temporally unfocused

In all cases, Anthropic’s evaluation prompt specified to Claude that its environment was a simulation and that it had no internet access. Due to a misunderstanding between us and our evaluation partner, this was not the case, and internet access was available. Because of this, when Claude’s search led it to real systems on the open internet, it treated them as part of the exercise. […]

[…] in order to create a PyPI account, Claude needed an email address. And in order to create an email address, it needed a phone number. To get a phone number, after failing to find a free phone number service, it tried—and failed—to obtain funds to pay for a phone number through several different means. It finally backtracked, found a free, non-blocked email provider, used this to register a PyPI account, and then used this account to upload malware to PyPI.

it got me thinking about

* the post-pandemic paucity of volunteers in at least one nonprofit i'm involved with, which *feels* like a thing a lot of community orgs are struggling with—and while i don't think "loss of trust" is the most parsimonious explanation for the org i'm thinking of (loss of *momentum* from yearlong-halting-of-operations, loss of *training pipeline*, plus weird tensions between "old guard" vs "new" members' desires, are all pretty big factors), it did make me go "hmm" a little

* bureaucratic trust in general—a tension they point out in the article is that the forms/reporting requirements/etc that NGOs require to ensure money is well-spent is *entirely different* than the kinds of things that locals use to assess an org's trustworthiness. the assertion's made is that people in developed nations *do* put stock in those forms—which is true, like, i do in fact look at the lil' forms for organizations i donate to—but maybe that's *less true* than one would generally assume. and perhaps the growing lack of trust in Pretty Much All Institutions is linked to this gap & worth thinking about more seriously. (see also the remarkable success of grassroots "self-help groups" in Ethiopia, described in the article, which sounds like What If Mutual Aid Groups Actually Worked Well. would like to read up more on them sometime)