LLMs and GPT can rate anything

LinkedIn posts, mails, documents (with images), or even books can be rated with GPT.

The evaluation grid can either be the one chosen by GPT or the one specified. When requested for the rationale, GPT does a good job of explaining its rating.

Since GPT can rate unstructured documents, it can logically rate more structured data in various domains. This includes banking or finance, where GPT assesses the risk of an individual or company defaulting on a loan, evaluates whether a resume matches a job description, or even acts as a dating machine by matching people according to their preferences. The sky is the limit! 

Since you can ask GPT for the rationale, technically, GPT can replace complex and expensive systems created and maintained by top-level engineers and scientists.

This shows that the real power in the near future won't be about who has the best engineers or AI algorithms, as these are becoming commodities (although the pace is frightening; who could have thought about the situation two years ago?).

The real game changer is who owns non-synthetic, logical, frequently updated, and strong databases.

LLMs are very good at making realistic synthetic data, but nothing beats real-world data (who could have thought that during COVID, toilet paper was scarce!).

I believe that businesses that capture data are likely to be the big winners in the new technology race that started two years ago.

My prediction is that retail and e-commerce will become even stronger.