You can still easily spot actual work, whether in website design or recruiter messages. Isn't that the whole schtick of the article?

But low-effort spam will always dominate, that's just the nature of it. You will always receive 10 times more messages that were only 1/10th of the work on the other side. How could it ever be otherwise?

That’s what makes this so annoying. I don’t mind a template, but the fake personalization gives it a whiff of deception in a way a pre-LLM simple template mass mailing doesn’t.

Sharing a trick I learned here on HN: I prefix my first name on LinkedIn with an emoji (green check box).

Whenever a message starts with “Hi (emoji) (first name)” it’s 100% spam.

So far they are not training to fix this (all they have to do is drop the prefix), which is somewhat surprising. Zero messages that address me correctly seem automated. Maybe because in today’s day people can legitimately have emoji in their names?

Emoji on linkedin is itself a tell of llms

I started doing this a while back, and will say that recently, they are stripping emojis.

it was nice while it lasted. I guess my next step is a jev+cron to sort good from bad. Browser automation here we come. Someone want to write a Firefox extension that runs in your open browser tab? I would pay for this.

What do you consider good?

whatever I told it was good. presumably it can be trained.