I tried Vibe Physics. This is what I learned.

Quick Overview

The presented video explores the feasibility of using AI models, specifically GPT-5, Grok 4, Gemini 2.5, Gemini DeepThink, and Claude Opus 4.1, to solve a physics problem related to the Navier-Stokes equation and General Relativity, ultimately concluding that while the idea is brilliant and creative, the inherent mathematical differences between the two fields make a direct mapping unworkable.

Key Points: The core idea is to embed Navier-Stokes (NS) dynamics into General Relativity (GR) using a specific stress-energy tensor and coordinate system, then leverage Penrose's singularity theorem to prove a blowup. The video ranks AI models based on their ability to handle this complex physics problem, with GPT-5 and Grok 4 performing better than Gemini 2.5, Gemini DeepThink, and Claude Opus 4.1. A fundamental mismatch exists between the kinematic blowup of NS on a fixed background and the geometric breakdown of GR background itself, making the proposed approach unworkable. The incompressibility condition of the fluid in NS is identified as a primary barrier, preventing the energy-density blowup needed for gravitational singularity. The video critiques the potential of AI in solving such complex theoretical physics problems, noting that while AI can assist in research and analysis, it cannot generate novel conceptual breakthroughs or perform the specialized abstract mathematical reasoning required for a Millennium Problem. The presenter suggests that while the idea is conceptually elegant, it faces monumental technical and conceptual hurdles due to the fundamental nature of the equations and the 'dissipative character' of NS versus the 'conservative geometric structure' of GR. The video concludes that the attempt to directly map NS to GR is problematic due to these fundamental differences, but acknowledges the creativity and brilliance of the thought experiment.

Context: The video discusses a user's ambitious physics idea: to solve a Millennium Problem by connecting fluid dynamics (specifically the Navier-Stokes equation) with General Relativity (GR) using AI models. The user proposed using AI to find solutions within GR that mathematically map to the Navier-Stokes equation, aiming to prove a singularity using Penrose's theorem. The presenter evaluates this idea, comparing the performance of several AI models and analyzing the theoretical challenges involved.