Assuming your samples are not biased, 1000 subjects are generally far more than are necessary to demonstrate an effect. People who complain about sample size are generally not that well-educated in statistics.

A well-educated person in statistics would also mention that it requires a certain class of distribution. This is one of Nassim Taleb's favourite subject (imagine computing the average net worth of a random group of people and suddenly Bill Gates is among them)

Which would become a biased sample. It’s not as if nobody doing statistical analysis has never seen outliers before.

A sample size of 198 as per this study is more than sufficient to draw pretty strong conclusions.

The issue is not the sample size, it's that studies like these almost always involve a very homogenous population of young college students.

You mean WEIRD.

(Western, Educated, Industrialized, Rich, Democratic)

But why this matters is there a challenge judging intelligence cross cultures?

>But why this matters is there a challenge judging intelligence cross cultures?

I don't know for sure, but my own anecdotal experience is that yes, there most certainly are challenges when a person from one culture assesses the intelligence of someone else from another culture.

It would be nice to know whether this is supported by scientific evidence, or whether this is simply my own personal bias at play.

Also not replicated that I can see.

except they can be

Why on earth? It seems to me that the brain itself is the very disease they're trying to preserve here. Not to mention that, as usual, only certain super-rich people could probably afford it—which, to me, seems like one of the worst and most incurable "diseases" there is. Just my humble opinion, coming from my broken brain ;-)

> NTransformer High-efficiency C++/CUDA LLM inference engine. Runs Llama 70B on a single RTX 3090 (24GB VRAM) by streaming model layers through GPU memory via PCIe, with optional NVMe direct I/O that bypasses the CPU entirely.

untested:

https://github.com/xaskasdf/ntransformer

...sending myself an email

Email as inbox - do you ever actually process it, or does it just pile up with everything else?

if it's worth i extract the core value out of the "conversation" with myself. could be a page in any knowledge base, a snippet in my ide, or whatever suits well... then the email thread goes either into archive or into data nirvanah.

If you are a dark mode addicted like me:

go for NORD theme

https://github.com/EliverLara/Nordic

and I love this icon set (white)

https://www.xfce-look.org/p/1277095

for more NORD integration have a look here:

https://www.nordtheme.com/ports

have fun

I'd choose Zukitre better. No dark theme or a light one blinding your eyes. Pretty neutral, gray.

As for the icon theme, Elementary XFCE works perfectly well with Zukitre. If not, ePapirus or Papirus itself. Simple and flat but contrasted, the opposite to a good chunk of flat themes today, where you can't guess where the buttons start and end.

Once you get used to that theme the Night Mode it's useless as I you can just spawn

     sct 5500  #or xsct

at daytime, or

     sct 3500 

at night time.

xsct/xsct will work with any window manager, too. And the Zukitre themes blend really well with minimal window managers as CWM, i3, DWM and the like, as it has neither curves nor gradients.

UX flaws:

Switching between C and F does not change the label of the field accordingly.

I can enter negative, unrealistic values or no values at all.

Maybe round (floor, ceil) the temperature results in steps of 5 (most devices cannot be adjusted in 1 degree steps)

Improvements: Timer (1st alarm at 1/2 the cooking time. 2nd alarm when done)

Great but I don't want meet myself letting AI write 10k lines of "code" for me

Maybe https://www.tuxedocomputers.com/ is of some interest to you.

Be prepared to pay with your personal data and receive nothing! unless you order a subscription

* "Receive what you already know"

It's a common pattern with me. No matter what test online, I already know the outcome.