martin biehl

Good question, maybe a) cause we can’t design scientific results to be engaging (more than they are), which is something you can do with some of those things b) to many people science is associated with some kind of establishment -> bad for vibes c) (in)voluntary smugness of scientists…

I thought digitalisation enables error correction and therefore arbitrarily low error communication over arbitrary distances and time intervals. Doesn’t that make it special? Can this be done with analogue methods?

Symmetries at the origin of hierarchical emergence Many systems of interest exhibit nested emergent layers with their own rules and regularities, and our knowledge about them seems naturally organised around these levels. This paper proposes that this...

Preprint time:
“Symmetries at the origin of hierarchical emergence”
arxiv.org/abs/2512.00984

On how symmetries generate hierarchical macroscales and shape the structure of our beliefs, making high-dimensional inference tractable

Are you interested in doing a funded Phd with me on AI in Games at the University of Hertfordshire, UK.

Check out the details below and feel free to share. Deadline for applications is 14th of November.

docs.google.com/document/d/e...

It is an enormous pleasure—and a dream, really—to share that the University of Bamberg has approved an initiative dedicated to mathematical approaches in #ConSci. The initiative is called Bamberg Mathematical Consciousness Science Initiative, or BAMΞ for short.

bamxi.org #BAMΞ @uni-bamberg.de

Yes, that’s a great thing to try and explain if the laws of physics determine everything. But if they don’t what is there to explain? I don’t see it.

What is the appeal of assuming that there are things that aren’t determined by physics as a starting point for any explanation? Wouldn’t assuming that everything is determined by (unknown) laws of physics and trying to explain our subjective reality from there be more interesting?

The four phases of the glider next to each other require 4 x 5 cells horizontally if you count the border a la Randy Beer. But then you still have 4 cells left when 3 would be more symmetric. 7 cells vertically could look ok with centred glider. Maybe snapshots of a moving glider? Wrapping around?

Agreed. I think the reason downward causation is controversial to some is just because within a causally closed dynamical system / the assumptions of reductionism it seems unlikely to exist in any meaningful way.
Outside of reductionism I think it’s bad terminology for some kind of Dennetian stance

We don’t know the true microstate or its true dynamics. It’s just a property of the concept of microstate that it determines the full state of the universe. Maybe sounds useless but allows us to define any closed dynamical system as a toy universe and think about what causation should be within it.

This is true e.g. if importance is measured in predictive power. But I think in the reductionist stance, by assumption, there are no constraints unless they are imposed by the microstate itself. Because there isn’t anything else than the microstate (lowest level) of the one and only true universe

From a quick look it seems they define a weaker notion of downward causation that’s more a form of causal “compatibility”. That’s much less controversial in my opinion. Afaik the controversial version is where the macrostate indeed has causal power beyond the microstate.

For downward causation that isn’t just a kind of stance/compression you need the current macrostate to determine something about the next microstate that the current microstate does not determine. It’s not clear that the example shows this. The microstate of our brains probably contains the rules.

Is it known that negative probabilities are also more memory efficient than using density operators, quantum channels and measurements? Or maybe that doesn’t actually work?

Naively I would think it is irrelevant whether it is generated. If it’s good enough it’s good enough. Maybe there should be standard way to disclose ai generated content though. This would make it clear that the author contributed prompting more than anything else. Not disclosing is like not citing.

Since my Spotify recommendations are mostly boring I started listening to online radio. NTS radio is a good station I think. I listen to them via an app called Mixcloud. You can find the radio station there and then subscribe to its feed and then pick from different djs with genre hints. Not too bad

No clue where this useless link came from … sorry about that.

Don’t you need to assume the existence of a synchronisation map? For a long time it wasn’t mentioned as an assumption. Then it got mentioned, which made sense to me. Is it now gone again?

After a multi-year collaboration with Martin Biehl, @mattecapu.bsky.social and @nathanielvirgo.bsky.social, I’m stoked to share the first of (hopefully) many outputs:
“A Bayesian Interpretation of the Internal Model Principle”
arxiv.org/abs/2503.00511.

1/

Sadly, these methods work against any kind of institution, fascist or not…

Just a minor fact that doesn't seem to be stated in many places - maybe it's too obvious to state:

In the absence of knowledge of what a probability distribution is, people sometimes like to pick the maximum entropy distribution consistent with the knowledge they do have.

More generally you […]