HNHacker News
TopNewBestAskShowJobs

Chinjut

3,384 karma · joined June 23, 2014

submissionscomments
Chinjut··on Hi everyone yes, I left OpenAI yesterday
It was always like this ("Now the system just makes shit up and says 'Yes you are right' when you catch it on BS."). The scales are just falling from your eyes as the novelty fades.
Chinjut··on Researchers have found a faster way to do integer linear programming
"The problem’s dimension influences the dimension of this shape: With two variables it takes the form of a flat polygon; in three dimensions it is a Platonic solid, and so on."

It certainly needn't be a _Platonic_ solid! The author must have meant to write merely "solid" or "polyhedron" or some such thing

Chinjut··on Post-apocalyptic life in American health care
It is odd to write this and not in one's conclusions realize or call attention to how this "Whose responsibility is this?" problem all goes away with single payer. In countries with single payer healthcare, this problem never comes up.
Chinjut··on Ninth Dedekind number found by two independent groups
There needn't be a layer such that all layers below are completely white and all layers above are completely blue.

For example, consider a cube. It has four layers. We could say the bottom-most layer is the one with (0, 0, 0).

Above that, there's a layer which contains (0, 0, 1), (0, 1, 0), and (1, 0, 0) [the entries with a single 1].

Above that, there's a layer which contains (1, 1, 0), (1, 0, 1), and (0, 1, 1) [the entries with two 1s].

Above that, there's a layer which contains (1, 1, 1).

We could make (0, 0, 0), (0, 0, 1), (0, 1, 0) and (0, 1, 1) white, while making (1, 0, 0), (1, 1, 0), (1, 0, 1), and (1, 1, 1) blue.

Note that all the blue values will have a 1 in the first position, while all the white values will have a 0 in the first position, ensuring monotonicity.

But both the "single 1" and the "double 1" layers have a mix of blue and white within them. So there's no single transition layer, as you put it.

Chinjut··on What does it mean for a monad to be strong?
Given a monad M on a category C, it is among other things a functor from C to C, which means that we have functions from Hom(c, d) to Hom(M(c), M(d)) for each pair of objects c and d in C. [Here, by Hom(x, y), I mean the set of morphisms from x to y in C.]

If C is a so-called "closed category", not only is there an actual set Hom(x, y) of morphisms from x to y, but also, there is internal to C a corresponding object x → y.

We say a monad is "strong" if there are morphisms from x → y to M(x) → M(y) [note that these are morphisms in C between objects of C] which suitably correspond to the aforementioned functions from Hom(x, y) to Hom(M(x), M(y)) [note that these are morphisms in Set between objects of Set].

This is the distinction between having an operator external to your programming language that rewrites functions from x to y into functions from M(x) to M(y) (plain monad) vs furthermore having this operator itself represented internal to your programming language by a function from x → y to M(x) → M(y) (strong monad).

(This notion of strength really doesn't have to do with monads specifically, and applies more generally to functors. It's the same as what people call being an "enriched functor".)

The notion of a strong monad on a closed category C goes beyond just the plain notion of a monad on C, because it asks for some structure to exist internal to C which otherwise only has to exist in Set. But if C itself is Set, this is automatic. This is why monads on Set are automatically strong.

This is all muddied in the way people in Haskell world talk about monads and functors, because they never talk about monads on arbitrary categories, or functors between arbitrary categories. They never talk about categories at all because they only ever have one category in mind. They only ever talk about monads on Set, or functors from Set to Set [or rather, "Set" where "sets" are taken to be Haskell types and "functions" are taken to be Haskell functions. But this amounts to about the same thing, if you only ever work internal to "Set"]. This can make for some confusion in translating between how Haskell people use category-theoretic jargon and how category theorists use it.

But there are more categories in the world beyond just the ambient category Set. For a simple example of a closed category and a non-strong monad on it, consider the four-element Boolean algebra with 0 < p, q < 1. We can construe this as a category, where there is a (unique) morphism from any value to any equal or larger value. This category is in fact cartesian closed; we have Boolean negation such that p and q are each other's negations, and more generally, the interpretation of x → y becomes the material implication operation such that x → y = (NOT x) OR y.

If we define on this the operation M which sends 0 to itself but sends all other values to 1, we find that this M is indeed functorial: Whenever x ≤ y, we have M(x) ≤ M(y). And this M is furthermore a monad: It's easy to verify that x ≤ M(x) and M(M(x)) ≤ M(x) for all x. [All the other equations that go into being a monad are trivial in this partial order context where all parallel morphisms are equal]

But this M is NOT a strong monad! Even though it's true from our external perspective that x ≤ y entails M(x) ≤ M(y), that is not visible internal to this category, so to speak, in that we do not in general have that (x → y) ≤ M(x) → M(y). For example, p → 0 = q, but M(p) → M(0) = 0, and we do not have q ≤ 0.

Another example along the same lines is where we take our category C to be Set^2, whose objects are pairs of sets and whose morphisms are pairs of functions, all composing component-wise, and equip it with a monad which sends (empty set, empty set) to itself and sends everything else to (one element set, one element set).

There are also less trivial, more complicated examples, but this should perhaps get the idea across of what a non-strong monad. Of course, it requires thinking about categories in general to begin with.

Chinjut··on Scientists may have found mechanism behind cognitive decline in aging
I'm having a very hard time understanding this comment. Perhaps my cognitive decline has begun.

Update: I see I should rewatch "Life of Brian".

Chinjut··on About math limitations
What is this capitalization?
Chinjut··on Twitter to limit Direct Messages by non-paying users
For those who cannot access Twitter, the official announcement is this: "We'll soon be implementing some changes in our effort to reduce spam in Direct Messages. Unverified accounts will have daily limits on the number of DMs they can send. Subscribe today to send more messages".
Chinjut··on Notes on Puzzles
In the described "perpetual check", why does the king not move to h2 to escape the perpetual check?
Chinjut··on Why is the volume of a cone one third of the volume of a cylinder? (2010)
If you call multiplying the Fourier transform by a linear function "high-pass filtering" and dividing by a linear function "low-pass filtering", then I understand what you are saying. But I'm not accustomed to calling those high-pass filters and low-pass filters. I think of high- and low-pass filters as multiplying the Fourier transform by a step function (ideally, or some close approximation of such).
Chinjut··on Why is the volume of a cone one third of the volume of a cylinder? (2010)
Yes, that's exactly what happens. Think of a point in an n-dimensional hypercube and the pyramids made by connecting it to the n many opposite faces. These pyramids each have volume 1/n times the hypercube's volume. And in the same way, in n dimensions, any tapering figure has volume given by its height times its base divided by n.
Chinjut··on Why is the volume of a cone one third of the volume of a cylinder? (2010)
Differentiation is high pass filtering? How so? Is low pass filtering followed by high pass filtering the same as leaving a signal unchanged?
Chinjut··on Twitter Is DDOSing Itself
Do you enjoy having weekends off from working on somebody else's goals?
Chinjut··on A semester of Haskell-language-server: an internship report
Haskell didn't have structs? What could you possibly mean by this?
Chinjut··on Don Knuth plays with ChatGPT
Why are those the token boundaries? "mirth" decomposes as "m" + "irth"? "grins" decomposes as "gr" + "ins"?
Chinjut··on Texas professor fails entire class from graduating- claiming they used ChatGTP
Yes, I see in retrospect I was misinterpreting the comment I was replying to and have edited it slightly accordingly.
Chinjut··on Texas professor fails entire class from graduating- claiming they used ChatGTP
Why does anyone trust ChatGPT to correctly answer this question? It's a text prediction engine, not an all purpose correct answer generator. A yes or a no answer is the sort of text that would follow a question like this, certainly.
Chinjut··on The Two Cultures (1959) [pdf]
"The second law of thermodynamics" is also a rather undescriptive name. Which one is called the first law, which one is called the second law, which one is called the third law, etc? I can't keep it straight myself.
Chinjut··on The Two Cultures (1959) [pdf]
If you asked people "Will a broken glass mend itself spontaneously?", they will all get it correct. That's not the sort of question people fail to be able to answer. The sort of thing question people fail to be able to answer is "What is the Second Law of Thermodynamics?".
Chinjut··on A Programmer's Introduction to Mathematics
It's common to say that constructive mathematics is continuous, as it cannot prove the existence of a discontinuous function. One can certainly reason constructively about the various things traditionally considered continuous mathematics. I find your opposite perspective on it a bit odd.
Chinjut··on “Twitter is going great” is being sunset
Twitter is currently down. https://news.ycombinator.com/item?id=35778041
Chinjut··on Steven Spielberg: ‘No film should be revised’ based on modern sensitivity
He's not called that in the book. People constantly say he is, with an air of authority, but go read it yourself. Go do a search on the Project Gutenberg text. He's just called Jim. Not once is there any instance in the book where anyone takes his name to be anything other than Jim. I don't know where this popular misconception was invented. Yes, the racial term is said in the book, but not as part of his name.
Chinjut··on A $300k Salary Feels Like $100k in the Priciest US Cities
I advise anyone making $300k to give half of it to me. Then you will end up at $150k, which is $50k ahead of where you started.
Chinjut··on Nyquist Frequency
Whoops, I should've pulled the division by x out of the definition of f. The example I had in mind was [sin(4πx) - 2 sin(6πx) + sin(8πx)]/x. [Another good example is [sin(6πx) - 2 sin(8πx) + sin(10πx)]/x, whose frequencies are between 3 Hz and 5 Hz, thus a bandwidth of 2 Hz, but sampling at 4 Hz or even 8Hz gets all zeroes.]

Anyway, the details on that example don't matter, the Wikipedia graph and article makes things more clear.

Chinjut··on Nyquist Frequency
That is, it's not quite as simple as saying you just need to sample at any frequency at least twice the bandwidth. Rather, it's the more complicated behavior described by this graph: https://en.wikipedia.org/wiki/Undersampling#/media/File:Samp.... That is, the general rule is that the ratio of the highest frequency in the signal to half the sample rate, and the ratio of the lowest frequency in the signal to half the sample rate, have to lie within an interval of consecutive natural numbers.

When the lowest frequency is zero, this is the familiar rule that the sample rate has to be at least twice the highest frequency in the signal. But more generally, it's more complicated.

Chinjut··on Terence Tao on GPT-4
Looking at the actual input/output prompts he shared (https://mathstodon.xyz/@tao/110178324935054737), this seems like not very much. These tasks seem really trivial regardless. Entering line breaks that were removed when copy and pasting from a PDF, and looking up the most basic functionality in the spreadsheet software he was already using.
Chinjut··on Nyquist Frequency
Consider a signal whose value at x seconds is f(2x) - 2 f(3x) + f(4x), where f(x) = sin(2πx)/x. Considering that the absolute frequencies of f(x) are uniformly distributed from 0 to 1 Hz, the absolute frequencies of this total signal should be constrained to between 2 and 4 Hz. Thus, a bandwidth of 2 Hz. But if we sample at 6 Hz (three times the bandwidth!) including x = 0, we'll get all zeros.

Granted, we might say that from the perspective of the complex Fourier transform using signed frequencies, the frequencies of this signal actually range over [-4 Hz, -2 Hz] U [+2 Hz, +4 Hz]. But I'm not sure that's the interpretation you had in mind.

Let me know if I've screwed anything up here!

Chinjut··on Twitter is now blocking Matt Taibbi from search results
The claim is that no posts by Matt Taibi show up on Twitter search. And that is true. No posts made by Matt Taibbi can be found via Twitter search. Yes, his profile comes up and you can click through to it. But his posts themselves are not indexed by search in the usual way. If you want to find a post Taibbi made months ago, for example, good luck doing so.
Chinjut··on If parrots can talk, why can’t monkeys?
I believe there is significant scientific skepticism as to whether Koko was communicating to the degree popularly portrayed.
Chinjut··on Twitter's Recommendation Algorithm
Perhaps that's related to this line. Though perhaps this is just used for observing metrics. https://github.com/twitter/the-algorithm/blob/7f90d0ca342b92...
← PreviousPage 8 of 32Next →