This blog post is about a conversation I had with Scholar GPT on cosmology and the existence of God. I’m interested to hear what others say about the scientific data points I use to make the argument and the logic I use to connect the dots. Scholar GPT seemed to accept both. I am also interested to hear what people think of the new research ideas proposed under a Theistic framework. I believe this would open up research to new ideas and new learning. What do you think?
1 Like
To the contrary, your encounter with Scholar GPT shows that there is no scientific evidence for the existence of God. Yes the universe we live in began 13.8 billion years ago and will expand for trillions of years until a heat death. But the finite existence of our universe is not evidence for the existence of God. Science is neutral on the existence of God and proceeds without the assumption that God exists.
*and proceeds without the assumption that God does exist
7 Likes
Hi Ron
Thanks for posting this.
I completely agree that methodological naturalism can lead to inaccurate models in certain cases especially around models that try to explain origin events such as some of the ones scholar GBT listed.
The suggestions made are unsurprisingly vague and lacking in concrete detail. There’s nothing really actionable there. The problem of producing testable hypothesis from the idea of God is a major issue - and the reason why there is no significant scientific evidence for God.
But you really can’t expect much from an LLM on this topic. It’s a product of AI-related research but it is not actual AI.
2 Likes
You know, whenever – and it is not often – I am tempted to try to gain insight into a question that bridges the gaps between philosophy, science and theology by consulting some sort of text-generating oracle incapable of actual cognition, I just reach for the good ole Magic 8 Ball. It never fails.
3 Likes
Me: Vilenkin’s model has not been widely accepted because it cannot be tested.
Me: … I think we should begin by attempting to fit more of the data we already have into a theistic framework to see what new insights might come from that.
So the reason for not accepting Vilenkin’s model is because it cannot be tested?
Wait a moment whilst I count the number of testable theistic frameworks:
Zero, … … … [sound of crickets chirping]
So what we have discovered is that Scholar GPT, whilst very knowledgeable, isn’t too bright.
“Christian Apologetics – good enough to fool a dim LLM” – has a nice ring to it. 
2 Likes
Interesting! BUT you should consider that LLM’s are a sophisticated sort of autofill, and in that sense that will tend to fill in the best prediction it can make of what you want to see.
After some interaction you asked …
Do you want to guess what conclusion I draw?
And GPT happily filled in the blank, correctly interpreting the conclusion you already hold.
The point being, this is a fun exercise, but a flawed sort of inference. I think we could start with the same initial proposition and lead GPT down the path to conclude God does not exist. It might be interesting to change the discussion at various points to see how that influences GPT’s conclusions. For instance, at the point …
Scholar GPT: Yes, I agree. Vilenkin’s model, while theoretically interesting, cannot be tested due to the impossibility of replicating or observing conditions of “no matter, no space, and no time.” Embracing such untestable hypotheses goes against the principle of empirical testability in science, as emphasized by Newton’s “Hypotheses non fingo.”
What happens is we instead ask about empirical testability of God? I’m not suggesting we actually do that, only that we would likely see different responses from GPT.
4 Likes
From the cited article found in Free Thinking Ministries:
Me: Do you agree that when Nature does not exist, then natural explanations are impossible?
Scholar GPT: Yes, I agree.
But, of course, nature does exist. It exists in abundance. Paraphrasing Jean Paul Sartre’s masterwork, Nausea nature is all around us, it is in us, it is us. It births us and suffocates us and, ultimately, destroys us.
I hear these inane arguments on occasion. In fact, to call them inane is overly generous. Thinking that you are going to find your God by “conversing” with a machine is tantamount to a lunatic howling at the moon, with one important exception, the Moon knows better not to howl back…
1 Like
You know Mattel still makes them…
1 Like
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.