In some contexts, a bit can refer to a boolean variable, a flag. In other contexts, it may refer to the voltage at a certain point, or any number of other things. But when you are talking about bits/s then it's a measure of information.

These are not the same thing because the amount of information contained in a bit is not always equal to one bit of entropy.

Yes, but as you know, this implies that the information is already available. You can use that knowledge to create a compression algorithm, or to define a less redundant file format. That's very practical.

We can also be a bit philosophical and ask: How much information does a backup contain? The answer could be: By definition, 0 bits. That's not a useful answer, which implies a problem with the application of the definition.

A more interesting question might be: How much information does a file contain, that stores the first 1 million digits of the number π?

If they had heard of it, we'd probably get statements like: "It's just statistics." or "It's not information. It's just a probability."

It's a bit like asking what is the difference between the letter "A" and ink on a page in the shape of the letter "A". Of course, first one would have to explain how they are usually not different at all.

BTW, I don't know what you mean by "external storage data". The expression doesn't make sense.

Wrong. They are called the same because they are fundamentally the same. That's how you measure information.

In some contexts, one wants to make a difference between the theoretical information content and what is actually stored on a technical device. But that's a fairly subtle thing.

Here's a link to Caltech's press release: https://www.caltech.edu/about/news/thinking-slowly-the-paradoxical-slowness-of-human-behavior

Here's a link to the actual paper (paywall): https://www.cell.com/neuron/abstract/S0896-6273(24)00808-0

Here's a link to a preprint: https://arxiv.org/abs/2408.10234

Someone's tooting their own horn.

I checked on google maps and it wouldn't be a problem. About 1km away, there's a recycling station and a cemetery with lots of room.

KI-Gesetz/AI Act: http://data.europa.eu/eli/reg/2024/1689/oj

Die Kennzeichnungspflicht gilt erst ab August 2026. Sie gilt auch nur für Anbieter/Nutzer, wenn ich richtig verstehe. Die Inhalte zu verbreiten ist also legal, auch wenn man die Kennzeichnung entfernt. Das heißt, außer Satirikern und Memern werden nur manche Open-Source-Sachen ein Problem bekommen.

Schwer zu sagen. Traditionell ist Computer Vision ein Teilbereich von KI, was einen großen Teil der "normalen" Bildbearbeitung einschließt. Die Definition im Gesetz ist so gehalten, dass das alles abgedeckt wird. Aber das Gesetz wird natürlich von eher wenig kundigen Leuten ausgelegt, sodass man sich nicht darauf verlassen kann, dass Begriffe technisch korrekt verstanden werden.

Ich hab noch mal nachgeschaut. "Diese Pflicht gilt nicht, soweit die KI-Systeme eine unterstützende Funktion für die Standardbearbeitung ausführen[.]" Das automatische Ausschneiden könnte also unproblematisch sein, allerdings ist so eine satirische Bearbeitung kein Standard. Vielleicht meint "Standardbearbeitungen" nur solche, die von den KI-Systemen in Kameras bei jeder normalen Aufnahme gemacht werden.

Dürfte ein Rohrkrepierer werden, aber das ist wahrscheinlich auch das Beste.

Seriously, why are they producing all these great promo shots for him? If they had done that for someone like Bernie Sanders, the US would be in much better shape now.

(explaining the title reference)

The only reason you need to know someone's name is to talk about them behind their back.