---
title: "My AI Rube Goldberg Machine"
description: "Explore how 4 AI models analyze SaaS metrics differently: comparing net income \u0026 cash flow margins across Claude, GPT, Gemini \u0026 Perplexity. See which AI wins."
categories: ["fundraising"]
keywords: ["AI analysis","SaaS metrics","net income margin","cash flow from operations","Claude AI","Gemini AI","Perplexity AI","ChatGPT","Tomasz Tunguz","venture capital"]
ai_summary: "This post compares how four AI models analyze SaaS financial metrics, revealing discrepancies in their calculations."
date: 2024-10-29
lastmod: 2026-07-23
canonical_url: https://www.tomtunguz.com/mech-turking-ai/
author: "Tomasz Tunguz"
---

[In yesterday's post](https://tomtunguz.com/if-ai-keep-accelerating-impact-business/), I calculated the profitability of public software companies. To calculate these figures, I built a little [Rube Goldberg](https://en.wikipedia.org/wiki/Rube_Goldberg_machine) machine. 

I didn't download the data into Excel. Instead, I complexified things by sending the analysis to 4 AIs to see if they would agree.

The inspiration : many companies have used Amazon's [Mechanical Turk](https://www.mturk.com/) to crowdsource tasks, & pick a consensus answer across three workers to improve accuracy. 

Why not try this across 4 AI workers instead?


![4 robots with flags](https://res.cloudinary.com/dzawgnnlr/image/upload/thnh8tbolnep6ydiywhc.jpg)
Prompt :  "calculate the average net income margin and cash flow from ops margin from this data set" plus the data set. Note that CFOM isn't a simple average but requires dividing cash flow from ops by revenue beforehand. 

| Model | NIM, % | CFOM, % | 
| --- | ---:| ---:| 
| Claude | 4.99 | 27.31 | 
| Gemini | -9.29 | 16.2 | 
| Perplexity | -8.67 | 14.4 | 
| ChatGPT | - 9.29 | 1,433.01. / 14.9% |    
| My Analysis | -9.29 | 16.2 | 

Gemini scored top marks for tabulating correctly on both columns. ChatGPT did well with NIM but "forgot" to complete the additional division step, which I corrected with a follow up, but still not the right figure. The other systems missed the mark altogether.

It would be a mistake to draw any broad conclusions from my little experiment.

But in this case, consensus doesn't yet work as a strategy which means I still need to double check calculations myself. 

At some point, AI will mechanize the illusory [Mechanical Turk](https://en.wikipedia.org/wiki/Mechanical_Turk) & I'll restart my Rube Goldberg math machine with confidence.