Evaluation is what matters

In continuing with the idea (or my theory) that the ability to evaluate/critique/curate the output is the critical skill required to extract value out of hashtag#genAI (more so than prompting), once again Ethan Mollick nails it: it you can't tell the difference between good and not-so-good in a given domain or field of knowledge, you won't be a power user or even a decent user, but sometimes that is enough, depending on the use case.

🤔 For example, if you are the end consumer of AI-generated art(ifacts), such as music, then it doesn't matter whether you can distinguish good from not-so-good as you are the one enjoying the output. You are in fact the best person to know what you like.

😱 If you are using hashtag#genAI to write mission-critical code and can't tell the difference between robust and brittle, that's another issue.

The critical evaluation skill is therefore dependent on the requirements of the output's end consumer.