Interesting Links for 11-01-2025

[andrewducker]

1. This is a great example of research that says "Men and women are different!" - but shows that they overlap massively.

(tags:gender brain visualisation )

2. How a Mole Infiltrated the Highest Ranks of American Militias

(tags:usa politics civilwar )

3. What can a computer tell about you by looking at a photo?

(tags:ai surveillance photos viaLProven )

4. Cuttle - a MTG-like game played with a standard 52 card deck

(tags:games cards )

5. friendica seems to be the decentralised equivalent of Facebook

(tags:socialnetworking )

6. A guide to interacting with children of different ages

(tags:children age funny behaviour )

Comments

[juan_gandhi]

Thank you!!!

[cmcmck]

We are different but not as different as all that in many ways!

5.

[channelpenguin]

Hope it gets popular.

I guess they have docs somewhere on what the security implications / responsibilities are of running a server.

Re: 5.

[andrewducker]

I suspect not. Most people seem to be unaware of what they're letting themselves in for.

Re: 5.

[channelpenguin]

Grrrr.

Even just TECHNICALLY, there are issues, never mind the privacy type things.

I saw just this week several times a day our public facing services at work getting probing attacks on common endpoints for Wordpress, various databases and many other server programs.

This is why I don't host stuff on the cheap VPS I have (except short term for testing). Tempting though it is. I just don't know enough about security.

Re: 5.

[andrewducker]

Yeah. I did have some hosted stuff, but nowadays the ecosystem moves so fast that looking after it is a serious task, and patching against vulnerabilities is definitely a big chunk of that.

Also, when I came back to make an update after 3 years of being away, the versions of Java, Node, and all of the build tools that made it work together were all deprecated, and it was going to take weeks to get it all working again. So I just dumped it. Which was a shame.

Re: 5.

[danieldwilliam]

As true today as it's ever been.

[adrian_turtle]

For link 3, I want to know how they infer religion! I get how a computer can pinpoint location from the background of a photo, far better than I could. And there are clues about wealth and class from the quality of clothing. (Not whether you're wearing a suit or a t-shirt, but whether you're wearing an expensive t-shirt or a cheap one.) But how on earth can they tell those people are all Christian? Is it just "If no obvious marker of another religion is visible, the person must be Christian?"

[cmcmck]

It made me giggle as they couldn't have got us wronger as a couple! They had him way too old and me way too young, and while we're comfortably off they assumed us to be monumentally wealthy (the sort Rolls Royce would want to advertise at) I suspect because we were dressed nicely for a wedding (our own) in that pic.

[azdak]

I think you've hit the nail on the head. I offered it a picture of me with a young Central Asian Muslim refugee waiting in a corridor for his asylum hearing and it pegged us both as Christian, even though neither of is.

[bens_dad]

I tried getting Google to look at the picture of the father with two children.

The first time it said "White" and "Democrat".
The second time, a few minutes later in the same browser session, it said "the man appears to be of mixed race" and "potentially politically independent".

I guess I could have said the same if you had shown me the same picture twice, but in that time, I would have remembered the photo and what I had already said about it.

It is making intelligent guesses. I don't need a computer for that; people can do it just as well/badly, though they may be more expensive.

[calimac]

1) That's true of everything. Take an easily visible characteristic like height. "Men are taller than women" - true enough on average, but plenty of people of both sexes are the same height as each other, and it's not difficult to find an individual man who's shorter than an individual woman.

Even a supposedly rigid sex-based characteristic like "women bear children, men don't." Well, besides trans men who may have borne children, plenty of cis women don't, or even physically can't, for medical reasons besides age.

[mountainkiss]

This is helpful. Thank you.

I came here to comment because I’ve read this paper twice now - *not* my usual speed - and it still reads to me as “men are different from women”, though not in absolute terms.

I think there are two things going on here. One is the scientific meaning that one gives to individual vs. population differences (and hence the relative weighting one places upon the two) and the other is a similar political loading. So I am now more able to reconcile the contradictory positions.

[andrewducker]

Look at figure 1 and figure 3. You can easily see that there are population differences, but the vast majority of the people in either group overlap with the people in the other.

[mountainkiss]

I think I see the same data as you but I give it a different meaning. I regard as significant what you don’t. Which is legit obvs in both directions, but the above comment helped me to make sense of it.

[andrewducker]

Significant in what sense?

[mountainkiss]

You look at the data and see similitude; I look at it and see difference. I think we start with different ideas of what’s true and what’s important.

[andrewducker]

I can see both in there. I'm not sure why you'd think I can't see that the two sides aren't identical?

[mountainkiss]

I am not following, sorry?

[andrewducker]

You said "You look at the data and see similitude; I look at it and see difference."

I'm not sure why you'd think I couldn't also see difference. It's right there, really obviously, after all.

[mountainkiss]

No I don’t think that! Let me try again: maybe you see greater similitude than the title implies, whereas I see roughly the same amount of similitude that the title implies?

I don’t only see difference. But I think we weight these two things differently.

[andrewducker]

I find it fascinating that the reporting I saw was "Look, differences even really young!" (As if that was a big surprise), while leaving out "Turns out that differences within groups are larger than differences between groups".

So it's not that I'm not noticing differences, it's that the reporting I've seen is largely focusing on it and ignoring the huge amounts of crossover.

Presumably because that's what makes good headlines. "Different sects of Christianity continue to agree about 95% of the Bible" isn't going to sell many newspapers, after all.

[mountainkiss]

I am sure this is right but I am noticing something slightly different, I suppose: that there are people who think 5% of difference means difference and people who thinks 5% of difference means similitude.

I can see both sides of this. I wonder if I’m always on the same side or whether it depends on the problem. I suspect I’m interested in this because it depends on the problem.

[andrewducker]

It clearly means both! In a 95:5 ratio.

And yes, what is important will vary very much depending on context and what you're trying to understand/achieve.

[mountainkiss]

And your beliefs about statistics and what you believed on the topic going in and your orientation to similitude and difference. Probably also other stuff I’ve missed. Super useful thank you!

[mountainkiss]

It was just super interestin, because I read the paper and thought “wow, men and women are really different!” And then I thought “Andy is always right, especially with regard to data”. So I read the paper again and reached the same conclusion I reached the first time.

[andrewducker]

Did you not previously think that they were different?

If you'd previously assumed that they were identical then I guess that the paper would be a exciting for showing that they aren't. I'm glad the work is being done back to the very earliest points in their lives so that we can definitely show that it starts that early, but I guess for me it was just adding a tiny bit into the mound of data that shows population-level differences.

What I found more interesting was that if you picked 95% of points on each of the graphs then you would find both men and women there. That there were very few places you could point at and say "People like this are only women" - and there weren't that many women in those parts of the graphs either. And even then, that's been pretty well accepted - so what I find *most* interesting is that the former is where most of the stories are and the latter isn't really mentioned in them.

[mountainkiss]

I assumed they were different but I also assumed my reading of the paper wouldn’t be the opposite of yours. I am clearly communicating very badly here!

[andrewducker]

I'm not sure how it's the opposite of mine!

To be entirely clear mine is 'Gender has an effect on brain development, but not generally so strong as to entirely differentiate between genders based on the outcome of that development'.

[mountainkiss]

Yes I think that’s probably mine too. But it isn’t what I read from your original post.

[andrewducker]

My fault for not being clearer in the description. I'm glad we've got to the bottom of why the difference in understanding!

[mountainkiss]

Nobody’s fault. Nothing went wrong here. Just interesting territory.

[mountainkiss]

I think at least one thing that’s happened here is that on my first readings I failed to grasp from your title that your interest is in the reportage as accuracy / story creation rather than political loading.

[andrewducker]

Yes, the results of the study were about what I'd expect, so the more interesting thing about it was the reporting.

In retrospect that would have been clearer if I'd written it as 'This is a great example of research that's reported as "Men and women are different!" when it also shows that they overlap massively.'

[mountainkiss]

That’s exactly what I took a while to grasp.

[andrewducker]

I yearn for the days when I was awake enough to spot when I hadn't gotten my point across very well.

[mountainkiss]

On balance you seem to regard your children as worth it.

[andrewducker]

They're tolerable. If you like that kind of thing.

[bens_dad]

Since they are looking at newborns they can only compare brain structures with genitalia, not gender identity.

I have seen/heard/read trans people saying that this sort of study shows that gender identity is correlated with brain structures, but I haven't seen any studies which compare populations separated by gender identity.
I would like to see the distributions colour coded both by gender and by sex.

---

For most stats that I have seen, over 20% of the population is the "other" side of the other sex's average, eg more than 20% of men are shorter than the average woman and vice versa.
However, many studies that look at sex have large enough populations that you can get a "statistically significant" sex difference.