“How to build a truth engine”

[drumdude]

The following important film is currently showing at SXSW:

While political polarization and COVID rage across the world unleashing even wilder conspiracy theories, misinformation, and fake news, several teams of experts have set out to fight back against the rise of information warfare that threatens our democracies. We follow award winning investigative journalists, fact-checkers, scientists and engineers in their efforts to develop groundbreaking technology and understanding of the human psyche to help bring societies back to a common understanding of reality.

https://m.imdb.com/title/tt12026606/


That last sentence, to me, is key. The big problem with Mormonism is that it simply conflicts with the common understanding of reality. The community of Christ has shown that it is possible, even if it’s painful, to return back to reality.

I am grateful for those who dissect Mormon misinformation, but it is an enormous amount of work. The backyard professor, RFM, John Dehlin, etc all take hours and hours to debunk apologists and church leaders.

The possibility that A.I. can be leveraged to point out all the flaws in an Interpreter article, for example, could be truly miraculous.

[Failed Prophecy]

That last sentence, to me, is key. The big problem with Mormonism is that it simply conflicts with the common understanding of reality. The community of Christ has shown that it is possible, even if it’s painful, to return back to reality.

You do realize that every organization distorts reality in their own way, including the CoC?

I am grateful for those who dissect Mormon misinformation, but it is an enormous amount of work. The backyard professor, RFM, John Dehlin, etc all take hours and hours to debunk apologists and church leaders.

Not really, it's all variations on a theme. There's very little that is new in Mormonism that requires fresh research to debunk.

The possibility that A.I. can be leveraged to point out all the flaws in an Interpreter article, for example, could be truly miraculous.

You seem to have a very misguided view of what A.I. can do. A.I. literally has no concept or truth nor of facts nor of reality. It only recognizes patterns and can give you statistical probabilities of patterns repeating or being similar to other patterns. Even LLMs, the current fad, just make answers based on what is statistically likely to follow sentences that have already appeared in its corpus of texts. Your prompt is just the entry point into a chain of statistical reasoning and the "answer" is just the next chain in that set of probabilities.

The way you talk about A.I. seems more akin to the Mormon view of the Holy Ghost. Maybe go to church on Sunday to experience that truth engine? I won't be joining you (on any Sunday), but maybe it will be beneficial?

[drumdude]

That last sentence, to me, is key. The big problem with Mormonism is that it simply conflicts with the common understanding of reality. The community of Christ has shown that it is possible, even if it’s painful, to return back to reality.

You do realize that every organization distorts reality in their own way, including the CoC?

I am grateful for those who dissect Mormon misinformation, but it is an enormous amount of work. The backyard professor, RFM, John Dehlin, etc all take hours and hours to debunk apologists and church leaders.

Not really, it's all variations on a theme. There's very little that is new in Mormonism that requires fresh research to debunk.

The possibility that A.I. can be leveraged to point out all the flaws in an Interpreter article, for example, could be truly miraculous.

You seem to have a very misguided view of what A.I. can do. A.I. literally has no concept or truth nor of facts nor of reality. It only recognizes patterns and can give you statistical probabilities of patterns repeating or being similar to other patterns. Even LLMs, the current fad, just make answers based on what is statistically likely to follow sentences that have already appeared in its corpus of texts. Your prompt is just the entry point into a chain of statistical reasoning and the "answer" is just the next chain in that set of probabilities.

The way you talk about A.I. seems more akin to the Mormon view of the Holy Ghost. Maybe go to church on Sunday to experience that truth engine? I won't be joining you (on any Sunday), but maybe it will be beneficial?

Much of what humans do (including your explanation here) is repeating common knowledge in response to something someone else says. A LLM can repeat a much larger set of common knowledge than you or I can. In response to, say, an interpreter article, it can debunk much faster than a human can.

In many cases the interpreter articles are so poorly written that they don’t even warrant a human response. Most of what posters here do to refute them is follow the footnotes and see if they match what the Interpreter author claims they say. That’s pretty easily in the wheelhouse of a LLM.

[Failed Prophecy]

You do realize that every organization distorts reality in their own way, including the CoC?



Not really, it's all variations on a theme. There's very little that is new in Mormonism that requires fresh research to debunk.



You seem to have a very misguided view of what A.I. can do. A.I. literally has no concept or truth nor of facts nor of reality. It only recognizes patterns and can give you statistical probabilities of patterns repeating or being similar to other patterns. Even LLMs, the current fad, just make answers based on what is statistically likely to follow sentences that have already appeared in its corpus of texts. Your prompt is just the entry point into a chain of statistical reasoning and the "answer" is just the next chain in that set of probabilities.

The way you talk about A.I. seems more akin to the Mormon view of the Holy Ghost. Maybe go to church on Sunday to experience that truth engine? I won't be joining you (on any Sunday), but maybe it will be beneficial?

Much of what humans do (including your explanation here) is repeating common knowledge in response to something someone else says. A LLM can repeat a much larger set of common knowledge than you or I can. In response to, say, an interpreter article, it can debunk much faster than a human can.

In many cases the interpreter articles are so poorly written that they don’t even warrant a human response. Most of what posters here do to refute them is follow the footnotes and see if they match what the Interpreter author claims they say. That’s pretty easily in the wheelhouse of a LLM.

Go ahead and continue to believe in magic I guess. LLMs don't behave that way. There's no way, even in principle, for an LLM to judge if the truth values of an article are supported by references. They can make up BS, they can't detect it.

[Moksha]

There's no way, even in principle, for an apologist to judge if references support the truth values of an article or a parallel. They can make up B.S., but they can't detect it.

Slightly altered.

[Failed Prophecy]

There's no way, even in principle, for an apologist to judge if references support the truth values of an article or a parallel. They can make up B.S., but they can't detect it.

Slightly altered.

Touche!