I’ve spent a chunk of the last few weeks doing what a lot of people are doing right now: looking for work.
Nothing to announce, no contract signed. But a few things have happened along the way worth writing down, not because they’re about me, but because they’re about what hiring looks like right now from the other side of the table.
Three moments stood out. Two I said no to. One said no to me, and by the time it did, I wasn’t sorry.
The repo I didn’t clone
More than one recruiter this month asked me to clone a GitHub repo and run it on my own machine as a “technical test”. All of them were for roles adjacent to crypto or blockchain.
I said no every time, partly because of what they were asking for, and partly because of how they were asking for it.
Here’s the part worth knowing if you haven’t come across it yet. There’s a scam running right now, tracked by Microsoft’s own security team, where fake recruiters pose as crypto or AI companies, send over a coding “assessment”, and the repo itself is the attack. Clone it, run it as instructed, and it quietly goes looking for your saved passwords, your crypto wallets and your API keys.
I’m a product designer. Nobody legitimate books a design interview and then asks you to execute someone else’s code on your own laptop. That mismatch was the first flag, before I’d even thought about malware.
A real interview process doesn’t need you to run someone else’s code on your own machine. Saying no costs you nothing when the process is legitimate.
The interview I never turned up to
One process wanted a recorded interview conducted by an AI interviewer instead of a person.
I said no before it started. An interview isn’t only them evaluating me. It’s my one real chance to ask questions back and work out whether the place is right for me too. A script running on a server can’t answer those, and nobody could tell me who reviews the recording once it exists.
I’m clearly not alone. A report from Greenhouse, covered by Fortune, found 63% of US job seekers have now sat an AI interview, and 38% have walked out of a hiring process rather than do one. Of the ones who went through with it anyway, 51% were then ghosted or are still waiting to hear back.
The “friendly chat” that wasn’t
A different process booked a video call on Google Meet, titled “Call for Senior Product Designer”. The meeting description asked for cameras off.
If voice was really all they wanted, a phone call would have done that. Instead they booked a video call and then asked nobody to use the video, which is a strange amount of effort to go to for less than a phone call gives you.
The questions came one after another, fully scripted, with no space for me to ask anything back. Partway through, they mentioned that if this stage went well there’d be three more rounds after it: an AI-automated one, then a third, then a final stage. Four in total.
I asked about that. They told me this first call wasn’t really an interview, just a friendly chat.
Calling forty minutes of scripted questions deciding whether you reach round two of four a “friendly chat” is taking the piss. Whatever you call it, it’s an interview, not a chat. It didn’t go anywhere in the end, and by the time they told me, I wasn’t sorry. My heart had left the process a while before they did.
The one that got it right
Not every process has been like that. One conversation this month felt completely different, and the difference was simple: I got to ask real questions, and they actually answered them. It felt like two people deciding something together, not one side interrogating the other.
That’s the whole difference between a good interview and a bad one right now. Not the AI, not the number of stages, not even the video call that turns into a phone call halfway through. Whether both sides get to ask something.
What this means if you’re hiring
If you’re the one running a process rather than sitting through one, here’s the pattern worth noticing. None of the companies above think they’re doing anything wrong. The scripted call was probably built with good intentions: consistency, fairness, a way to compare candidates evenly. The AI interviewer probably saves someone real hours every week.
I made the case a few weeks ago that AI is a paintbrush in this job, not the painter. The same holds here. AI can filter a flood of applications. It can’t build the two-way trust that makes someone want to work with you, and it can’t make up for a process that never lets the candidate ask anything back. That part’s still a human job, and skipping it costs you exactly the people you most want to hire, because the good ones have other conversations happening at the same time, and they notice.
If your process wouldn’t survive a candidate asking a question back, that’s worth looking at, whatever it saves you in recruiter hours.
So that’s the update. No big news, still looking, still having some strange conversations along the way.
I do have capacity right now. Feel welcome to book a call, I’d love to hear all about your project and how I can help.
Comments