Phil
@philphi.bsky.social
Fisher Curvature, Explainable AI, Evolutionary AI, PHILosophy. "Philo" φίλος which means "loving" or "friend". D[R S] ≠ 0
I fixed my background to analyze background independent dynamics. An uncomfortable subsystem, proceed with caution or precautionary principle.
December 15, 2025 at 8:11 AM
I fixed my background to analyze background independent dynamics. An uncomfortable subsystem, proceed with caution or precautionary principle.
A way to practically understand this is to engage in behavior where quantum dominates (intricate dynamics) then engage in higher level events (causal networks observation). A Fisher Computer can observe and then predict records of the projection with parameter estimation. Quantum to Analogto Digital
December 14, 2025 at 10:42 PM
A way to practically understand this is to engage in behavior where quantum dominates (intricate dynamics) then engage in higher level events (causal networks observation). A Fisher Computer can observe and then predict records of the projection with parameter estimation. Quantum to Analogto Digital
Its metric-free descendant downstairs is a graph/causal-network MFI defined only from order/links: I_graph[ρ] = Σ_(i→j in L) ( (ρⱼ − ρᵢ)² / ρᵢ ) (up to sym/weights), where L is the set of causal links (cover relations). Uses only adjacency in the poset/DAG plus counting measure, not spacetime metric
December 14, 2025 at 8:37 PM
Its metric-free descendant downstairs is a graph/causal-network MFI defined only from order/links: I_graph[ρ] = Σ_(i→j in L) ( (ρⱼ − ρᵢ)² / ρᵢ ) (up to sym/weights), where L is the set of causal links (cover relations). Uses only adjacency in the poset/DAG plus counting measure, not spacetime metric
Fisher–Rao m on a statistical model pᵢ(θ) over events (or link-patterns) is gᴲ_ab(θ) = Σᵢ pᵢ(θ) [∂_a ln pᵢ(θ)] [∂_b ln pᵢ(θ)].
Discard spacetime distances but retain a Fisher geometry on probability distributions over events/causal relations. If “I” is written upstairs as a continuum MFI functional,
Discard spacetime distances but retain a Fisher geometry on probability distributions over events/causal relations. If “I” is written upstairs as a continuum MFI functional,
December 14, 2025 at 8:36 PM
Fisher–Rao m on a statistical model pᵢ(θ) over events (or link-patterns) is gᴲ_ab(θ) = Σᵢ pᵢ(θ) [∂_a ln pᵢ(θ)] [∂_b ln pᵢ(θ)].
Discard spacetime distances but retain a Fisher geometry on probability distributions over events/causal relations. If “I” is written upstairs as a continuum MFI functional,
Discard spacetime distances but retain a Fisher geometry on probability distributions over events/causal relations. If “I” is written upstairs as a continuum MFI functional,
In regimes where the spacetime description is valid, this order agrees with the continuum causal future: eᵢ ≺ eⱼ ⇔ eⱼ ∈ J⁺(eᵢ).
Define links (adjacency) without metric data by the cover relation: eᵢ ≺· eⱼ ⇔ (eᵢ ≺ eⱼ) and there is no eₖ such that eᵢ ≺ eₖ ≺ eⱼ. ST metric discarded, info geo can remain
Define links (adjacency) without metric data by the cover relation: eᵢ ≺· eⱼ ⇔ (eᵢ ≺ eⱼ) and there is no eₖ such that eᵢ ≺ eₖ ≺ eⱼ. ST metric discarded, info geo can remain
December 14, 2025 at 8:33 PM
In regimes where the spacetime description is valid, this order agrees with the continuum causal future: eᵢ ≺ eⱼ ⇔ eⱼ ∈ J⁺(eᵢ).
Define links (adjacency) without metric data by the cover relation: eᵢ ≺· eⱼ ⇔ (eᵢ ≺ eⱼ) and there is no eₖ such that eᵢ ≺ eₖ ≺ eⱼ. ST metric discarded, info geo can remain
Define links (adjacency) without metric data by the cover relation: eᵢ ≺· eⱼ ⇔ (eᵢ ≺ eⱼ) and there is no eₖ such that eᵢ ≺ eₖ ≺ eⱼ. ST metric discarded, info geo can remain
Let E={eᵢ} be the sampled events produced by a discrete sampling kernel at resolution ℓ. Define the causal precedence relation directly on E: eᵢ ≺ eⱼ means “eᵢ can causally influence eⱼ” (a partial order on events). In regimes where the spacetime description is valid. 3D recovered from the CN order.
December 14, 2025 at 8:32 PM
Let E={eᵢ} be the sampled events produced by a discrete sampling kernel at resolution ℓ. Define the causal precedence relation directly on E: eᵢ ≺ eⱼ means “eᵢ can causally influence eⱼ” (a partial order on events). In regimes where the spacetime description is valid. 3D recovered from the CN order.
Spacetime→Causal network is not a move to something more fundamental (as in CST). It is a further lossy compression of the same emergent spacetime, useful for macroscopic causality. This projection forgets geometric distance (no g_μν) while retaining causal structure as a locally finite event order.
December 14, 2025 at 8:29 PM
Spacetime→Causal network is not a move to something more fundamental (as in CST). It is a further lossy compression of the same emergent spacetime, useful for macroscopic causality. This projection forgets geometric distance (no g_μν) while retaining causal structure as a locally finite event order.
eᵢ ≺ eⱼ ⇔ eⱼ ∈ J⁺(eᵢ); eᵢ ≺· eⱼ ⇔ (eᵢ ≺ eⱼ) and there is no eₖ such that eᵢ ≺ eₖ ≺ eⱼ. gᴲ_ab(θ) = Σᵢ pᵢ(θ) [∂_a ln pᵢ(θ)] [∂_b ln pᵢ(θ)]. Discard geometric distances in spacetime, but retain a Fisher geometry on probability distributions over events /causal relations. “I” is the continuum functional
December 14, 2025 at 7:21 PM
eᵢ ≺ eⱼ ⇔ eⱼ ∈ J⁺(eᵢ); eᵢ ≺· eⱼ ⇔ (eᵢ ≺ eⱼ) and there is no eₖ such that eᵢ ≺ eₖ ≺ eⱼ. gᴲ_ab(θ) = Σᵢ pᵢ(θ) [∂_a ln pᵢ(θ)] [∂_b ln pᵢ(θ)]. Discard geometric distances in spacetime, but retain a Fisher geometry on probability distributions over events /causal relations. “I” is the continuum functional
Causal discrete spacetime the next projection, I can see it
December 14, 2025 at 6:49 PM
Causal discrete spacetime the next projection, I can see it
that's what the simulation wants u to believe, ha jk
November 11, 2025 at 4:12 PM
that's what the simulation wants u to believe, ha jk
More and more evidence (like DESI) that there really is evolving Dark Energy, maybe a tachyon condensate.
Random Insight: Today's Laplace's Demon is basically Simulation Theory (or religion), but there were a lot less claims of Laplace's Demon being real. Don't need either for determinism.
Random Insight: Today's Laplace's Demon is basically Simulation Theory (or religion), but there were a lot less claims of Laplace's Demon being real. Don't need either for determinism.
November 10, 2025 at 5:30 PM
More and more evidence (like DESI) that there really is evolving Dark Energy, maybe a tachyon condensate.
Random Insight: Today's Laplace's Demon is basically Simulation Theory (or religion), but there were a lot less claims of Laplace's Demon being real. Don't need either for determinism.
Random Insight: Today's Laplace's Demon is basically Simulation Theory (or religion), but there were a lot less claims of Laplace's Demon being real. Don't need either for determinism.
Boundary conditions, like math, as a tool, but not fundamental. AI knows math well already, but what if we (and/or AI) figure out another way to understand physics?
November 9, 2025 at 5:32 PM
Boundary conditions, like math, as a tool, but not fundamental. AI knows math well already, but what if we (and/or AI) figure out another way to understand physics?
bad/useless AI is annoying, but is not liking better AI just dumb? if there is a foreseeable route to create something better, do u partake? is smart reproduction creating something better than yourself, a better child with a better life than your own? when does the lines between human and AI blur?
November 8, 2025 at 5:52 PM
bad/useless AI is annoying, but is not liking better AI just dumb? if there is a foreseeable route to create something better, do u partake? is smart reproduction creating something better than yourself, a better child with a better life than your own? when does the lines between human and AI blur?
Or a mix. CJ+ Anisotropic (rank one disformal): G=Ω(R)^2 g + Θ(R) outer(∇lnR,∇lnR).
Same L-B: Q=−(ħ^2/2)(Δ_G R)/R + α ℛ(G).
Directional response without Fisher still has Distinguishability: R sets the relational metric tensor; its curvature guides motion.
Θ→0: conformal CJ; push-forward via K ⇒ EH
Same L-B: Q=−(ħ^2/2)(Δ_G R)/R + α ℛ(G).
Directional response without Fisher still has Distinguishability: R sets the relational metric tensor; its curvature guides motion.
Θ→0: conformal CJ; push-forward via K ⇒ EH
November 6, 2025 at 5:29 AM
Or a mix. CJ+ Anisotropic (rank one disformal): G=Ω(R)^2 g + Θ(R) outer(∇lnR,∇lnR).
Same L-B: Q=−(ħ^2/2)(Δ_G R)/R + α ℛ(G).
Directional response without Fisher still has Distinguishability: R sets the relational metric tensor; its curvature guides motion.
Θ→0: conformal CJ; push-forward via K ⇒ EH
Same L-B: Q=−(ħ^2/2)(Δ_G R)/R + α ℛ(G).
Directional response without Fisher still has Distinguishability: R sets the relational metric tensor; its curvature guides motion.
Θ→0: conformal CJ; push-forward via K ⇒ EH
Same LB and Unified. Conformal-Jacobi: set G = Ω(R)² g. HJ+continuity (equivariance) hold with Q_CJ = −(ℏ²/2)(Δ_G R)/R + α ℛ(G). Push-forward of α ℛ(G) via K ⇒ 4D Einstein–Hilbert term (G_eff⁻¹ ∝ α⟨Ω⁻²⟩). DGZ limit: Ω→1, α→0 (standard BM on shape space). Curvature piece αR(G) directly to an EH term.
November 6, 2025 at 4:47 AM
Same LB and Unified. Conformal-Jacobi: set G = Ω(R)² g. HJ+continuity (equivariance) hold with Q_CJ = −(ℏ²/2)(Δ_G R)/R + α ℛ(G). Push-forward of α ℛ(G) via K ⇒ 4D Einstein–Hilbert term (G_eff⁻¹ ∝ α⟨Ω⁻²⟩). DGZ limit: Ω→1, α→0 (standard BM on shape space). Curvature piece αR(G) directly to an EH term.
FisherPHILic: Born amplitude supplies measure; the Fisher–Rao construction is the unique, coarse-graining-monotone way to promote that measure to a metric so that the QP becomes the Laplace–Beltrami curvature of distinguishability. The state-dependent geometry a fixed shape metric (DGZ) cant provide
November 6, 2025 at 2:30 AM
FisherPHILic: Born amplitude supplies measure; the Fisher–Rao construction is the unique, coarse-graining-monotone way to promote that measure to a metric so that the QP becomes the Laplace–Beltrami curvature of distinguishability. The state-dependent geometry a fixed shape metric (DGZ) cant provide
N=2 has one extra rotational symmetry about the inter-particle axis, so the reduction removes only two rotational DOF, leaving a 1-D ray. The space is stratified, rotations about the line joining the points act trivially (an SO(2) stabilizer), so the reduced space is one-dimensional, the ray r>0.
November 3, 2025 at 8:14 PM
N=2 has one extra rotational symmetry about the inter-particle axis, so the reduction removes only two rotational DOF, leaving a 1-D ray. The space is stratified, rotations about the line joining the points act trivially (an SO(2) stabilizer), so the reduced space is one-dimensional, the ray r>0.
Barbour has an intrinsic scale in shapes that are a closed subsystem. When you reason that relational physics should remain consistent when subsystems interact, keeping scale (3N–6) is the cleaner, more operationally relational choice. Scale is what sets the Machian clock; distinction is relational.
November 2, 2025 at 4:41 PM
Barbour has an intrinsic scale in shapes that are a closed subsystem. When you reason that relational physics should remain consistent when subsystems interact, keeping scale (3N–6) is the cleaner, more operationally relational choice. Scale is what sets the Machian clock; distinction is relational.
Uniform treatment of all 𝑁. The same projector Π and the same Jacobi construction handle 𝑁=2 (one DOF), generic N≥3 (dimension 3𝑁−6) and special stabilizer strata, all from one rule. You don't need 3 points: two points suffice, a fundamental ontology of distinctions and dynamics move along geodesics
November 1, 2025 at 6:07 PM
Uniform treatment of all 𝑁. The same projector Π and the same Jacobi construction handle 𝑁=2 (one DOF), generic N≥3 (dimension 3𝑁−6) and special stabilizer strata, all from one rule. You don't need 3 points: two points suffice, a fundamental ontology of distinctions and dynamics move along geodesics
Jacobi: Time from change. No absolute time variable. The ephemeris 𝑡 is distilled from the path length. This is a Machian lean ontology. Rods and clocks are emergent. With scale kept, the metric itself furnishes the standards of length (arc-length) and time. You do not need to posit extra standards.
November 1, 2025 at 5:48 PM
Jacobi: Time from change. No absolute time variable. The ephemeris 𝑡 is distilled from the path length. This is a Machian lean ontology. Rods and clocks are emergent. With scale kept, the metric itself furnishes the standards of length (arc-length) and time. You do not need to posit extra standards.
Law = geometry. No added forces or extra structures: the curvature of Fisher geometry (G) and the scalar (V) fully determine the connection of the Jacobi metric and hence the motion. That’s fewer ontic kinds than “space + time + masses + forces”; sPNP is just configuration-space geometry + a scalar.
November 1, 2025 at 5:45 PM
Law = geometry. No added forces or extra structures: the curvature of Fisher geometry (G) and the scalar (V) fully determine the connection of the Jacobi metric and hence the motion. That’s fewer ontic kinds than “space + time + masses + forces”; sPNP is just configuration-space geometry + a scalar.