This seems really cool and impactful. I just wonder how it would be received by a public scared of vaccines. The yuck factor is very very high. But it's all natural except for a little generic engineering. Hard to say which way it falls.

It's how LLMs write. The tweet / article is written by an LLM and that's how it does.

"Three layers."

As a reader, I appreciate reading writing that lacks large amounts of spelling mistakes. Everyone agreeing on spelling seems like a useful monoculture, like driving on the same side of the road.

But I don't feel the same way about AI writing. It feels totally different in a way that good spelling does not.

Even if I liked the style, I would object strongly to that style quickly becoming a monoculture.

We're on a path to a style optimized for shallow attention maximization becoming the majority of text we read.

Interesting to think what could be if cephalopods raised their young instead of leaving them to completely fend for themselves. It would start intergenerational knowledge transfer, i.e. culture. Maybe selection pressure then trends towards group cooperation instead of going it solo.

I still think there would be huge barriers to "civilization" as I think you mean? (Do any of the apes have "civilization"?).

The real problem with cephalopods is their lifespan. For their age, they are almost as smart as humans, the problem is that they don't live past the age of 5 years.

I would argue that not having any overlap between generations is a bigger problem. It guarantees no accumulation of knowledge.

Agreed. If they were social enough to form large communities of unrelated families, it would also fix the generation overlap. But they don't do that either. They seem to be in a weird evolutionary dead end for intelligence.

Orcas do this already.

I mean there might be a already a civilization that is in the building that will peak 100k years later, and we just don't know about it.

having arms is probably an advantage in developing tool-use, but who knows!

I wouldn't be surprised if humans behaved the same way when playing the same game?

Like even if you brought me into a room and told me I was controlling "real nuclear weapons" I wouldn't believe you.

I think is an important point, and I don't see it mentioned in the article or the paper (though I skimmed the latter).

They are aware of what they are and how they are used. They're told to act as AI assistants. And there's theories of them being aware of their answers influencing their training.

So surely they must be able to reason that they're not literally controlling weapons of mass-destruction with their answers.

I understand your point, but I will say that many of the other policies also are not capable of being truly objectively enforced either ("I know it when I see it"). An effort could be made to define an approximate boundary.

Perhaps the moderator comment about "obvious smell" being killed is defining that line and the line allows lots of fairly obviously AI edited prose through. We've chosen a boundary and it's different from what I'd like (and think is best in the long run for humanity), but I will accept it and move on.

What would they do if the natural birthrate were to tip it over the threshold? (Perhaps unlikely at current birthrates, but given that laws last long times, perhaps worth considering?)

The initiative does not mention what. Because the initiative's sponsor are anti-migration.

Yeah I'm aware -- just pointing out that this threshold could be reached by other means and the options for enforcing it if it happens via birthrate are all pretty grim.

For the record, I'm not against the use of AI during the writing process. It can be used to make the final product better without eliminating the human authors style. I am against everything I read turning into one monoculture of AI produced prose, even if humans put some thought into what the prose is communicating.

Even if I thought that the style of AI produced prose was good, I still wouldn't want it to be the only style of writing I ever read unless I open a book written before 2022.

HN is big enough that its standards could propagate beyond these walls.