I'm tired of looking for jobs and rewriting my CV for every job opening. But thanks to ChatGPT, I was able to automate this and now it outputs them in Markdown format so I can send them to jobs straight away.
I would like to add a search for existing jobs and auto-responses to them in the future. But to do that I need to realise that someone is interested in this, so please leave a fitback
I'm torn. On the one hand, this is a great application of ChatGPT and a natural response to HR departments' ML-based filtering.
On the other hand, this is just another step in the deadification of the internet. It's a feedback loop: an overwhelming amount of junk leads someone (HR in this case) to produce an algorithm to filter it. This algorithm by its nature has a high false positive rate, so people game the system, initially by hand and then eventually automatically. Automation leads to even more junk and even more aggressive filtering.
It's the same pattern that plays out all over the internet, and the end result is that high-quality non-gamed content becomes nearly impossible to find in the mess.
Not that you're wrong to build this step—it's the natural and rational reaction to the system in which we find ourselves. I just wish there were a way to stop the cycle and get back to human interaction.
I understand. I'm wondering if I should do AI for HR to catch my own AI resumes and cover letters. It's kind of unethical, I guess.
On the other hand, it's very hard to find a job for non-top professionals right now. And I'd be happy to help them.
It's an extremely difficult ethical issue. I don't think I can find an answer to it very quickly.
HR teams need to filter, wind up using ML systems to distill us down to keywords.
Applicants start using ML.
Now we have computers creating applications to be reviewed by computers for a job to get hired where you use ML to try to shortcut the amount of work actually done, while your health insurance claims are judged first by ML and your company's stock is traded by ML...
The instructions need significant tuning. You should construct a few test cases, ranging from minor tweaks (close alignment in CV/job) to highly divergent (librarian applying to medical director of heart clinic).
With a couple quick attempts, it completely changed all of my 8 jobs over 20 years to include mentions of the target job, even if it meant adding unbelievable lies like running a scuba certification program and improving training using my knowledge of library sciences. The instructions may need examples of how to remain truthful and to just skip roles that are hopelessly unrelated instead of fabrication.
You could even suggest it adopt a multi-step process with a self-evaluation if that helps. I made a recipe reformatter that worked that way and it improved the output by finding places it accidentally dropped or changed an ingredient's amount, for example.
Do the settings need to be changed or something so it's public? I just tried but it only shows me the basic info and won't let me chat. I have a paid sub so I should be able to use it if it's properly permissioned.
This is some kind of problem at OpenAI.
I have rechecked all the settings several times. I can see 30+ chats.
But some are working and some are not. Maybe because of the country of the account =(
I built a similar thing and found it to add qualifications quite often, especially when you had it write a cover letter. It was still helpful though, just like with every use case you need to check what ChatGPT is telling you.
I'm working on it. Often ChatGPT lies a bit, as their main concern right now is that you pass the HR screening. But I think this can be tweaked in the instructions
Maybe, if it’s also trained on my experience. Then, it would say, No. the ad is referring to something else when they say ‘works well in a team’ and not your weekend CounterStrike LAN parties. Or, Yes, your twelve Python scripts does qualify you for a senior developer position at this lemonade stand. (Whatever)
On the other hand, this is just another step in the deadification of the internet. It's a feedback loop: an overwhelming amount of junk leads someone (HR in this case) to produce an algorithm to filter it. This algorithm by its nature has a high false positive rate, so people game the system, initially by hand and then eventually automatically. Automation leads to even more junk and even more aggressive filtering.
It's the same pattern that plays out all over the internet, and the end result is that high-quality non-gamed content becomes nearly impossible to find in the mess.
Not that you're wrong to build this step—it's the natural and rational reaction to the system in which we find ourselves. I just wish there were a way to stop the cycle and get back to human interaction.