Urgent.News

What's breaking now, across thousands of outlets.

AI

Claude Fable 5.2 Solves Black‑Hole Paradoxes – The AI That Just Changed Physics

Claude Fable 5.2 Cracks Two Black‑Hole Mysteries – What It Means for Physics and the Real World Lead On 19 Nov 2026 Anthropic unveiled Claude Fable 5.2 , an LLM that generated the first AI‑driven, peer‑review‑ready solutions to two long‑standing open problems in black‑hole physics. Within hours, the announcement topped Hacker News, earned 252 up‑votes, and appeared in Google News feeds worldwide.…

Claude Fable 5.2, developed by Anthropic, represents a significant milestone in artificial intelligence, as it has successfully solved two long-standing problems in black-hole physics. This development has garnered considerable attention within the scientific community and the general public, with the announcement quickly gaining traction on platforms like Hacker News and Google News.

The two problems tackled by Claude's Large Language Model (LLM) are the information-loss paradox for evaporating micro-black holes and the existence of a regular, rotating semiclassical solution that adheres to the weak cosmic-censorship conjecture. These conundrums have puzzled theorists for over five decades. Remarkably, Claude's solutions to these problems are presented as fully analytic constructions, which have been independently verified by experts at prestigious institutions such as the Institute for Advanced Study and the Max-Planck Institute for Gravitational Physics.

To provide context, consider a graduate student named Maya working on a complex problem involving the semiclassical Einstein equations with trace-anomaly corrections. This task typically requires extensive computational resources and time, as it involves solving a system of coupled ordinary differential equations (ODEs) that describe an axisymmetric metric ansatz.

Maya's primary challenge is to maintain regularity at the horizon and ensure asymptotic flatness at infinity while adhering to the weak cosmic-censorship conjecture (WCCC).

Enter Claude Fable 5.2. Upon receiving Maya's prompt—"Derive an axisymmetric solution to the semiclassical Einstein equations including the type-A trace anomaly. Enforce regularity at the horizon and respect WCCC."—the LLM swiftly generates a symbolic expression for the metric functions (f(r,θ)) and (g(r,θ)) in closed form. It also provides a step-by-step derivation, along with a custom Python routine that combines gradient-based optimization with adaptive step sizing for "AI-accelerated shooting methods."

This allows Maya to integrate outward from the horizon and obtain a convergent solution within seconds.

Maya then verifies Claude's generated code in her Mathematica notebook, reproducing the analytic metric and cross-checking it against two independent numerical relativity codes. This process demonstrates how Claude Fable 5.2 can serve as a valuable research assistant, offering creative insights and computational shortcuts that would otherwise consume significant time and effort for a human researcher.

However, it is essential to note that human experts still play a crucial role in performing final sanity checks, interpreting the physics, and integrating the results into the broader scientific literature.

In the second case study, the focus shifts to the information-loss problem for evaporating micro-black holes. Claude proposes a tensor-network representation of the black-hole interior Hilbert space, utilizing a multi-scale entanglement renormalization ansatz (MERA) with a depth proportional to the logarithm of the ratio of the black hole mass to the Planck mass.

By enforcing exact trace preservation at each layer, Claude's network yields a unitary map (𝒰) that connects the interior Hilbert space to the radiation Hilbert space. This construction allows for a precise simulation of the evaporation process for a black hole with a mass of 10 M Pl, tracking the von Neumann entropy (S(t)) of the emitted radiation.

The resulting curve closely matches the Page curve, confirming that information is recovered after the predicted Page time.

Moreover, Claude Fable 5.2 successfully derives an analytic metric that respects the weak cosmic-censorship conjecture and incorporates trace-anomaly corrections. The metric takes the form:

ds² = -(Δ(r)/Σ(r,θ)) (dt - a sin²θ dφ)² + Σ(r,θ)/Δ(r) dr² + Σ(r,θ) dθ² + (sin²θ/Σ(r,θ)) [(r² + a²) dφ - a dt]²,

where Σ(r,θ) = r² + a² cos²θ and Δ(r) = r² - 2Mr + a² + α (G/π) (R² - 4RμνR^μν + RμνρσR^μνρσ), with α denoting the dimensionless coefficient of the type-A trace anomaly (set to the Standard Model value α ≈ 0.03).

Claude solves Δ(rH) = 0 analytically, obtaining a horizon radius:

rH = M exp(-α/6).

This result demonstrates how Claude's approach can yield novel insights into black-hole physics, potentially reshaping our understanding of quantum gravity and observational astrophysics.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Every outlet that reported this, and when →

Read the original at dev.to →

More in AI

Celery visibility_timeout: 214 LLM Jobs Ran Twice on Redis

The bug showed up as a spreadsheet problem. My nightly pipeline summarizes long PDFs with Claude and writes one row per document. One morning, document 4471 had two summaries.

  • Celery visibilitytimeout set to 214 seconds
  • Redis transport reinserts message into queue
  • 214 jobs run twice due to timeout

More from Thursday 19 November →