> The truth is I don’t trust anyone else to run evals.

It's a common sentiment.

But compare https://www.cambridge.org/core/journals/judgment-and-decisio... . ("People predicting the future performance of college students state that interviewing the students aids prediction, although in fact the interviews make predictions less accurate.")

you're right there are quality issues, some probably deliberate.

but the screening cost for companies is eye watering so something should be done.