Ahh, interesting argument! I wasn’t thinking about the argument that these firms might (e.g.) slightly accelerate economic growth, which might then cause an increase in x-risk (if safety is not equivalently accelerated). In general I feel sufficiently unclear about such considerations—like maybe literally 50:50 equipoise is a reasonable prior—that I am loath to let them overwhelm a more concrete short-term impact story in our cost-benefit analysis, in the absence of a clear causal link to a long run impact in the opposite direction, as you suggest in the article.
In this case I think my argument still goes through, because the claim I’m objecting to is so strong—that there is in some sense a >50% probability that every reasonable scenario has all three firms being negative.
Ahh, interesting argument! I wasn’t thinking about the argument that these firms might (e.g.) slightly accelerate economic growth, which might then cause an increase in x-risk (if safety is not equivalently accelerated). In general I feel sufficiently unclear about such considerations—like maybe literally 50:50 equipoise is a reasonable prior—that I am loath to let them overwhelm a more concrete short-term impact story in our cost-benefit analysis, in the absence of a clear causal link to a long run impact in the opposite direction, as you suggest in the article.
In this case I think my argument still goes through, because the claim I’m objecting to is so strong—that there is in some sense a >50% probability that every reasonable scenario has all three firms being negative.