115 karmaJoined


On my reading, "human-level automated alignment researcher" means a system that is human-level at alignment research, but not AGI. You can take the position that in order to be human-level at alignment research, it will need to be AGI, but I don't think that's necessarily true, and in any case it's certainly not obvious. For myself, I keep being surprised at how capable systems can get at particular abilities without being fully general. (Years ago I wrongly believed that AGI would be necessary for artificial systems to reach the level of language capability they have right now; back in the 70's, Hofstadter wrongly believed AGI would be necessary for superhuman chess ability; etc.)

I was thinking for example of my energy and water bills, which both list various ways that one can pay. Looking at my energy bill now, I see that bank transfer is listed as the first option, followed by debit/credit card (if one calls on the phone), cheque, and finally direct debit. I find bank transfer the easiest for usage-based utilities, as I prefer to pay exactly what I owe rather than make estimated monthly payments as I'd need to do with direct debit... I'm a little paranoid about whether I will be refunded in a timely manner for overpayments, or if I will be hit with late fees if the estimate is too low. I didn't actually realize about the extra protection afforded in the case of direct debits, though, thanks! Good to know.

My invoices for accountancy services also list bank transfer as an option, though they also allow other ways to pay... I agree it would be weird for a big company to have bank transfer as the only option. Private contractors (repairmen etc.) have sometimes just given me an invoice with their bank details only.

Distrust any attempt to get you to make a bank transfer, you have less recourse when it comes to reversing these compared to a card payment.

In my experience living in the UK (8 years), bank transfers are extremely commonly used by small and large companies alike, and even for interpersonal transactions (say, paying your half of the bill to an acquaintance who has picked up the bill at a restaurant that wouldn't split the check), just because of the convenience and speed of them. Every utility and rent bill I've ever gotten has included bank transfer details as a payment option, sometimes as the only payment option. It's true that there's usually no recourse to reversing such a transfer, so obviously you need to exercise reasonable caution, but "distrust any attempt to get you to make a bank transfer" seems a bit much.

In addition to the 80,000 Hours career guide already mentioned, it's worth noting that they can schedule a free 1:1 call to help you think through your options, help you work out next practical steps etc.: https://80000hours.org/speak-with-us/

I recently filled out the Airtable form, but was surprised to see when I got my e-mail receipt that many of the answers I provided did not appear.

How would you suggest that I and others affected by this proceed? Thanks!

[edit: extraneous information removed]

Folks employed at universities might consider theconversation.com. Pieces published there are sometimes picked up by other outlets. They especially like it if your proposal links into something currently in the news. They prefer being pitched first but I have succeeded in the past with a full piece ready-to-go. They will expect you or a coauthor to be an 'expert' in the topic you are writing on, but this is broadly construed; I would just fill out my profile in a way that shows you have some academic background relevant to the topic.

sidenote: There has been an argument that 'radically transformative AI' is a better term for the Industrial Revolution definition, given the semantic bleaching already taking place with 'transformative AI'.

Good thought, thanks. The "Work as Meaning Inventory" (https://journals.sagepub.com/doi/full/10.1177/1069072711436160 , items in Table 1) might be relevant, as it also has elements of 'what do you prefer' as well as a 'greater good motivations' subscale.

Load more