Thank you for the clarification. In your opinion, what is the level of knowledge or particular signal you’d expect as sufficient for a “generalist” to have, and is it equally necessary among the roles (e.g. how much would it differ for otherwise skilled operations vs. research management professionals )? Would be great to talk if it’s more comfortable and if you’re open to.
Btw, I just thought I should say that I really appreciate you folks writing this post. I don’t want you to think that I disliked your article just because I disagreed on one point (which is how things can sometimes come off).
I should say that relatively little management experience (largest AI Safety ANZ was when I ran it was me and Yanni part-time, a part time ops contractor and an intern), but that said key crux for me is this:
• Option 1: hire someone and severely limit their promotional potential, acknowledging the weird dynamics this might create • Option 2: hire someone with a reasonable level of value alignment and ability to understand strategy
Option 1 might work for specialist roles (ie. if an org needs an accountant, that person might be fine only ever being an accountant). It’s worth noting, that even if you do this, there’s still a cost from bringing them into the field insofar as someone else may hire them to do a role they’d be ill-suited for.
In terms of understanding strategy, it’s important to realise that different people have wildly different worldviews and ways of seeing the world. You can collapse these down to a few dot points and tell yourself that you understand the different perspectives, but you’d just be kidding yourself (I’ve made this mistake myself in the past).
I’m pretty busy, but feel free to ping me in like two weeks.
In terms of understanding strategy, it’s important to realise that different people have wildly different worldviews and ways of seeing the world. You can collapse these down to a few dot points and tell yourself that you understand the different perspectives, but you’d just be kidding yourself (I’ve made this mistake myself in the past).
This is a curious point to explore further, thank you for raising it and describing vividly. If I understand you correctly, it complements one of our observations—that worldviews might be a tricky thing to evaluate, which raises the question of what exactly AIS organisations look for when they look for alignment, and how the job seekers need to obtain or signal it. That’s what we’d love more data on, alongside with our more general questions.
Thank you for your encouragement and being open to further contact.
Sorry, autocomplete got me. I meant mentorship. I’ll update
Thank you for the clarification. In your opinion, what is the level of knowledge or particular signal you’d expect as sufficient for a “generalist” to have, and is it equally necessary among the roles (e.g. how much would it differ for otherwise skilled operations vs. research management professionals )?
Would be great to talk if it’s more comfortable and if you’re open to.
Btw, I just thought I should say that I really appreciate you folks writing this post. I don’t want you to think that I disliked your article just because I disagreed on one point (which is how things can sometimes come off).
I should say that relatively little management experience (largest AI Safety ANZ was when I ran it was me and Yanni part-time, a part time ops contractor and an intern), but that said key crux for me is this:
• Option 1: hire someone and severely limit their promotional potential, acknowledging the weird dynamics this might create
• Option 2: hire someone with a reasonable level of value alignment and ability to understand strategy
Option 1 might work for specialist roles (ie. if an org needs an accountant, that person might be fine only ever being an accountant). It’s worth noting, that even if you do this, there’s still a cost from bringing them into the field insofar as someone else may hire them to do a role they’d be ill-suited for.
In terms of understanding strategy, it’s important to realise that different people have wildly different worldviews and ways of seeing the world. You can collapse these down to a few dot points and tell yourself that you understand the different perspectives, but you’d just be kidding yourself (I’ve made this mistake myself in the past).
I’m pretty busy, but feel free to ping me in like two weeks.
This is a curious point to explore further, thank you for raising it and describing vividly. If I understand you correctly, it complements one of our observations—that worldviews might be a tricky thing to evaluate, which raises the question of what exactly AIS organisations look for when they look for alignment, and how the job seekers need to obtain or signal it. That’s what we’d love more data on, alongside with our more general questions.
Thank you for your encouragement and being open to further contact.