Ethan Mollick tests GPT on Linear A, warns it's 'researchslop' until GPT-6 cracks it

emollick · x · 2026-09-15

Ethan Mollick shared an experiment prompting GPT to tackle Linear A, the undeciphered ancient script. He tempered expectations: unless GPT-6 says "I can now read all Linear A and here are the translations," this is just "researchslop." He also admitted his prompting technique is likely terrible and advised readers not to copy it.

Original post →

More from Models

Models channel →