Personally, the metadata that I would value (and several of these are already possible with the AI filters) :
- Article length (word count, time required to read)
- Full text available or not (after NewsBlur’s full text extraction)
- Readability score
- Contains embedded videos or audios?
- Listicle? Clickbait?
- Is this content AI-generated (low perplexity)?
- Sentiment - positive or negative?
- Using the two new categories: is this article widely read or a long read?
- EDIT: one more - language of the article (e.g., NYT sometimes publishes Spanish language items in the main primarily-English-language feed)
The metadata that I think others might value:
- Trigger words and content - e.g. mental health related
- EDIT: one more, does it contain a code block?