Where do you stop once you go down this rabbit hole? Which faith(s) get their views injected in? Christian? Muslim? Hindu? Pagan Gods? Should I get the perspective of the follower's of Thor when I ask a question?

Note that you can always ask for the religious perspective you're interested in. IME with the religion(s) I grew up in or know a great deal about, the LLMs are pretty good at answering accurately and respectfully. Nearly all the products already offer you tools to personalize the output for you too if you want to inject your faith into the answers, so it's not like the LLMs won't give you a religious perspective if you want it.

Look for the followup article where they wrap the mundane observation around the politics drumbeat (people who use AIs may be exposed to political beliefs you don't believe in) once they notice the article's view count.

Is certainly a typo, off by an order of magnitude. 600 years ago the Christian Church had already been around for more than 1000 years. Young earth creationists believe the earth is ~6000 years old

> [...] LLMs rely on internal universes derived yet decoupled from reality. Religions that deify their interpretation of their scriptures instinctively know this model and how to use it.

I stopped reading at this sentence. If you go to the source (https://www.vatican.va/content/leo-xiv/en/encyclicals/docume...), you can see it's available in eight languages, none of them being Latin. In fact, I read elsewhere a few days ago that one of the novelties of this one is that, unlike all the preceding ones, it's not written in Latin; the Wikipedia article (https://en.wikipedia.org/wiki/Magnifica_Humanitas) also says that ("The encyclical was the first to be published without an official Latin version. This followed a recent change to Vatican regulations permitting such documents to be drafted in other languages.[4]").

If the article gets it this wrong already in the third paragraph, it's not worth reading any further.

Everything that does not contain the universe itself has to do that.

Have you ever seen a cat and then realised it was just a plastic bag. That;s you synchronising your internal model to updated observations.

There's a cheap shot at Anthropic suggesting that AI might be more than the sum of its parts.

There are patterns to science and patterns to religion. To delare something is or is not is the domain of religion. Science always gives you something that is incomplete, it's the best we know right now, it corresponds to theories and tests.

It is a very scientific thing to say you could possibly have a conscious AI. Because the possibility exists through not being convincingly ruled out. It is possible that matter can produce consciousness because we are made of matter and if we call the thing we have consciousness then that suggests things made of matter can be conscious.

It is also scientific to analyse models to see what they represent and to present the possibility that their representations might correspond to an experience.

None of that is says that is definitely the case.

Once you get to saying 'It's obvious that A LLM could never produce consciousness' or 'they are not really thinking' or 'they don't understand anything' you are talking in the absolutes that only religion can offer.

Science will frequently state 'There is no evidence of ...'. It is frequently entirely reasonable to act on the presumption that something doesn't exist in that situation, but that is not the same thing as refusing to accept that it could. If evidence turns up science will change its view, if it does not, then it fails to be science.

You see poor science exhibited in off-hand dismissals of perpetual motion machines. If someone presents a claim that they have made a perpetual motion machines, you don't respond, as many do, by saying it is against the laws of thermodynamics and therefore the machine doesn't work. That is applying a faith in the discoveries that are more than they say, you are treating science like a religion by doing that. A scientific response to a perpetual motion machine is to say "I expect you are wrong in some way, because well tested rules suggest this cannot happen" you may then ask about various implications working from the assumption that the machine does work. If they cannot show any of those implications actually happening it suggests that the machine doesn't not work (basically you can ask "where does the energy come from'). If they can show solid proof of the implications, good science would then accept this as a counterexample to the laws of thermodynamics. That has yet to happen, the laws seem fIrly robust. But a perpetual motion machine does not fail to work because of the laws of thermodynamics, it fails to work because of the way the universe works. The laws of thermodynamics are just a very good model of how it works. If the perpetual motion machine worked. It would be the laws, not the universe itself that was incorrect.

Science allows for possibility, and searches for evidence.

Religion just states what some want to believe is true.

In the same way, outright declarations of the capabilities of AI now and in the future are the Domain of religion.

The scientific view is, as always, "here's what we have evidence for, so we don't think this particular AI is conscious because it doesn't show what we would expect for that." There is no absolute here. If something unexplainable with the current system comes to light then science must adapt, that is what makes it science.

They say it's because "LLMs rely on internal universes derived yet decoupled from reality" but so do humans and my cat. I don't think those are religions.

I mean, sure. Some communities, like the Amish are pretty famous for that attitude. It is a respectable option, but the practical upshot is it just means there are a bunch of jobs you can't do. There is a bonus that you get to be part of a pacifistic community because your military branch isn't going to be in a position to defeat anyone.

So, exactly like fundamentalist religion.

A classic interview question used to be "where do you see yourself in 5 years?". What I'd like from the tech industry (and maybe also from the US government) would be an answer to "where do you see humanity in 5 years?" - because right now, I see lots of hype and panic, but little explanation what kind of vision should be realized with AI. Well, except overtly dystopian visions that could come right out of Marx' writings or a William Gibson novel. But even those aren't really self-consistent.

I see only benefits. If a creator exists, they will come back to reignite the spark of evangelism like they "did" before. This might give us a bunch of new stories and an update of the bible to match modern life. If a creator does not exist, then we are freed of all doubt and can continue with our lives without worrying about ending in hell, etc.

Let’s rewind:

- Doomsayers with Yudkowsky et. al. at the helm calling to bomb datacenters

- “Roko’s Basilisk”

- General AI psychosis and treatment of AI as something with consciousness. Any attempts of rebuttal are met with lazy “but how do you know you are conscious. haha got ‘em”

- Inquisitorial purges of non-believers: “our CEO mandates you shall use AI in your work, or else”

- Communion with a God: “I was talking to ChatGPT and I realised <something personally profound>”

- “You are just using it wrong” aka “You are not praying hard enough”

- Cult Executive Officers

- Promise of incoming salvation with AGI taking over and making a Heaven (or Hell - caveat emptor) on Earth

-----

You may not want AI used as a proselytizing pipeline into home and office. Others do.

In many ways, AI is a religion. Not because it requires belief in a utopian future through a dystopian present, or that it’s used by very powerful people to get more power, or that nobody can define what it actually is. All these things are overlaps on the Venn diagram, but the biggie is that LLMs rely on internal universes derived yet decoupled from reality. Religions that deify their interpretation of their scriptures instinctively know this model and how to use it.

-----

is not actually a coherent train of thought, although you can see why it might seem to be to a language model.