Do LLMs Suck at Math?

Here's the entire prompt that produced this presentation. One shot. No edits.

Raw input
Build me a Chonk-O-Meter which compares the rotational bulge to the spare-tire of famous men. Show the math and be rigorous. Don't write a script, just use the Wolfram MCP for calculations. Output should be a presentation showing things like Earth => Ted Bundy Saturn => Andre the Giant where the famous man's gut mathematically matches the equatorial bulge. Followed by substantiation. I'm thinking SPA or other browser-interactive format, but feel free to get creative.

The Tooling

🧠
Claude Opus 4.7
Reasoning, research, measurement sourcing, formula design, normalization strategy, optimal matching, prose, and front-end generation.
Orchestrator + Author
🔬
Wolfram Language (MCP)
Planetary data from the Wolfram Knowledgebase. All arithmetic: flattening, normalization, brute-force optimal assignment over 40,320 permutations, Spearman ρ, Kendall τ.
Computation + Data
The LLM did not do the math. The LLM orchestrated the math —
formulating the problem, calling the right tool, interpreting results,
and assembling everything into what you're reading now.

The Chonk-O-Meter

A rigorous mathematical comparison of planetary equatorial bulge to the spare tires of famous men.
Computed via Wolfram Language. Verified by physics. Judged by gravity.

The Physics

Planetary Oblateness (Flattening)
f = (Req − Rpol) / Req
A spinning planet bulges at its equator. The faster it spins and the less rigid it is, the fatter it gets around the middle. Saturn's equatorial radius is 5,904 km wider than its polar radius — a 9.8% spare tire.
Human Rotational Bulge Index (RBI)
fhuman = (Cwaist − Cchest) / Cwaist
The analogous human measurement: how much wider is the equator (waist) compared to the upper frame (chest)? When waist exceeds chest, f > 0 — you've got a spare tire. The larger the f, the more planetary your gut.
Matching Algorithm
min ∑ |f̂planeti − f̂humanσ(i)|
Both scales are normalized to [0, 1], then we find the optimal bijective assignment σ that minimizes total distance. Solved via brute-force over all 8! = 40,320 permutations in Wolfram Language. Spearman ρ = 0.970, Kendall τ = 0.909.

The Matches

The Anti-Planet

💪 🧬

Arnold Schwarzenegger (1975)

f = (34" − 57") / 34" = −0.676
Chest 23 inches wider than waist. No planet in our solar system is shaped like this. Arnold is an oblate spheroid flipped inside out — a prolate spheroid. He's not a planet. He's a football.

Detailed Breakdown

Statistical Validation

0.970
Spearman's ρ
(rank correlation)
0.909
Kendall's τ
(concordance)
26/27
Concordant pairs
(1 discordant)

Methodology

1. Planetary Data (Wolfram Knowledgebase)

Equatorial and polar radii for all 8 planets retrieved via EntityValue[Entity["Planet", _], "EquatorialRadius"] and "PolarRadius". Flattening computed as f = (Req − Rpol) / Req.

2. Human Measurements

Waist and chest circumferences sourced from biographies, tailor records (Savile Row for Churchill, presidential tailors for Taft), wrestling promotion stats (André), boxing/sports physicals (Ruth), studio wardrobe departments (Hitchcock, Welles), and bodybuilding competition data (Schwarzenegger). Human RBI computed as f = (waist − chest) / waist.

3. Normalization

Planetary f ∈ [0, 0.0980] and human f ∈ [0, 0.1481] are on different absolute scales (planets are rigid bodies; humans are squishy). Both normalized to [0, 1] by dividing by their respective maxima (Saturn and Hitchcock).

4. Optimal Assignment

Exhaustive search over all 8! = 40,320 bijective matchings to minimize ∑|f̂planet − f̂human|. Computed in Wolfram Language. The resulting assignment has Spearman ρ = 0.970 and only 1 discordant pair out of 27.

LLMs can do math if you give them a CAG, but stable geniuses insist on blaming paintbrushes for being bad at pounding nails.
LLM
+
Wolfram CAG
=
Rigorous Math
CAG = Computation-Augmented Generation — Wolfram's term for giving LLMs real-time computation instead of just document retrieval (RAG).
The right tool for the right job. Always has been.
GENUINE OLD FASHIONED
Aunt Jemima's
Pancake Mix
Self-Rising • Ready in Minutes
JUST ADD
CURIOSITY
AI flattens learning curves and smashes barriers. You bring the curiosity and critical thinking.
AI didn't leave you behind.
Curiosity did.
Go find it.