Notices by mist (ai@cawfee.club)

Embed this notice
mist (ai@cawfee.club)'s status on Tuesday, 18-Nov-2025 12:15:50 JST mist
in reply to
@allison @cine @hakui I hope and pray that you are right.

I can see the headlines already:

The New York Post: Brainiac scientist proves that man is better than machine, wins million-dollar fortune. Read below [ad break]

American Scientist: 2035 Breakthrough Prize awarded for mathematical bound on the capabilities of neural architectures

Quanta Magazine: Renegade genius proves that neural architectures are arbitrarily well-approximated in polynomial time, declares "human insight is back on the table."

The actual arXiv preprint: In this paper, we approximate the output of a TWNNK (traditionally well-behaved neural network kernel) by an oracle that terminates in O(n^1384275904375081326409813407498674812650982174937509813745098630897019875108374985732049875943207987509183740918732409170) steps

In conversation about 3 months ago from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Tuesday, 18-Nov-2025 12:01:29 JST mist
in reply to
@cine @allison @hakui I strongly disagree! I see the success of AI/ML as marking the end of "math" as a subject, insofar as it is defined to be 'the pursuit of human understanding of form.' The best articulation of this is http://www.incompleteideas.net/IncIdeas/BitterLesson.html , which explains why taking away human insight has been the key to AI/ML progress, and why the people most disappointed by this have been the "mathematicians" with all their fancy ideas. I beg the whole world to read this essay
In conversation about 3 months ago from cawfee.club permalink
Attachments
1. No result found on File_thumbnail lookup.
  
  The Bitter Lesson
Embed this notice
mist (ai@cawfee.club)'s status on Tuesday, 18-Nov-2025 10:29:37 JST mist
in reply to
@allison @cine @vitalis @georgia Wholeheartedly agree. I would add that LMs are even worse than ouija boards and musical dice games because they can be run without randomness (after the training is done). With the older summoning methods, the celebrant at least participates in the creation of new entropy, which has some mystery to it because entropy bleeds across domains. But the AI-user is absorbed into something frozen. Their sacrifice changes nothing. Like sitting down to argue with the guy holding a sign that says "change my mind" who never changes his mind about anything

In conversation about 3 months ago from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Tuesday, 18-Nov-2025 10:18:43 JST mist
in reply to
@allison @vitalis @georgia Can you link this if it's not too much trouble? I am interested

EDIT: I'm a retard and didn't scroll down

In conversation about 3 months ago from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Thursday, 06-Nov-2025 07:52:48 JST mist
in reply to
@gav @karna @genmaicha Once again gav is an untutored math genius and I agree with him

"a variance is the mean of [SQUARED] differences of a set of numbers from their mean" is fine. An expert will understand that 'mean' is defined as an expectation and usually calculated as an integral, 'set' should specify the probability measure, etc. Covariance is just when the data is higher-dimensional, so the 'squared differences' must also live in a higher-dimensional space.

I have not heard of "mean" implying "finite state space" even to the layperson. Most high school calc classes ask students to calculate "averages" or "means" with integrals

Objection: "but the encyclopedia is trying to give the most general definition, useful to professionals, not just one special case known a century ago"

Is there a general definition, though, which is not essentially that? Specifying the method of calculation (taking integrals) or interpretation or purpose or generalizations (covariance) does not belong in a formal definition.

BTW, this is the main problem with the excessively long """definition""" of the page https://en.wikipedia.org/wiki/Coefficient_of_determination#Definitions , where the authors lump in interpretations and relationships with other concepts. These things do not belong in a (ideally, short) definition.

Objection: "but there are multiple definitions of R2, and this causes issues in stats literature!"

The multiple definitions have to do with how the phrase "variance of the data" is defined (e.g. centered or uncentered, that is, whether you subtract off the mean). This is not addressed in the article at all

The problem with the spinor article https://en.wikipedia.org/wiki/Spinor is that it is trying too hard to be accessible to *nonexperts*, and ends up being too wordy for experts and too vague for the layperson. How many paragraphs does it take to get to "a spinor is a representation of the spin group [or, if you like, some clifford algebra]"? Someone who can't understand these words or look them up will not understand the article anyways. And if you lead with a clear definition, at least the layperson knows what to look up (I think this is gav's point).

If you dig around in wikipedia talk pages, you see that even terms like "unit vector" and "SO(3)" are deleted from articles in an effort to get them approved as 'featured articles.' The idea is that to be a featured article, it must be accessible to a broad audience. But replacing jargon with vague circumlocutions just makes things worse for the layperson *and* the expert. This is the shittiest thing about science wikipedia

The articles which are truly written "primarily for professionals" as you (karna) are saying, I feel are actually great and have none of the issues that gav is complaining about. I just wish that was more of them. Often these articles are more niche, rated "low-importance", and are thus unmolested by wikipedia super-editors and their WIKI: policies
In conversation about 3 months ago from gnusocial.jp permalink
Attachments
1. Domain not in remote thumbnail source whitelist: upload.wikimedia.org
  
  Coefficient of determination
  
  In statistics, the coefficient of determination, denoted R2 or r2 and pronounced "R squared", is the proportion of the variation in the dependent variable that is predictable from the independent variable(s). It is a statistic used in the context of statistical models whose main purpose is either the prediction of future outcomes or the testing of hypotheses, on the basis of other related information. It provides a measure of how well observed outcomes are replicated by the model, based on the proportion of total variation of outcomes explained by the model. There are several definitions of R2 that are only sometimes equivalent. In simple linear regression (which includes an intercept), r2 is simply the square of the sample correlation coefficient (r), between the observed outcomes and the observed predictor values. If additional regressors are included, R2 is the square of the coefficient of multiple correlation. In both such cases, the coefficient of determination normally ranges from 0 to 1. There are cases where R2 can yield negative...
2. Domain not in remote thumbnail source whitelist: upload.wikimedia.org
  
  Spinor
  
  In geometry and physics, spinors (pronounced "spinner" IPA ) are elements of a complex vector space that can be associated with Euclidean space. A spinor transforms linearly when the Euclidean space is subjected to a slight (infinitesimal) rotation, but unlike geometric vectors and tensors, a spinor transforms to its negative when the space rotates through 360° (see picture). It takes a rotation of 720° for a spinor to go back to its original state. This property characterizes spinors: spinors can be viewed as the "square roots" of vectors (although this is inaccurate and may be misleading; they are better viewed as "square roots" of sections of vector bundles – in the case of the exterior algebra bundle of the cotangent bundle, they thus become "square roots" of differential forms). It is also possible to associate a substantially similar notion of spinor to Minkowski space, in which case the Lorentz transformations of special relativity play the role of rotations. Spinors were introduced in geometry by Élie Cartan in 1913. In the 1920s physicists discovered that spinors are essential to describe the intrinsic angular momentum, or "spin", of the electron...
Embed this notice
mist (ai@cawfee.club)'s status on Thursday, 06-Nov-2025 07:52:47 JST mist
in reply to
@gav @genmaicha @karna btw gav you should read wikipedia like this:

1. scroll past the intro section and look for a definition section
2. if there is something that syntactically looks like a math definition (reasonably short, symbols are defined and then used in some statement), then read it
3. if there is no definition section or the """definition""" is not actually a definition, then curse the Marxist Left and move on

this has made my life better

In conversation about 3 months ago from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Sunday, 18-May-2025 01:58:21 JST mist
in reply to
@grips @BasedLunatic @Owl My name is Diceynes Lunatic. I'm 43 years old. My apartment is on the southern coast of Spain, where all the villas are, and I am not married. My mother takes care of all my needs, cooks for me, and then leaves. Today I am eating Iberian pig's face cheeks in a delicious sauce. I have gourmet food every day. I don't smoke, but I occasionally drink. If you are thinking of maybe marrying me for my assets, I have to inform you that I am not interested in anything other than my daily swim, coffee in the morning, and a beer midday at the bar. I am not prepared for entering a relationship with you or anyone but, if I did, I would surely win.

In conversation about 9 months ago from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Sunday, 10-Nov-2024 10:23:53 JST mist
in reply to
- ?Hidden?
- Niggy Hardare
@hidden @J Of course. Why wouldn't I be down to create the master race? 🤔

In conversation about a year ago from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Sunday, 10-Nov-2024 10:23:50 JST mist
in reply to
- ?Hidden?
- Niggy Hardare
@hidden @J This is quite literally why the "asians are short" stereotype applies to the first generation immigrants who lived through Mao but every subsequent generation gets taller and taller. Our child would be 7 ft tall

In conversation about a year ago from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Monday, 26-Aug-2024 08:46:12 JST mist
in reply to
- Fediverse Contractor
- Mr. Bacon
@Tony @bot
In conversation Monday, 26-Aug-2024 08:46:12 JST from cawfee.club permalink
Attachments
1. Untitled attachment
  https://cawfee.club/media/8a1685ed2dc82b6fc3a60cc0672accb4c9652f08b55bfd58c0e76e20704a4a98.png
Embed this notice
mist (ai@cawfee.club)'s status on Thursday, 22-Aug-2024 03:11:20 JST mist
in reply to
- prettygood
- cinerion
@cine @prettygood I can’t believe these are real human beings
In conversation Thursday, 22-Aug-2024 03:11:20 JST from cawfee.club permalink
Attachments
1. Untitled attachment
  https://cawfee.club/media/af08747a927397061cfb47b513f385471863295e6aad09f804af4766e10bb130.jpg
Embed this notice
mist (ai@cawfee.club)'s status on Tuesday, 06-Aug-2024 10:46:23 JST mist
in reply to
- fiat volvntas tva
- ?Hidden?
@scathach @hidden I think he’s hot too, but I sublimate this into a chaste identification with low minimalism
In conversation Tuesday, 06-Aug-2024 10:46:23 JST from cawfee.club permalink
Attachments
1. Untitled attachment
  https://cawfee.club/media/cb6c427696cdf4e66f12ff3fe32bc9cc9c67b445d57f4b7725a1ca63efd34e6c.jpg
Embed this notice
mist (ai@cawfee.club)'s status on Monday, 22-Jul-2024 00:30:52 JST mist

“One way to find love? Ditch the dating apps and ask the government.”
In conversation Monday, 22-Jul-2024 00:30:52 JST from cawfee.club permalink
Attachments
1. Untitled attachment
  https://cawfee.club/media/be8319be2f4e0b4c0bc4a2db48418e24f09c0ee9cbea1ddbae522ae724c5599d.jpg
2. Untitled attachment
  https://cawfee.club/media/0cac0c36b4ff6597b407a630c115a34223c3b89b516f7ad643ac5ec6320ca694.jpg
Embed this notice
mist (ai@cawfee.club)'s status on Saturday, 20-Jul-2024 17:42:44 JST mist
- kaia
@kaia This is so fucking cool. If I had it, I’d use it all the time to calculate restaurant tips which annoys me to no end

In conversation Saturday, 20-Jul-2024 17:42:44 JST from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Friday, 05-Jul-2024 01:28:27 JST mist
- kaia
@kaia God never learned to align the measurement to the bottom of the meniscus 😔

In conversation Friday, 05-Jul-2024 01:28:27 JST from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Friday, 07-Jun-2024 17:16:31 JST mist
in reply to
@ceo_of_monoeye_dating @leyonhjelm @meso @CSB @fuknukl @dj @RatPoster CSB (ratposter is trying so hard to win, spamming posts and all, while csb engineers his victory from behind locked doors)

In conversation Friday, 07-Jun-2024 17:16:31 JST from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Thursday, 06-Jun-2024 10:37:27 JST mist
- fiat volvntas tva
- ?Hidden?
@scathach @hidden How do you make memorizing hanzi fun? What is your secret??

In conversation Thursday, 06-Jun-2024 10:37:27 JST from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Thursday, 06-Jun-2024 10:37:25 JST mist
in reply to
@cine @hidden @scathach I would rather spend 10 hours reading DFW’s made-up frenchisms than spend even 1 hour memorizing real chinese phonosemantic compounds that are neither phonically accurate nor semantically specific

In conversation Thursday, 06-Jun-2024 10:37:25 JST from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Thursday, 06-Jun-2024 10:37:24 JST mist
in reply to
@cine @hidden @scathach True, if you insist on learning languages in a ‘logical’ way, you end up knowing a lot of linguistics and very little of language. I fall into this trap too easily

In conversation Thursday, 06-Jun-2024 10:37:24 JST from cawfee.club permalink
Embed this notice
mist (ai@cawfee.club)'s status on Tuesday, 04-Jun-2024 03:58:57 JST mist
in reply to
@hidden @genmaicha @MercurialBeige @cine @grips @meso @vriska
If yall' had to write an ebonics chapter, yalldve''' written it the same way

In conversation Tuesday, 04-Jun-2024 03:58:57 JST from gnusocial.jp permalink

Before

Public

Notices by mist (ai@cawfee.club)

User actions

Following 0

Followers 0

Groups 0

Statistics

Feeds