.
comment in response to
post
In many-worlds: There are exponentially many real worlds
Bayesian: There is 1 real world, and the "exponentiality" is a type of illusion
Bohmian: The other probabilities are kinda real, but one of them (what actually occurs) is more real than the rest