I think this trend has been around for a while now (it started to become more noticable for me at least a couple of years ago).

At first there seemed to be a correlation between how 'cool' the project was and the number of emoji, but now it seems like it's as expected as just having a README itself.

I've definitely seen _more_ decorated READMEs, and I can't help but feel like there's an inverse correlation between emoji count and readability.

Not specifically about readmes/GitHub repos, but I've noticed some LLMs like Sonnet and GPT4.1 are really enthusiastic about doing emoji-prefixed lists for some reason.

Trained on too many JS libraries.

Not necessarily a helpful thing, In fact I think that we were to use this to create dynamic prompts then it increases exponentially

When ever I see tons of emojis in a list/faq/readme I first think of LLM output.

I do as well when I see:

# Code Comments Every Few Lines

It's probably LLM generated. Adding a fun/cool factor to the project. I created a Chrome Extension where you can "emojify" any text with a right-click. https://emoji-bot.com

Basically any LLM

AI, I presume (but I’m not sure) that the code “agent” they are using creates it.

Tool is an insult when applied to a human...