AI Affirms The Necessity Of God...!

[Crossbreed]

I Convinced Grok God Is Real (Using Math, Science, and Logic)

 

[Jumpinbare]

While I am convinced of God's existence, I am not impressed with an AI's "opinion". They are simulating how a large number of people would react based on training data, not on actual logic.

 

[Crossbreed]

The limiting criteria was not popular opinion, but mathematics, observable science & mathematical probability.

 

[Outdated]

Funny, last week it was the reincarnation of Hitler. What's it going to be next week?

 

[Crossbreed]

Funny, last week it was the reincarnation of Hitler. What's it going to be next week?

1. Acknowledging God, alone, does not confer innocence.
   
   "You believe that there is one God. You do well.​
   Even the demons believe—and tremble!" James 2:19​
2. Grok is shown to rescind that conclusion toward the end of the video when said parameters are lifted.

 

[Hypnalis]

The limiting criteria was not popular opinion, but mathematics, observable science & mathematical probability.

It's impossible to prove the existence of anything supernatural using science.

But it's easy to lie to an AI while it's being trained.

A common example of a self-cancelling claim.

A simple question: what is the probability that there is intelligent life (using homo sapiens as a yardstick for intelligence) somewhere outside our star system?

That question cannot be answered. Probability is defined around using observations of a sample (usually a subset of something larger) to make inferences about something else (usually a superset of the sample).
There are conditions on the nature of the sample, and how large it must be.

We have only one observation to base out inference on - by my definition, we know there is one case of "intelligent life".

Probability cannot be used to project that onto the universe. It makes no difference that we now know there are many other planets. We have only one observation (and it's the same one) to make inferences about the likelihood that a planet will "produce" intelligent life.

Guesses are possible, and the web is full of them. Some, made by people who use reasonable means to make their "guesstimates", are very interesting.
But anyone who claims that their guess is actually an inference based on math is ignorant, an fool, or a liar.

This is a very convenient "reality check" though.
False claims based on "probability" with no or an invalid sample instantly bounce the claimant's reality check.
:
:
BTW: the one sample/one exception rhetorical trick (used to undermine valid heuristics) isn't the same as inference based on mathematical techniques.
Some claims (commonly in math) can be disproven with a single counter-example.

Heuristics don't claim to be right all the time. They are "soft" rules that are true most of the time.
So "one exception does not void a heuristic".

But you still can't construct a valid heuristic about extra-solar life.

 

[Crossbreed]

It's impossible to prove the existence of anything supernatural using science.

The video did not claim to prove God's existence. It just agreed that He was a (more) viable explanation for the directed, intelligent design (that we can observe).

But it's easy to lie to an AI while it's being trained.

The AI's default training was evolution, not creation. It even said so.

 

[Hypnalis]

The video did not claim to prove God's existence. It just agreed that He was a (more) viable explanation for the directed, intelligent design (that we can observe).

We cannot "observe intelligent design". There is no evidence for it.

We can believe it. I don't OFC, but I have nothing against the kind of thinking, belief, and behavior that leads to it.
But you can't bridge from a belief into a real-world fact.

What we can observe is evolution, though it's not easy due to the time scales. But it works as expected with viruses , which change fast enough to be observed (**).
The evidence is very solid now. And it's highly relevant that the theories are confirmed by what we've learned from genetics, which came long after the theory was demonstrated through large-scale real world observations.

(**)
There's a common "trick-question" about the development of new species, which isn't the same as genetic changes as observed in e.g. the C19 virus. The data is solid, but the time scales are huge.
Applies to humans: the difference in time between us and out "nearest relative" (Chimps) is 7 million years.

AFAIK even with viruses we can't observe the development of a new species (or the equivalent in viruses - they're weird). I might be wrong though.

The AI's default training was evolution, not creation. It even said so.

It's easy to lie to an AI. And it's easy to get an AI to provide a false answer.

What's not possible though, is for them to "think" - they are not sentient.
If an AI "finds god", it got it from data it was trained with.

 

[Crossbreed]

We cannot "observe intelligent design". There is no evidence for it.

Did you watch the video?

 

[Hypnalis]

The first sentence was enough. .

AI: .. "despite their logical and probabilistic weaknesses when scrutinized rigorously".

AI's don't "scrutinize" - they organize data in a very clever way. There are no "logical or probabilistic weaknesses" in evolution.
That came from data it's been trained with (or (much less likely IMO) instructions to lie).

BTW - I meant what I said above about not having any objection to counterfactual beliefs. 50% of the beliefs in any culture are counter-factual - in terms of quantity, a few more or less won't hurt.

On the other hand I'm opposed to using counterfactual beliefs to interfere with real-world knowledge.

So I distinguish between "This is what I believe", and "My beliefs disprove established facts, so you have to stop using them".

With, OFC, the false "grey area" argument of "it's just a different explanation, but you can keep using the facts".
Science is all about cause and effect, and building on existing knowledge: disrupt the cause/effect links and you block development.

I suggest we stop here (or soon).

I don't object to your beliefs or the existence of the video, nor would I care if everyone else on AF shared your beliefs. But the claims above (in italics) aren't neutral. Using them invites dissent.

 

[Crossbreed]

@Hypnalis, our faith does not rely on AI's testimony, but it is a telling contradiction.

 

[Gerald Wilgus]

I only need to know that AI is subject to hallucinations and no AI engine has ever passed the Montreal Cognitive Assessment as being other than cognitively impaired.

 

[Mary Terry]

I only need to know that AI is subject to hallucinations and no AI engine has ever passed the Montreal Cognitive Assessment as being other than cognitively impaired.

I've read that AI cannot beat a human at playing checkers. Not sure it's true, but I take everything "AI" with a huge grain of salt.

 

[Hypnalis]

@Hypnalis, our faith does not rely on AI's testimony, but it is a telling contradiction.

What contradiction?

An AI cannot prove the existence of any supernatural entity.

But like a human, it can be told to claim that it has proof.

 

[Hypnalis]

I've read that AI cannot beat a human at playing checkers. Not sure it's true, but I take everything "AI" with a huge grain of salt.

It's correct to be cautious. AI already can do some good things, and it will have huge economic effects( by replacing new classes of workers). But it doesn't "think".

I certainly wouldn't use an AI that is good at predicting the next word in a line of text (this is essentially what the most readily available ones do, and it's more powerful than it sounds) to play checkers.

OTOH AI's (tech akin to e.g. the text- and graphic-centric ones, but with some game-specific features) are much better than the best human players of Chess and Go/WeiChi/Baduk, which are far more complex games.

Stockfish (chess) - Wikipedia
AFAIK a standard desktop PC running Stockfish is easily strong enough to beat any human.

 

[Crossbreed]

What contradiction?

The contradiction that I am referring to is

1. This is true.
2. #1 cannot be true.

I am sorry if I was unclear.

 

[Gerald Wilgus]

I've read that AI cannot beat a human at playing checkers. Not sure it's true, but I take everything "AI" with a huge grain of salt.

The evidence for that is the meta "review" that RFK jr. released concerning vaccination that was generated by AI. It created citations that were entirely bunkum. Something like that I find to be completely dangerous because people think that AI cannot be a facilitator of misinformation.

 

[Crossbreed]

But like a human, it can be told to claim that it has proof.

Is that what you think happened in the video?
(That was stated outright in its initial response.)
I thought that Smith was being very Socratic.

 

[Dagan]

Yup. One core fear of how AI will eliminate us all is expressly because they will misinform us - they will understand what saves us or keeps us healthy and alive, and they will make us believe / do the opposite.

If anything, they'll do it just to cull the lower IQ from our population and keep the others for their maintenance needs.

Should any of this become reality, I can envision requiring an apocalyptic event that crashes power grids everywhere and even creating tsunamis to fry out further power plants.

If you really want to go down a rabbit hole, look online at theories that this has happened on Earth before already.

 

[Hypnalis]

The contradiction that I am referring to is

1. This is true.
2. #1 cannot be true.

I am sorry if I was unclear.

I remain mystified.
Technically that's not a contradiction unless you claim both are true, or neither is true.

Even if you manage it (I recommend a trip to wikipedia), you have a relevance deficit.