A few jobs ago, I worked as a system architect at an org that wanted to build a new IoT product. The tech stack had to be Erlang and FreeBSD, for a cool combination of network protocol decoding and decent platform-level security. There was one issue though: we were in Quebec city, which is far from a mecca for Erlang nor BSD, and the only two companies I knew of at the time with that tech stack were WhatsApp and some small Italian company that did audio translation of some kind (I think?)
In short, we were gearing up to hire people we knew we would have to train regardless of their skillset. The people behind the project had picked a tech stack we ended up assuming we could not find good fits for in the first place. That constraint, however, ended up being a kind of a blessing.
We had to plan for training, and we knew we couldn't do some coding test super easily—we did have a small screening set of questions to pick people who had some familiarity with either functional programming concepts, network programming, or FreeBSD as an OS—which forced us to really structure the interviews differently. If you were good enough in one of them and willing to learn the rest, we could teach the rest, so long as you were willing.
Rather than testing people and figuring out whether they knew enough to get in, we explained the situation, had a good chat with them (we still wanted to figure out whether the workplace would fit their expectations) and flat out said: Look, we're going to train you in what you don't know. What is it you'd like to get better at?
And because of that [honest] framing, we had very candid answers. People would openly admit the things they were weak at, the parts they felt were shaky, and things they knew about and were confident with. And regardless of whether candidates were junior or more senior, we generally got answers that were true and accurate.
It stuck with me that treating people as responsible and human, but also that we wanted to invest in them to make us successful as well, made the whole fucking game of weird-ass tech interviews mostly useless. It was so straightforward and nice in comparison.