Unpopular opinion: most career assessments wouldn't survive basic scientific scrutiny

Someone in a thread here who used to build matching-site infrastructure said something that's stuck with us: personality tests can have reliability (consistent results) without validity (actually predicting anything real). They're fun to talk about at parties. They're not something you should base a career decision on.

Here's the uncomfortable part: if these tests actually worked, companies would use them for hiring decisions at scale. Most don't, because HR departments already know what the research shows — these tools are entertainment dressed up as insight.

We're not claiming PotenAI has this fully solved either. But we refuse to ship the same thing with a shinier AI wrapper. Our output has to reference a user's actual answers back at them — not abstract trait labels that could apply to anyone. If we can't point to something real you actually said or chose, we don't ship it.

Is this harder to build and prove than a 20-question quiz with a nice UI? Yes. Do we think the industry's current bar is embarrassingly low? Also yes.

Tell me why I'm wrong in the comments — genuinely want the pushback.

🌐 https://nopick.site — waitlist open, early access coming soon.