YHK - yet another hacker news clone

This is one of my bigger issues with generative AI.

It used to be easy to spot effort: a cool logo on a business card, an interesting website design, a recruiter email that speaks to a specific experience I have on my LinkedIn profile. This used to indicate that someone was willing to do a bit more than the bare minimum and gave them some legitimacy.

Now I see any of these things and I have to do the extra work to try to figure out if it was AI generated.

These are truly maddening. I’ve recently fallen prey to another trick - email from a famous-ish tech person starting a new company and asking to chat, but then responding with interest just enters you into the normal pipeline with a hiring manager. Nothing wrong with that pipeline, but don’t misrepresent the outreach as an opportunity to chat with an influential person when it’s actually just an automated recruitment email.

I get tons of these. I suspect they must run on crappy local models to save money, because the ones I get are also terrible at drawing a simple line connecting my experience with the job.

I look at them with a mixture of pity and contempt. They are examples of "In the land of the blind, the one-eyed man is King" work: i conjecture that this type of "shallow customization" is going to be most persuasive with recipients who have very little exposure to slop and/or the kinds of things human recruiters used to say when visually scanning some stale copy of a résumé while on the phone with the prospect.

Like robocars killing people, it feels worse when it's AI.

I mean, before it was humans using template text with dynamic fields; in some cases the first time the sending human would actually see the message would be if you responded.

But now you can get all the way to the first call booking without a human ever being troubled to participate in the interaction.

In my experience, LLMs have not improved the quality of targeting (let's assume for each person doing outreach there are at least a few people who actually want the messages being sent), and my inbox is absolutely flooded with AI generated campaigns. I have stopped unsubscribing and I just mash the spam flag.

Invest in your actual network. It makes ignoring recruiters extremely easy. I try to remind myself of this.

> That's the gap.

That's one of the shortest LLM tells I've seen.

We need to have RCAPTCHA where the recruiter is quizzed on job images.

CLICK ON ALL IMAGES OF DEVOPS PIPELINES

CLICK ON ALL THE OKR IMAGES

I assume it’s like how scams have typos to self-select for people who are not paying attention. Presumably any candidate who cares is not worth going after

I actively report these as spam (they are) and block the email. +1 on insulting intelligence.

> they probably haven't spent more than 10 seconds looking at my profile to see if it’s a fit.

I doubt they even glance at your profile, probably scraped and stored somewhere that's sold to these AI recruiting / sourcing companies the recruiters use.