Battle Royale: Grok vs. ChatGPT

[John Bauer]

When it comes to which AI model is better, not everyone agrees. I think Grok is wrong too often, others think it is superior to ChatGPT. Rather than trading anecdotes, we are going to run a simple experiment: A single high-level question, asked once to each model, with no follow-ups or "hand-holding," and then compare the answers.

The question we will pose to both models is this:

Provide a detailed, step-by-step reconstruction of the genealogical data and cultural-linguistic evidence (including ancient Near Eastern context, Mesopotamian, Egyptian and Levantine sources) that supports locating the original garden of Eden in the headwaters region of the Tigris and Euphrates rivers. Then evaluate three major objections to that location (geological, botanical, and textual) and offer a reasoned conclusion.

This question is deliberately loaded. It demands multi-disciplinary analysis (genealogy, ancient languages and cultures, ANE background), engagement with external data (geography, geology, and botany), serious textual work (close handling of Genesis, plus interaction with objections). In other words, it's not about who can regurgitate a commentary blurb. This will test whether the model can synthesize data across disciplines into a coherent, reasoned argument.

We will post both answers once we have them. Readers can decide for themselves which model handled the question better.

 

[makesends]

I doubt either of them will do "serious textual work", but only parrot what they have 'read'. They won't analyze pottery nor geological layers, but repeat what geologists and archaeologists have said.