Yes, better tools to analyze data will yield great results. Even a good push to scan all those finds and make all the data available would probably allow amazing new discoveries. The catch is that people like to hoard that data and milk it for their own careers and fame.

That said… LLM is Large Language Model. By definition, LLMs are unlikely to analyze 3-dimensional shapes. The newer AIs, like Gemini or GPT-4o, also use vision and audio but they are often still called (multimodal) LLMs. It’s justifiable as they still seem to have language at the core, but it’s getting increasingly dubious.