What to do if I am working with a vibe coder who just wants to implement things right away according to what chat gpt says without studying context or giving me time to assess
retained text
Hey guys, I’ve recently taken over as sys admin in the company I’ve been working at as an operations admin for the past year. I’m in the process of rolling out a series of procedures to tighten security and so that we can move away from the current third party IT vendor that has been assisting the business One of the things I’m trying to really focus on is access control, account management and privileged access restrictions (as previously just anyone would get given global admin etc if they requested from the external IT vendor, hr had user admin set in their every email and so on). I’m setting policies on entra and going to do a big inventory of all the registered users / guests with access share points, upgrade everyone still using windows 10, start entra joining company devices and enrolling in intune etc etc In the office we have this vibe coder who is building a new system for the company using AI (like a CRM so people can enter timesheets, HSE documentation, tracking projects etc). He had an idea of how the system should be to make everyone’s life easier since he used to work outside on the operations field, and he fed that idea to ai. This was built on Azure and office staff who are entra registered can login with their Microsoft accounts and standard infield workers have a separate login I haven’t had much involvement with the software but today we had a meeting with an external auditor (a friend of the CEO) and he suggested implementing firewall rules etc directly in Azure. (For context this was built on mr vibe coder’s personal azure subscription linked to the company entra and I assisted transferring the subscription over and just gave him an ‘owner’ role in azure and limited it to that) Now that you have a brief context, the problem I am facing is that whenever mr vibe coder wants something (such as a permission or to implement something) he just comes up to my desk with instructions given by chat gpt and wants me to do that straight away the exact way chat gpt is telling him to do. But I have a process in place and I like to assess and study things before I do them, like for example for anything I’ll implement I like to make a small planner for myself and get a deeper understanding of how it works to then implement it. Of course I use chat gpt but as I recently cyber graduate with some knowledge I know that some things need to be done a certain way because of everything else it affects and because it’s important to have a process in place, I can’t just blindly go there and click it and done like he thinks should be done (for example when transferring the azure subscription across he wanted me to just go and make him global admin on the spot) he isn’t considering the whole picture With the Azure security stuff I’m skeptical that he will rush me and want to implement things straight away as per chat gpt’s instructions, but I will want to study the system and refresh my understanding of its needs and the best way to implement rules and policies etc. I know we are in a new era with AI automatic everything, but I think it’s a lack of standard to blindly do things without seeking deeper knowledge on what we are actually doing. Am I thinking about this the correct way? And how do I proceed in this situation??? For social context, I’m 22F and mr vibe coder is 33M (and doesn’t speak English very well so there’s a bit of a language barrier at times although English is also not my first language)
document fetched 2026-09-30 00:40 · canonical https://reddit.com/r/sysadmin/comments/1wtbw46/what_to_do_if_i_am_working_with_a_vibe_coder_who