kkoncevicius

I have to agree with your point about EDA. The library is neat, but even the example of covariance matrix animation is a bit contrived.

Every pixel has a covariance with every other pixel, so sliding though the rows of the covariance matrix generates as many faces on the right as there are pixels in a photograph of a face. However the pixels that strongly co-vary will produce very similar right side "face" pictures. To get a sense of how many different behaviours there are one would look for eigenvectors of this covariance matrix. And then 10 or so static eigenvectors of the covariance matrix (eigenfaces [1]) would be much more informative than thousands of animated faces displayed in the example.

Some times a big interactive visualisation can be a sign of not having a concrete goal or not knowing how to properly summarise. After all that's the purpose of a figure - to highlight insights, not to look for ways to display the entire dataset. And pictures that try to display the whole dataset end up shifting the job of exploratory analysis to a visual space and leave it for somebody else.

Thou of course there are exceptions.

[1]: https://en.wikipedia.org/wiki/Eigenface

Maybe one lesser known is changing to visual mode within operation. A few examples (^ marks the cursor):

    one two three
                ^

'db' would leave letter "e" behind. But 'dvb' would include the "e" as well.

Another example, say we want to delete from "two" to "five".

    one
    two
     ^
    three
    four
    five
    six

'd/five' would leave a "t" as well as "five", but 'dV/five' would delete all necessary lines. This helps targeting lines without using relative line numbers.

I had the same experience with using a paper notebook. There are various systems and tips about how and what to write in a notebook to become more productive. But the best way for me was just to use it as a sheet of paper and write about what feels necessary at the moment with no structure whatever.

There is a "hide" button which lets you hide certain posts from the front page. Maybe that's a good solution for your case?

Your comment mirrors my experience with both close friends and a spouse. One time my now ex-wife asked why I don't share my feelings more. When I did she said she felt unsafe and we started talking about her instead. In my anecdotal experience men are routinely trained not to talk about their troubles and emotions. Even if I had some form of emotional support I am not sure I would know how to open up. And I am not sure I would want to.

I think you are taking yourself too seriously.

I am suspicious of that driver effect being caused solely by sun exposure. If that were the case we would see the same effect in all drivers.

Anecdotal, but the parts of my skin that see more sun exposure (forearms) look healthier compared with parts that are more covered (upper arms). I wonder if maybe gradual sun exposure starting from spring and increasing towards the summer is healthier compared to sudden intense tanning of the area that is otherwise constantly covered by clothes.

Another alternative is MacPorts. I've been using it since my first Mac (8 years ago), and haven't found a reason to change.

To me it's quite telling that even John Anthony West is among those who say this structure is natural. J. A. West was famous for his theory about Sphinx showing signs of water erosion and hence possibly being 12,000 years old.

I have an alternative, more generous explanation of your 2nd point - the people you talk about haven't done much reflection about consciousness yet and are so immersed in it that they cannot separate their own conscious experience as an entity to talk about. Like fish and water. Just like you said - it was also your position when you were younger. The people you label as having a different interior experience might be in the same position as your younger self.

To me it seems your reply is conflating consciousness with perception. The perception is the consciousness. Auditory illusion, for example, is just signals to the senses and your senses miss-representing the inputs. The consciousness is the part which is aware of these sensations. If they are accurate or not - is not the point. The point is that you are aware of them.

Just wanted to say how much I like your answer. To me one of the bigger puzzles is why people have such different takes on consciousness. To some the idea of p-zombies is immediately clear. To others it is nonsense. But during my conversations about the topic I was never able to explain it adequately to someone sceptical. And from my perspective they (the sceptics) always conflate being conscious with talking, learning, thinking, remembering, etc.

To me it seems that the hard problem of consciousness can be stated a lot simpler, like so:

How can we tell if another person is conscious or not?

As far as I see, this is not possible and will never be possible. Hence the "hard problem".

Same for R

A long term storage that can also be used as a trap wire to catch pathogens.

[1]: https://en.wikipedia.org/wiki/Neutrophil_extracellular_traps

Might be relevant and interesting:

https://www.thenewatlantis.com/publications/getting-over-the...

Crystallography Open Database:

http://www.crystallography.net/cod/result.php?year=2024

http://www.crystallography.net/cod/1570196.html

Genecards: https://www.genecards.org/cgi-bin/carddisp.pl?gene=OR52L1

Scroll down to see lots of different kinds of data about a particular selected gene, with a navigation menu sticking at the top.

love the separate quotes for "AI" and "features"

"The reign of quantity and the signs of the times" - Rene Guenon

Since it is on topic - sharing useful link I have in my bookmarks about transformations in general: http://fmwww.bc.edu/repec/bocode/t/transint.html

The Reign of Quantity and the Signs of the Times - Rene Guenon (1945).

An esoteric book that summarises everything as a direction which moves away from quality and towards quantity. Here is an excerpt that uses this system to analyse modern workplace anonymity:

"[...] tendency to uniformity demands that individuals shall be treated as mere numerical 'units', thus realizing equality by a leveling down, for that is the only direction in which equality can be reached 'in the limit' [...] Anyone who wonders what happens to the individual in such conditions will find that [...] he is so to speak reduced to his substantial aspect, and this amounts to saying that he becomes scarcely more than [...] 'a body without a soul'. From such an individual the qualitative or essential aspect has indeed almost disappeared ('almost', because the limit can never actually be reached); and [...] the individual really no longer has any 'name' that belongs to him, because he is emptied of the qualities which that name should express; he is thus really 'anonymous', but in the inferior sense of the word. This is the anonymity of the 'masses' of which the individual is part and in which he loses himself, those 'masses' that are no more than a collection of similar individuals, regarded purely and simply as so many arithmetical 'units'. 'Units' of that sort can be counted, and the collectivity they make up can thus be numerically evaluated, the result being by definition only a quantity; but in no way can each one of them be given a denomination indicating that he is distinguished from the others by some qualitative difference."

Not big into poetry, but recently I really enjoyed "Song of the Strange Ascetic" by G.K. Chesterton, so will just share it here:

https://www.chesterton.org/the-song-of-the-strange-ascetic/

I think we mostly agree. If you are interested, there are developments that reduce the fluctuations of clock predictions - like principal component based clocks [1], where the authors claim average difference between replicate measurements of 1.5 years.

And the study you are citing (8 to 12 year errors) - seems to report maximum errors, not averages. These might be caused by poor quality samples, so it's unfair to report it like that. Also the publication under discussion (epigenetic age oscillates) would answer where the fluctuations observed by the article you linked to come from - part of them are of course technical due to measurement error, but part of them are influenced by differences in sample collection times.

I also share your skepticism about intervention research and using "biological clocks" to measure how healthy you are. Curiously chronological clocks seem to not be easily affected by interventions. And I would guess this is why most people doing epigenetic aging for sport use biological clocks (like DunedinPACE). But then the question - if you are younger according to "biological age", but we still can measure your true chronological age accurately - are you actually younger in a meaningful sense.

[1]: https://www.nature.com/articles/s43587-022-00248-2

Epigenetic clocks are one of the (if not the) most accurate ways to estimate someone's real chronological age using molecular level data. So not sure what you mean by "extremely unreliable".

And not just this stuff but a lot of health metrics probably have this occult dependence on the time of day. For example MRIs [1].

[1]: https://www.nature.com/articles/s41467-023-42588-6

One of the stranger behaviours for me is that R allows you to combine infix operators with assignments, even thou there are no implemented instances of it in R itself. For example:

  `%in%<-` <- function(x, y, value) { x[x %in% y] <- value; x}

  x <- c("a", "b", "c", "d")
  x %in% c("a", "c") <- "o"
  x
  [1] "o" "b" "o" "d"

Or slightly crazier:

  `<-<-` <- function(x, y, value) paste0(y, "_", value)

  "a" -> x <- "b"
  x
  [1] "a_b"

We with Antoine Fabri created a package that uses this behaviour for some clever replacement operators [1], but beyond that I don't see where this could be useful in real practice.

[1]: https://github.com/moodymudskipper/inops

I agree more with James K on this, but see your point too. Maybe this should be abstracted away such that each person can set their own preference which would be applied across all websites. Similarly how some of them respect your theme (dark vs light), they could be made to respect the animations, so that all toggles behave the same way (fast animation / smooth animation / instant).

I would choose instant thou.

IDK about this. The model y = ax + b is not specified, parameters a and b have to be chosen by optimization. Now add regularization that some of those parameters are shrinked to 0 (lasso) and you have "automatated model specification" where only some parameters are left and others discarded.

And furthermore the models are always chosen from a predefined hypothesis set, so there can never be truly automated specification.