I would prefer nitrous oxide or xenon gas over argon, since those will make the pigs feel actively good as they are killed. But imo argon would be better than electric stunning, since it is likely just a slow drift out, whereas stunners could be painful.
Obviously, ideally no pigs would be killed without dire need, but that's not the world we live in.
Random thought, on what the "next chicken" could be: if someone was able to commercially grow porcini mushrooms, and preserve them such that they could be sold fresh.
They are really delicious, I had some freshly cooked on a mushroom foray and it was one of the best things I've ever tasted. I used to love meat before I went vegan, and I would put it up there with a great steak or great ribs.
I could also see them being healthier than chicken, but someone would have to actually check that.
But they go bad really fast and it's tough to keep worms out, and as I understand they're tough to grow.
This was super helpful and clarifying! Especially alongside @Yay Meristinoux's comment and your response. And I agree with @Toby Tremlett🔹 that this format worked much better than the usual undisclosed claude, normally it's tough to read but this was very clear and not jarring.
I don't really have much to add, I think you've shown to my satisfaction that it's a lot harder than I thought. It does make me wonder how traditional/for-profit companies deal with this issue tho, do they just bite the bullet of hiring less experience people in the hopes that eventually they will be net positive more often? Or do they just generally make so much more money that it's worth it for them to absorb more net negative (for the specific role) people?
On it being surprisingly hard, do you have a rough ballpark estimate of the number of hires you've seen that ended up being net negative?
If 0% of hires were net negative, I would guess that a lot of value is being left "on the table", while if 100% of hires were net negative, how is the org even still around lol? So the sweet spot is somewhere in between, and obviously it sucks to go through hiring someone who's, to be blunt, a dud, but it seems like more can get done in an org with a few superstars, a ton of solid folks and a few duds than an org with a few superstars who are kept from their potential cause they have to do work that probably even duds could handle if needed.
edit: another thought, have you considered hiring more with strict firing policies in the first, idk, 6 months? Or first year? Probably leads to a bad culture but idk
Just read the substack comment, I think that does clarify things, but my complaint isn't that lots of people get rejected, it's that people who seem like great fits even if they're not perfect fits get rejected and then the role gets left unfilled.
Yeah it's super important to be clear about what you're looking for and minimum requirements. I think many orgs have a bad habit of saying "even if you think you may not be a good fit, please apply!". They're targeting the amazing candidate who's anxious or has imposter syndrome, without considering the huge group of people who accurately think they shouldn't apply, but end up spending time on an application for something they had no chance of getting.
Hmm, on your question about what I'm unhappy with, in part I think that some bar hirers have unrealistic expectations of talent. There's the meme of a job application looking for 6 years of experience in a technology that's only been around for 3. And I think in turn, having the bar too high means that it is actually problematic to do bar hiring. In theory you could hire the perfect candidate who's 10x better than everyone else, and get 10x the value per dollar, but if refusing to hire anyone less than that means everyone else has to do what they'd be doing and miss other opportunities, then you probably end up worse off than just hiring the 1x person. There are probably situations where you might be better off hiring the 0.5x person, but imo the number of very smart, talented people trying to get into EA makes me think that ~5x people are getting missed due to the search for the 10x person. And I get it, we want to maximize the value per dollar, it's "EA" not just "A", but it's also important that we not turn it into a red queen race. I don't actually know what the best solution or tradeoff is either, and it probably varies a lot per situation.
Sorry just realized I should've read you substack comment before writing all this 😅 I'll take a look at that!
Love this article, it's such a solid point but I see it rarely discussed. We are trying to tell AI "do as I say, not as I do". Ask any parent how well that works.
Minor nitpick: "This is the exact principle that superintelligent AI cannot learn" maybe use "must not" instead of "cannot"?
I'm biased in that I found my current partner on veggly and she's great, but I would seriously recommend using it. Yes, it's pretty dead, but that's also kind of a good thing. Sort by new, check it once a week, and expand your surface area for luck, especially if you'd be aligned with someone interested in veggly enough to try it out
Sometimes people will say "two things can be true at once" in the context of talking about mutually exclusive things that can't both be true.
But sometimes they say that about things that seem mutually exclusive but actually can both be true.
And sometimes the literal reading might be mutually exclusive, but the point still stands.
And sometimes it's just said cause people like phrases about two things that are both obviously true and entirely compatible.
I would prefer nitrous oxide or xenon gas over argon, since those will make the pigs feel actively good as they are killed. But imo argon would be better than electric stunning, since it is likely just a slow drift out, whereas stunners could be painful.
Obviously, ideally no pigs would be killed without dire need, but that's not the world we live in.
Random thought, on what the "next chicken" could be: if someone was able to commercially grow porcini mushrooms, and preserve them such that they could be sold fresh.
They are really delicious, I had some freshly cooked on a mushroom foray and it was one of the best things I've ever tasted. I used to love meat before I went vegan, and I would put it up there with a great steak or great ribs.
I could also see them being healthier than chicken, but someone would have to actually check that.
But they go bad really fast and it's tough to keep worms out, and as I understand they're tough to grow.
This was super helpful and clarifying! Especially alongside @Yay Meristinoux's comment and your response. And I agree with @Toby Tremlett🔹 that this format worked much better than the usual undisclosed claude, normally it's tough to read but this was very clear and not jarring.
I don't really have much to add, I think you've shown to my satisfaction that it's a lot harder than I thought. It does make me wonder how traditional/for-profit companies deal with this issue tho, do they just bite the bullet of hiring less experience people in the hopes that eventually they will be net positive more often? Or do they just generally make so much more money that it's worth it for them to absorb more net negative (for the specific role) people?
On it being surprisingly hard, do you have a rough ballpark estimate of the number of hires you've seen that ended up being net negative?
If 0% of hires were net negative, I would guess that a lot of value is being left "on the table", while if 100% of hires were net negative, how is the org even still around lol? So the sweet spot is somewhere in between, and obviously it sucks to go through hiring someone who's, to be blunt, a dud, but it seems like more can get done in an org with a few superstars, a ton of solid folks and a few duds than an org with a few superstars who are kept from their potential cause they have to do work that probably even duds could handle if needed.
edit: another thought, have you considered hiring more with strict firing policies in the first, idk, 6 months? Or first year? Probably leads to a bad culture but idk
Just read the substack comment, I think that does clarify things, but my complaint isn't that lots of people get rejected, it's that people who seem like great fits even if they're not perfect fits get rejected and then the role gets left unfilled.
Yeah it's super important to be clear about what you're looking for and minimum requirements. I think many orgs have a bad habit of saying "even if you think you may not be a good fit, please apply!". They're targeting the amazing candidate who's anxious or has imposter syndrome, without considering the huge group of people who accurately think they shouldn't apply, but end up spending time on an application for something they had no chance of getting.
Hmm, on your question about what I'm unhappy with, in part I think that some bar hirers have unrealistic expectations of talent. There's the meme of a job application looking for 6 years of experience in a technology that's only been around for 3. And I think in turn, having the bar too high means that it is actually problematic to do bar hiring. In theory you could hire the perfect candidate who's 10x better than everyone else, and get 10x the value per dollar, but if refusing to hire anyone less than that means everyone else has to do what they'd be doing and miss other opportunities, then you probably end up worse off than just hiring the 1x person. There are probably situations where you might be better off hiring the 0.5x person, but imo the number of very smart, talented people trying to get into EA makes me think that ~5x people are getting missed due to the search for the 10x person. And I get it, we want to maximize the value per dollar, it's "EA" not just "A", but it's also important that we not turn it into a red queen race. I don't actually know what the best solution or tradeoff is either, and it probably varies a lot per situation.
Sorry just realized I should've read you substack comment before writing all this 😅 I'll take a look at that!
I'm not very happy with arguments about how talent constrained orgs are compared to how many talented people apply and get rejected. See also this comment by Scott Alexander https://forum.effectivealtruism.org/posts/B6d8Wzk4gNzHsXvdi/ai-safety-is-extremely-bottlenecked-on-grantmakers?commentId=n2Rd7RR4y9EnKP42F
Love this article, it's such a solid point but I see it rarely discussed. We are trying to tell AI "do as I say, not as I do". Ask any parent how well that works.
Minor nitpick: "This is the exact principle that superintelligent AI cannot learn" maybe use "must not" instead of "cannot"?
I'm biased in that I found my current partner on veggly and she's great, but I would seriously recommend using it. Yes, it's pretty dead, but that's also kind of a good thing. Sort by new, check it once a week, and expand your surface area for luck, especially if you'd be aligned with someone interested in veggly enough to try it out