←back to thread

446 points Teever | 1 comments | | HN request time: 0s | source
Show context
tptacek ◴[] No.45030497[source]
The controls summarized in the CNBC piece seem reasonable, or, if not that, then at least not all that onerous.

The controls in the actual proposal are less reasonable: they create finable infractions for any claim in a job ad deemed "misleading" or "inaccurate" (findings of fact that requires a an expensive trial to solve) and prohibit "perpetual postings" or postings made 90 days in advance of hiring dates.

The controls might make it harder to post "ghost jobs" (though: firms posting "ghost jobs" simply to check boxes for outsourcing, offshoring, or visa issuance will have no trouble adhering to the letter of this proposal while evading its spirit), but they will also impact firms that don't do anything resembling "ghost job" hiring.

Firms working at their dead level best to be up front with candidates still produce steady feeds of candidates who feel misled or unfairly rejected. There are structural features of hiring that almost guarantee problems: for instance, the interval between making a selection decision about a candidate and actually onboarding them onto the team, during which any number of things can happen to scotch the deal. There's also a basic distributed systems problem of establishing a consensus state between hiring managers, HR teams, and large pools of candidates.

If you're going to go after "ghost job" posters, you should do something much more targeted to what those abusive firms are actually doing, and raise the stakes past $2500/infraction.

replies(7): >>45030807 #>>45031024 #>>45031331 #>>45031407 #>>45031964 #>>45033787 #>>45034461 #
neonrider ◴[] No.45031331[source]
I'm not familiar with the process of passing a law. Is it one of those situations where the ask is open to negotiation? Like, if I want to be given a finger I first need to ask for the whole arm kinda deal? If it's the case, then as you said, perhaps the real ask is what's in the summary.
replies(2): >>45031472 #>>45035404 #
MichaelRo ◴[] No.45035404[source]
The devil is in the details. There is one interview process that is bulletproof but it's NEVER going to be adopted in mass by private companies: university / police academy admission exams.

Basically you have a set number of places, say 50 jobs and accept candidacies up to a certain date, when ALL candidates (say 1000 candidates) take the SAME exam, under the SAME conditions. They all get marked from 0 to 100% and top 50 of them get the job. If anyone of them drops out, the next in line is admitted. There can be litigations filed to dispute the mark and it's objective because the criteria is the same for everyone.

The perfect system already exists, and it's used here and there. My first intern job,out of the university, was such an exam at a small business. We were some 10 candidates, 5 or so were hired. My current big corporation employer uses the exact same approach for hiring interns, only now in today's shit market it's still some 5 jobs but 500 candidates.

The real problem is that the IT domain got filled and every year the universities and bootcamps and all churn more candidates. Gotta face the fact that most people who want to become cops, who compete at the cop entry exam, will never become cops. IT is the same now.

replies(1): >>45037370 #
rlpb ◴[] No.45037370[source]
This process only works when you're hiring for an entry-level role and also don't care about differentiating for anything that isn't on your exam.

I don't think it's possible to create such an exam for senior or leadership roles, where a candidate's (professional) background is the key differentiator. Say you have two candidates for a C-suite role. One was formerly with company X and demonstrates A, B and C attributes. The other was formerly with company Y and demonstrates D, E and F attributes. How would you have created an exam that differentiated between the two, without the benefit of hindsight?

replies(1): >>45038214 #
1. MichaelRo ◴[] No.45038214[source]
I would say, when you have 2, 3, 10 candidates, you don't need an exam. Problem is when you need to be machine gunning waves of assault soldiers. An exam seems better than the usual and increasingly sick alternatives: have people waste their time talking to AI, when it's obvious all that time goes down the drain.