I’m not sure whether Shannon’s information and Algorithmic information measure the same thing. You mean that they would be equivalent?

I know they are not. Yes, because Algorithmic information is about single objects.

Ah! Oh yes, you’re right. Shannon’s information is about events. Or rather, about occurrences of repeated events.

That’s why Shannon relies on the probability of events. And probabilities do not always apply. What’s the probability of π?

And yet, when probabilities make sense, both notions are equivalent. I’m not even sure of that.

Why? Because complexity is a minimum over all computations. And it may change through time.

You mean, it may depend on the context? Yes. A second occurrence of an event just requires a mere copy. It’s simple.

And yet, this second occurrence might be surprising if the event is rare. So it should carry much information. Yes. It was not totally clear to me. So I asked the teachers.

And...? I’ve been told that all this would become clear when we define algorithmic probability in Chapter 3.

That’s funny. Me, I’ve been told that I would understand all this in Chapter 5, when we talk about subjective information. Ok, let’s wait and see...