117 karmaJoined


Sorted by New


Thanks for flagging this Neil, good find! We've put this on the medium/low priority list for ARC & METR, since the EV is pretty low given our donors (but I think positive!). 


One little puzzle: given that METR is trying to stay independent from major AI labs, do we reject donation matches from Google?

My best tool is to become a connoisseur of what it's like to be shifting into reactive / fire-fighting mode, and make a craft of switching back to prioritizing. 

(responding to this post has the sort of dizzy pulling-away feeling that reactivity has, so I'm going to yolo submit and try to shift back to proactive mode) 

Thanks for writing this up! It also strikes me as a good overview and will probably be my default link if people ask me what ops is like. I like the specific examples across different orgs. It varies a lot! 

My ops buckets:

  • HR/Legal/Governance (lots of checklists driven by government compliance. Getting the right insurances, drafting company policies like misconduct, ensuring the board meets and approves the right things and discloses conflicts of interest)
  • Finance (Ensuring people get paid. Creating systems to track transactions, so that we can get accurate reports about the flow of money, predict how much we will spend for fundraising, prioritization, etc)
  • Basic logistics (Running an office and/or events, ensuring people have a place to work with working electricity, plumbing, and internet. Maybe some snacks :)

These buckets are specifically geared towards 'supporting beams', bases for me to ensure I have covered. Helps to keep priorities in order and avoid the always-putting-out-fires syndrome. 

That said, I'm probably wrong in my thinking about this, and it will vary widely across organizations and staff. Specifically I think I'm wrong in that priorities are more dynamic and diverse than this framework implies, so it's more an ever unfolding process of sorting out the prioritization of supporting beams / need-to-haves vs. nice-to-haves. 

Ah - maybe your post is making the point "if they would make a good senior hire, it seems fine to hire them in a junior position". Maybe I was getting confused by the term, I've seen people labelled 'overqualified' when they are above average on a few dimensions but not all of them. 

I'd have a harder time steel-manning a counterpoint to that. Maybe something about it not being stimulating enough so risking turnover... but that doesn't hold much water in my mind. 

Would you agree that, if Bob was more politically skilled, he would be a better fit for this position?

Yes... and no? 

Yes: it would be better re. 'overhead required'. If Bob foresees Carol's objections and takes her out to lunch and convinces her, this could save a bunch of management/board time. 

... and no: maybe Carol's concerns were legitimate and Bob was just very convincing, but not actually right. Fade to: Bob becomes CEO and the org is thriving but it's not really following the original mission anymore. 

I'm guessing Steve Jobs wanted people to convince him if (and only if) they were right. 'Right' meaning not just factually correct but probably also whatever Steve thought was good (whatever that was). 

So maybe if Bob was more politically skilled and also aligned with the mission of the organization? But aw geez now we're back to how it's hard to hire people aligned with the org. Hmm, that would probably cruxy too. Not sure how to measure it. 

Thanks Andrés, this helped me get oriented around the phenomenological foundations of what y'all are exploring.

Apparently my comment won a comment prize, which nudges me to carry on this conversation.

In general, I'm skeptical about "putting" people in leadership positions, especially when their colleagues don't want to be led by them

What if Bob has an ambitious project he's excited to run, and 4 out of 7 of his colleagues are excited by this project and want to be led by Bob on this, and Alice thinks it couldn't hurt to try, but Alice's cofounder Carol really doesn't like the idea and 2 of the 3 board members also don't like it? Carol et al. surface objections like 'it's not in the spirit of our mission' and 'I'm worried about the effects Bob's leadership would have on our culture'.

Maybe if the org had good culture and good leaders they could figure out how to thread the needle, give Bob's project a shot while addressing the concerns that Carol et al surfaced. 

But I guess the point is... all of that takes time. A lot of effort needs to be put into the work of coordinating around a contentious project. In a world where Bob was vetted as a senior  hire (which again, takes more time) he wouldn't have made the cut because of these concerns. But since he was vetted as a junior hire, people didn't think to consider 'the effects of Bob's leadership on our culture'. 

To be clear, I think a good hiring processes would sufficiently address these problems at the beginning, either vetting Bob as a senior hire and/or ensuring he understood the scope of the role.  A good hiring process would probably notice that Bob has the skills to fulfill the finance role, but does not have the skills to lead in the organization.

But... it's hard to make good hiring processes, it's hard to anticipate how this kind of thing will play out. In the face of this, I think it's somewhat reasonable for hiring managers to lean towards junior hires in some cases. Like if I 'just want someone to get this one set of things done reliably for the next two years" I might have less headache with a junior hire that shows up, does the thing, and goes home. 

If I hire an office manager and they start trying to reform my HR policies, this can be more of a headache than a help. I didn't think to vet them for their HR policy skills, but now they are feeling bored and upset that I'm not giving more detailed feedback on their proposals. But I don't have time right now. And then they leave the org because they feel unfulfilled, and we need to recruit for another office manager. Aagh! Let's just hire a junior person!

Even if I was a bad manager above, I think it's a pretty realistic situation and have the sense that a lot of orgs/managers have been burned like this, and so are more cautious about making senior hires.

To be clear, my all-things-considered view is pretty uncertain, and leans towards being willing to hire overqualified people (I also tend to be pretty bullish about hiring outside the standard EA demographic). Especially if you have decent management expertise, which would ensure e.g. good hiring processes. 

Looking back at my comment, I'm still a fan of this model:

the more agent-y the person you hire, the more you'll need to be careful about principal-agent problems

Something that feels maybe cruxy: Do overqualified hires have higher turnover? Higher management overhead?

One angle on how this could go poorly is something I call 'failure cascades' (a la information cascade). I'm excited that this has been incorporated as a concept in the EA Ops channel, and I think it would be valuable for EA consultants to keep it in mind. 

Roughly, a failure cascade could be:
> An EA consultancy conducts a search for a really good immigration law firm that they can use when helping EA orgs with immigration. They find a good law firm and proceed to help a dozen EA orgs with visas. Unfortunately it turns out this law firm misunderstood an important component of the H1B renewal process, and suddenly three years later a bunch of EAs across the ecosystem get kicked out of the USA. So a single judgment failure 'cascades' across the organizations, causing a relatively catastrophic situation for the community compared to a world where infrastructure/judgment was less centralized. 

Just got off the phone with Fidelity Charitable, they accept ETH with no minimums. (Also the two agents I spoke to were smart and efficient, average wait time of 8 min)

FYI  @MichaelDickens I just heard from Vanguard Charitable:

> At this time, Vanguard Charitable only accepts contributions of Bitcoin and Bitcoin Cash that are valued over $100,000.00.

Might be worth mentioning in the post.

Load more