Pawel Brodzinski的动态

So, how good of an idea is it to get ChatGPT to write your job application for a software development role, eh? After all, AI can write working code, so all the development tasks that were so prevalent in the past don't work anymore. And since ChatGPT can write an essay, it most definitely can write answers to open questions from the application form, can't it? We're running recruitment for another round of internships right now. And it's the first edition since generative AI became an everyday tool. In fact, we discussed how to structure our application form to give ourselves a fighting chance in figuring out genuine applications from AI-generated ones. I was against adding a message like "We'll know if you used AI to generate the answers" in our form. I didn't believe we would be able to do this. I was wrong. Here's the thing. The IT goes through a slowdown. Part of the reason is the general economic situation, but the other part is actually the dawn of AI, which is supposed to automate many programming jobs. Lots of tech companies fired their least experienced developers by dozens or hundreds. Landing an entry-level position as a developer became a challenge. We know all this, right? So, when you offer an entry-level development job, you get a ton of applications. We have already beaten our record of the number of internship applications by a factor of 5, and the form is still open! Why does it matter for AI-generated responses? I might have trouble deciding whether a single answer to a set of questions is genuine or generated by a bot. But when I see hundreds of them, suddenly (almost) all of AI-generated answers seem exactly the same. They use the same language constructs, identical out-of-the-context phrases that an actual human wouldn't use if they were actually trying to land a job at Lunar Logic. We have 3 open, non-technical questions. When I get a ChatGPT output, I know it by the time I'm in the middle of the first answer. I wouldn't have guessed it would be so obvious. Not to mention that it's actually way easier to identify an AI response, even without the benefit of comparing it to all others. Even in isolation, it's so obvious when someone didn't even try to write anything themselves. I mean, show me a person who literally writes the following in an application to a company that poses as a very informal one:? "Lunar's commitment to delivering high-quality, value-driven solutions aligns with my personal ethos of putting excellence and meaningful impact at the forefront of my work." What. A. Bullshit. So yeah, do yourself a favor, and instead of using ChatGPT to file your next job application, use an old-school spellchecker to make sure your English is fine. And show a genuine human in you. Because honestly, if we needed 200 job applications generated by AI, we wouldn't need to bother anyone. We'd handle it ourselves just perfectly.