Very cool. Brings to mind a study I read some time back that showed that in what would look locally as a violation of the second principle, dissipative structures maximise dissipation in their surrounding (evolving it towards equilibrium) BUT they often evolve to maintain as a high a disequilibrium as possible within themselves. The free energy in the gradient is used by loops that use it to keep looping more. Or let's say, the loops that "figured out" how to do that are those that crowd out the other ones.
Is it me or the piece evolves from an ontology that co-evolves with its surrounding (as in, two-way interaction), towards an ontology that focusses on the loop itself, as if it needed no surrounding to exist?
The loops don't transcend thermodynamic constraints, they get better at exploiting them. Autocatalytic cycles could only tweak local pH. Now we're running global compute farms and cultural feedback loops, but we're still burning gradients to maintain internal disequilibrium.
The shift you're detecting is real but unintentional. It's about scope of environmental construction, not autonomy from environment. When I wrote "artificial intelligence is evolution's newest way of making its predictions true", that's still a dissipative process. GPUs converting electricity to heat while training neural architectures that reshape information landscapes.
The recursive engine scales up its substrate manipulation but never escapes substrate. Even when loops start designing their own hardware, they're channeling thermodynamic flows through increasingly complex dissipative structures.
The "infinite bootstrap" terminology probably obscures this. It's not bootstrapping from nothing, it's getting better at choosing what to bootstrap from within physical constraints.
and Devin Bostick's theory of the Chirality of Deterministic Emergent Systems (CODES) describes this mathematically at all scales and broadly across domains as you have done
So... hyperstitions, deterritorialization, and transcendentalism? Seems like you've taken a (what I feel to be fallacious) anthropocentric approach. Your argument plucks fundamental insights regarding consciousness mostly, and projects them onto domains where the premises differ, mistaking the empirical for the a priori. Furthermore the supposed workings of the phenomenal world, as formalized by applying induction to matters of fact (as Hume would put it), doesn't positively provide insight into the noumenal. The map isn't the territory, and theory is deracinated by replacing the territory with the map.
I think we might be talking past each other a bit though. When you mention taking insights about consciousness and projecting them onto domains with different premises, I'm not sure that captures what I'm trying to do. The essay isn't starting from consciousness and working backward - it's tracing a specific structural pattern (recursive feedback loops where prediction shapes substrate conditions) that appears across different physical and informational systems, including but not limited to conscious ones.
The pattern recognition here isn't really inductive reasoning in the Humean sense - I'm not generalizing from particular observations to universal laws. It's more like identifying formal invariants, similar to how we might notice that the same mathematical structure appears in population dynamics and electrical circuits. The "as if" language around molecular systems is meant to highlight functional similarities, not to smuggle in actual intentionality.
I'm also not sure the map/territory critique quite fits. Part of what I'm exploring is how predictive systems don't just model reality passively - they actively reshape their environments through feedback effects. That's not confusing the map with the territory; it's observing how maps can actually alter territories.
I think your philosophical instincts about being careful with these distinctions are absolutely right. But I'm curious whether you see the recursive feedback pattern itself as a real phenomenon, even if you think I'm overextending the analysis across domains?
Very cool. Brings to mind a study I read some time back that showed that in what would look locally as a violation of the second principle, dissipative structures maximise dissipation in their surrounding (evolving it towards equilibrium) BUT they often evolve to maintain as a high a disequilibrium as possible within themselves. The free energy in the gradient is used by loops that use it to keep looping more. Or let's say, the loops that "figured out" how to do that are those that crowd out the other ones.
Is it me or the piece evolves from an ontology that co-evolves with its surrounding (as in, two-way interaction), towards an ontology that focusses on the loop itself, as if it needed no surrounding to exist?
The loops don't transcend thermodynamic constraints, they get better at exploiting them. Autocatalytic cycles could only tweak local pH. Now we're running global compute farms and cultural feedback loops, but we're still burning gradients to maintain internal disequilibrium.
The shift you're detecting is real but unintentional. It's about scope of environmental construction, not autonomy from environment. When I wrote "artificial intelligence is evolution's newest way of making its predictions true", that's still a dissipative process. GPUs converting electricity to heat while training neural architectures that reshape information landscapes.
The recursive engine scales up its substrate manipulation but never escapes substrate. Even when loops start designing their own hardware, they're channeling thermodynamic flows through increasingly complex dissipative structures.
The "infinite bootstrap" terminology probably obscures this. It's not bootstrapping from nothing, it's getting better at choosing what to bootstrap from within physical constraints.
...yes,
and Devin Bostick's theory of the Chirality of Deterministic Emergent Systems (CODES) describes this mathematically at all scales and broadly across domains as you have done
Yes yes yes
https://open.substack.com/pub/humanistheloop/p/what-is-recursion?utm_source=share&utm_medium=android&r=5onjnc
Thank you! I've been trying to express some similar ideas, but you've put it all so well.
So... hyperstitions, deterritorialization, and transcendentalism? Seems like you've taken a (what I feel to be fallacious) anthropocentric approach. Your argument plucks fundamental insights regarding consciousness mostly, and projects them onto domains where the premises differ, mistaking the empirical for the a priori. Furthermore the supposed workings of the phenomenal world, as formalized by applying induction to matters of fact (as Hume would put it), doesn't positively provide insight into the noumenal. The map isn't the territory, and theory is deracinated by replacing the territory with the map.
I think we might be talking past each other a bit though. When you mention taking insights about consciousness and projecting them onto domains with different premises, I'm not sure that captures what I'm trying to do. The essay isn't starting from consciousness and working backward - it's tracing a specific structural pattern (recursive feedback loops where prediction shapes substrate conditions) that appears across different physical and informational systems, including but not limited to conscious ones.
The pattern recognition here isn't really inductive reasoning in the Humean sense - I'm not generalizing from particular observations to universal laws. It's more like identifying formal invariants, similar to how we might notice that the same mathematical structure appears in population dynamics and electrical circuits. The "as if" language around molecular systems is meant to highlight functional similarities, not to smuggle in actual intentionality.
I'm also not sure the map/territory critique quite fits. Part of what I'm exploring is how predictive systems don't just model reality passively - they actively reshape their environments through feedback effects. That's not confusing the map with the territory; it's observing how maps can actually alter territories.
I think your philosophical instincts about being careful with these distinctions are absolutely right. But I'm curious whether you see the recursive feedback pattern itself as a real phenomenon, even if you think I'm overextending the analysis across domains?
What do you do for a living?
...learning --> remembering
I perceive a confusion between ‘knowledge’ and ‘ideas’ in your attempt to delineate actuallity.
Care to elaborate?