I feel like this sort of entirely misses the point of the post. Obviously there are people who will go work somewhere for less pay than some of their alternative options. The question is not whether those people exist, it’s whether there are enough of them with the qualities that Lightcone (and other EA orgs—this is not a Lightcone-specific post) is looking for such that Lightcone can “have its pick” of extremely qualified candidates without spending an excessive amount of time and effort looking for them.
First, Lightcone (and other EA orgs) are not hiring from the general pool of software engineers. They are hiring from the pool of software engineers who are already mission-aligned, which constrains them to maybe, if we’re lucky, a low five-digit number engineers in the US, a substantial fraction of whom don’t live in the Bay Area and probably don’t want to relocate. If I’m wrong about how strongly EA orgs filter on whether their desired candidates are mission-aware and mission-aligned, that does render a substantial chunk of my concerns moot, yes.
Second, Lightcone (and certain other x-risk focused orgs) are not in the business of trying to optimize on labor costs; they’re racing against the clock and want to be wasting as little time as possible on things that don’t directly advance their research agenda like “looking for & hiring people”. If this isn’t true, my argument is that it should be true. Spending an extra month looking for a similarly qualified candidate (or, worse, settling for a less qualified candidate) is a cost that has a dollar figure attached to it. If you think we live in a world with short timelines, it has something worse than a dollar figure attached to it.
I’m sure Lightcone would probably not have much trouble hiring from within the pool of engineers that comprise their personal network and who are amenable to doing direct work (and capable of it), if they wanted to do that instead of posting public job announcements. That pool, to the extent that it’s not already mostly doing direct work, is not going to be all that big.
Finally, our observation in reality is that EA orgs have been talking about how they’re talent-constrained for years. Maybe I’m wrong and this announcement will clear the market really fast! I’d love to see it. But, you know, just on priors, “not taking an obvious avenue to improve conversion rates” is a not a great sign.
I feel like this sort of entirely misses the point of the post. Obviously there are people who will go work somewhere for less pay than some of their alternative options. The question is not whether those people exist, it’s whether there are enough of them with the qualities that Lightcone (and other EA orgs—this is not a Lightcone-specific post) is looking for such that Lightcone can “have its pick” of extremely qualified candidates without spending an excessive amount of time and effort looking for them.
First, Lightcone (and other EA orgs) are not hiring from the general pool of software engineers. They are hiring from the pool of software engineers who are already mission-aligned, which constrains them to maybe, if we’re lucky, a low five-digit number engineers in the US, a substantial fraction of whom don’t live in the Bay Area and probably don’t want to relocate. If I’m wrong about how strongly EA orgs filter on whether their desired candidates are mission-aware and mission-aligned, that does render a substantial chunk of my concerns moot, yes.
Second, Lightcone (and certain other x-risk focused orgs) are not in the business of trying to optimize on labor costs; they’re racing against the clock and want to be wasting as little time as possible on things that don’t directly advance their research agenda like “looking for & hiring people”. If this isn’t true, my argument is that it should be true. Spending an extra month looking for a similarly qualified candidate (or, worse, settling for a less qualified candidate) is a cost that has a dollar figure attached to it. If you think we live in a world with short timelines, it has something worse than a dollar figure attached to it.
I’m sure Lightcone would probably not have much trouble hiring from within the pool of engineers that comprise their personal network and who are amenable to doing direct work (and capable of it), if they wanted to do that instead of posting public job announcements. That pool, to the extent that it’s not already mostly doing direct work, is not going to be all that big.
Finally, our observation in reality is that EA orgs have been talking about how they’re talent-constrained for years. Maybe I’m wrong and this announcement will clear the market really fast! I’d love to see it. But, you know, just on priors, “not taking an obvious avenue to improve conversion rates” is a not a great sign.