The Fed says this is a cube of $1M. They're off by half a million

I think it’s silly that English has these quirks, and it’s silly to engage with them as points of argument, which isn’t what I mean to do, but rather to show how my own thought process works, silly it may be. It’s okay to embrace silliness in the environment as long as it isn’t detracting from understanding. This thread is exploring the words, not arguing with each other or trying to convince the other, so it’s not at cross purposes to me. But I think I agree that there may be no disagreement?

Let’s not let the perfect be the enemy of the silly. ;)

https://en.wikipedia.org/wiki/Rai_stones

> The ownership of a large stone, which would be too difficult to move, was established by its history as recorded in oral tradition rather than by its location. Appending a transfer to the oral history of the stone thus effected a change of ownership.

> Some modern economists have viewed Rai stones as a form of money, and the stones are often used as a demonstration of the fact that the value of some forms of money can be assigned purely through a shared belief in said value.

...

> Rai stones were, and still are, used in rare important social transactions, such as marriage, inheritance, political deals, sign of an alliance, ransom of the battle dead, or, rarely, in exchange for food. Many are placed in front of meetinghouses, around village courts, or along pathways.

> Although the ownership of a particular stone might change, the stone itself is rarely moved due to its weight and risk of damage. Thus the physical location of a stone was often not significant: ownership was established by shared agreement and could be transferred even without physical access to the stone. Each large stone had an oral history that included the names of previous owners. In one instance, a large rai being transported by canoe and outrigger was accidentally dropped and sank to the sea floor. Although it was never seen again, everyone agreed that the rai must still be there, so it continued to be transacted as any other stone.

I discovered this when I had to test accuracy in a pose estimation model for a computer vision project involving crowds being counted by CCTV. Turns out in low-res imaging (even the best CCTV is a bit rubbish at long range due to wide angle lenses and so on), pose estimation models beat almost everything else for counting. Except... they're still not accurate.

I would spend hours going through hundreds of frames, counting people by hand, and then I'd compare with different pose models, and found they were always out, but they were out by the same amount.

Some models got confused by reflective surfaces, so would double count. Others would be out by ~20% every time for various environmental factors. The good thing about this is you could then easily show calibration. "OK, in that space, whatever the model says, half it, that's a better count", or "OK, in this image from this camera whatever the number the model says, increase/decrease it by 20%, and that'll be more accurate".

We ended up through human calibration being able to provide much more accurate "models" per camera than the models themselves.

It did mean counting hundreds of people per frame for hundreds of frames for dozens of cameras though... I had to abandon the trackpad for a mouse and just got settled in for a few days with DotDotGoose. Colleagues were surprised I had the patience for it. :-)

https://biodiversityinformatics.amnh.org/open_source/dotdotg...

I think the main thing I was responding to was this --

> That’s good that it’s unambiguous to you, as you happen to be correctly interpreting the meaning from the words as written, but I don’t read the context the same way, as in, your reading doesn’t always read as written, when I’m doing the reading.

-- which I (perhaps wrongly) took to be arguing against, or slightly misunderstanding, the claim of the person you were responding to. I don't think they were claiming that the words as written are inherently unambiguous, and I don't think it's a question of reading the context; I think it's just an idiomatic phrase that has a fixed meaning for those who natively use it. It's a bit like a dialect word; it's only ambiguous in the sense that people who don't speak the dialect won't know how to interpret it.

(It could turn out that I'm factually wrong about this, and that there are different groups who use the phrase in mutually contradictory ways! But so far I've only seen a split between groups who use it to mean "the Friday after this coming Friday" and groups who don't use it at all.)