[spuz]

I think the concept would be easier for me to understand if we talked about the inverse of entropy - i.e. some kind of measurement for the concentration of useful energy or "order". I think it would then be more intuitive to say that this measurement always decreases. Do we even have a word for the opposite of entropy?

[MaxRegret]

Negentropy? This is a concept in information theory, but maybe also in physics.

[winwang]

And it's closely related to the Gibbs free energy (available energy), which decreases with increasing entropy, all other things equal.

[bob1029]

> Do we even have a word for the opposite of entropy?

I've always referred to the inverse as "information" or "order".

[cshimmin]

"information" is not a good term for the opposite. Entropy is a well defined concept in information theory (and can be connected to the physical concept). More entropy means more information, not less.

[kgwgk]

> Entropy is a well defined concept in information theory (and can be connected to the physical concept). More entropy means more information, not less.

More entropy means more missing information.

https://iopscience.iop.org/book/mono/978-0-7503-3931-5/chapt...

[kikokikokiko]

As I understand it in information theory, "more entropy" equals "more information is necessary to fully describe this thing", but I may be wrong.

[kgwgk]

“more [additional] information is necessary to fully describe this thing” = “more missing information [in the incomplete description of the thing]”

[bob1029]

I agree. Order is a better term to describe it. Predictability. More entropy = more total possible states the system can be in.

I tend to use "information" to refer to a statistically-significant signal or data that the application/business can practically utilize. This is definitely not the same as the strict information theoretical definition.

[dangitnotagain]

Emergence.

If entropy is the distribution of potential over negative potential, emergence is that where an outcome creates potential (of some discrete domain).

The mystifying relationship with thermodynamic entropy is that through accelerating entropy in other domains (burning), the entropy in the primary domain (eternal drag) is supplanted by the potentials provided by the external domain.

[Timon3]

I might be completely off kilter here (please tell me if that's the case!), but what makes sense to me is to think about "how much entropy is left", as in "how much can entropy still increase between the current state and the highest entropy state". That flips the meaning to describe what you're talking about, and feels very intuitive to me.

[spuz]

When you say "how much entropy is left?" you make it sound like a quantity that is decreasing, not increasing. That seems incorrect, no?

[Timon3]

That's why I specified that it's meant as "how much can entropy still increase between the current state and the highest entropy state". If you can't read it in that sense, either ignore the shorter question or read it as "how much increase of entropy left". Or maybe "how much entropy left until max".

[nobody9999]

>"how much entropy left until max".

10^106 years[0] worth or so.

[0] https://en.wikipedia.org/wiki/Heat_death_of_the_universe

[danq__]

Energy should never be used to introduce entropy as the concept of energy itself is highly, highly complicated and abstract.

Entropy and energy are orthogonal anyway. You can understand entropy without the need to even use the word "energy."

Entropy is an aspect of probability, it is in fact a numerical phenomenon.

[vehicles2b]

I find it intuitive to think of such “ordered” distributions as having a higher compression ratio. (Ie compare the file sizes of zipping a file with just ones or just zeros vs zipping a file with a uniform, random mixture of ones and zeros.)

[Solvency]

Huh? I've never really understood this metaphor.

Take any photograph in Photoshop. First, save one copy of it as a compressed JPG.

Now, on the original, add a small densely repeating tiled pattern multiplied on top as a layer. Like a halftone effect, dot texture, whatever. Technically you're adding more order and less chaos. The resulting image won't compress as efficiently.

[dist-epoch]

The idea is to use the best compressor possible. So called Kolmogorov complexity.

[tnecniv]

Normally we care not about entropy but changes in entropy or entropy measured in comparison to a reference distribution, so you can just take the negative of that difference.

[kledru]

we do have a word -- "negentropy". Physicists also sometimes find it easier to talk about "negentropy" instead of entropy.

[knolan]

Exergy

[syntaxing]

Not trying to be cheeky but wouldn’t the opposite of entropy be work?

[cjs_ac]

Ectropy has been suggested, but the term is not in common use.

[dangitnotagain]

Potential