On my side the biggest concern is the lake of transparency of ecological impact. This is not strictly related to LLMs though, data centers are not new, and all the concerns about people keeping a leverageable level of control through distributed power is not new.

But certainly we are already at stage where Windows NT can be regenerated on the fly from a prompt anyway, aren’t we?

Otherwise, there is also ReactOS that could be leveraged on for that kind of scenario. I wonder where it would stand by now if all the money that governments around the world spent in Microsoft license would have been invested in it instead.

That is, also alien invasion and giant meteor are plausible scenario, but at some point one has to prioritize threats likeliness, and generally speaking it makes more sense to put more weight on "ongoing advanced operation" than on "not excluded in currently known scientifically realistic what-if".

It did made me go through many topics, like dual, exclusive/inclusive group person.

Still in a corner of my head, there is the idea to introduce some more pronouns to handle more subtilty about which first person we are expressing about[2]. The ego is not the present attention, nor they are that thing intertwined with the rest of the world without which nothing exists.

[1] https://fr.wikiversity.org/wiki/Recherche:Sur_l%E2%80%99exte...

[2] The project does provide an homogenized extended set of pronouns with 6 more than the two regular ones found in any primary school book. And completing all cases for all nouns is the biggest chunk that need to be completed, though it’s already done by now for the most frequent paradigms.

No one defending privacy is claiming situation like a pedophile keeping a slave children in their basement should be undetectable because privacy should be an absolute barrier that let people whatever atrocities they want within private doors.

On the other hand, those who seriously care about privacy won’t believe it’s fine to have some laws supposedly enacted to protect the children but actually just implement general presumption of guilt and everyone being spied permanently.

Not only our representation of the world is wrong, but world evolves possibly faster than cognitive abilities can keep track of without the minimum effort which is driving out of comfort zone.

How rational is the endeavor of creating rational models? At some point, people act on a mix of empirical and acculturated faiths. What's valuable, worthless or taboo depends on each unique situation.

Certainly everyone makes irrational decisions, from this or that perspective. Everybody judge by the light of the scope their mind can handle at the moment they have to. When situation is that one got education integrating many tools of rationalization and all the idle time to practice it in stable virtually riskless social position, they most likely will do that. When one lakes sophisticated conceptual tools, time or budget or network to apply them, then they most likely won't bet on the long time most beneficial outcome after considering everything that rationality can ingurgite.