You’d have to explain why that video is relevant and what it says that is worth responding to.

The argument would need to be better formulated to assess.

What exactly counts as artificial general intelligence? (I assume that’s what you/they mean) And when would we be able to say “it’s impossible” (since we can’t try every possible thing anytime soon)?

For example, does synthetic biology count as AGI? For example, in Alien, robots are organic constructs, not electromechanical but bioelectrochemical, yet entirely manufactured (in canon, they had real human prototypes on which they were modeled, but they make clear their minds are modified, hence Bishop is moral, via the installation of Azimov’s Laws, while the actual scientist who was his template was a total sociopath).

For instance, suppose we used a flesh printer (which will exist in our lifetime) and printed a brain and transplanted it into a recently deceased but life-supported corpse, and then a new conscious person awoke in that body. Does that count as AGI? And semantics aside, what effect would that have on the argument?

It is possible (and has even been proposed, though I don’t buy it) that AGI requires a specific kind of wetware to work (so it has to be neurons based on carbon molecules; it can’t be digital silicon circuits). Does that count as “AGI being impossible”? Or does it simply indicate the physics of consciousness is peculiar? On naturalism, the physics is the physics. So its merely being peculiar would not indicate God arranged it that way. If a rambunctious cat stumbles into its owner’s time machine a billion years from now and bumps into something that accidentally causes the world’s physics to work that way from the start, it’s still the way the physics works. There is no need of intelligent design or involvement. To argue that would require more evidence than just “the physics is peculiar.”

The only way the argument works, IMO, is in the very specific circumstance whereby printing a brain never works (something we are nowhere near to yet), and even though we fully understand the circuitry of a brain (something we are nowhere near to yet) we can’t get any program to emulate it (with any software or hardware) such that we literally run out of even ideas of how to try it (something we are nowhere near to yet). Then, perhaps, one can say that consciousness requires some intelligent being to hit the “on” switch and a good candidate for that being is a god of some kind.

But that really only gets you to supernaturalism. Because do we know the thing hitting the on switch is sentient? Maybe it’s not, yet still supernatural. Like the Tao or the Force or spontaneous psychic powers we never knew always resided in babies’ mothers, or something even weirder. In The End of Mr. Y supernaturalism is entirely generated by consciousness itself in an endless time traveling loop, such that there never is a God behind it all, but always just humans creating their own reality, complete with all the gods they imagine exist, in which case gods literally exist but are created by humans, indeed accidentally, and not the creators of humans, intentional or accidental. That sounds weird, but it is logically possible, and honestly, no weirder than modern theisms.

So the argument would need some serious rigorizing to get around all these analytical and empirical problems.

What’s that even trying to say? If we can’t create intelligence or moral reasoning neither can evolution, therefore there must be a non-physical source, therefore God?

If so, I think that might be covered here with regard to Swinburne: https://www.richardcarrier.info/archives/12814 (search for “moral awareness”)

What do you think of an argument of this type?

Honestly, if AGI proves to be impossible to create, that’s a pretty good argument for the existence of God on both moral and cosmological grounds.

It was made by a professor of pentecostal studies at Emory University. It looks intuitively very weak to me, but I can’t outline my critique clearly.

I did address that in this article, explaining why we now trust reason, and certain improved models of it (like formal logics and mathematics, and the scientiufic method).

It’s not circular because it operates on an external check. Some methods are confirmed to work because they get material results; others are shown not to work by not as reliably doing that.

So, for example, the scientific method is known to be successful by contact with reality; if it weren’r working, we couldn’t have sent men to the moon or bred genetically modified grain that grows in a drought. We couldn’t even successfully navigate our way to the bathroom if even our natural faculties weren’t roughly reliably contacting a reality apart from our minds.

You cannot explain this in any other way than appealing to a Cartesian Demon, which cannot survive any defensible review (see my article on our not being in a simulation for why).

As for whether we think in words or propositions, it’s both. Words are a computational process we build on top of the inherent one, which is propositional, when we define propositions (as distinct from sentences encoding them in words) as models of reality. The brain operates as a model builder. The models it builds are what we mean by propositions. And we innately think by navigating those models, and testing whether the model then plays out in reality or not, through contact with reality, as our external check.

A thought occurred to me when I was reading the part about propositions:

Do we think in words/images or do we think in propositions?

Apart from some visualizations, I usually form sentences in my head when thinking and not always in the same language, which is kind of funny.

I suppose sentences like that are useful as inputs into the next stage of thinking.

Now I wonder, are propositions the basis of thinking that get externalized as words/visuals?

Anyway, the main thing I wanted to post here is a part of a discussion that goes on when debating about AfR which is omitted in this article.

The omitted part is about justification of reasoning process in the first place.

You have in many ways, with great detail, explained how truth-detection capabilities vastly improve chances of survival and give evolutionary edge to organisms.

However, when explaining this, you are using the very reasoning faculty that is being put into question.

In other words, you are using reason to justify reason, which is circular.

Theists may say that given naturalism, you are on very shaky grounds.

For that, reliability of cognitive faculties is stronger given supernaturalism, since by definition God is perfectly rational. That’s the argument.

However, I see the same problem not being solved. They also need to presuppose reason to get to that conclusion.

As far as I see, there is no way out of this. Reasoning has to be taken as a fact, and it cannot be justified because there are simply no tools to do this. As believers would like to say, you have to have faith in reason.
Otherwise, we are just paralyzed from the start, and any discussion is impossible.

I’m really interested; what are your thoughts on this issue?

I like this sneak peak into your process. 🙂

The syllogism you just made here is false and obviously false.

“All good meals are the product of causal-deterministic systems.
Bad meals are also the product of causal-deterministic systems.
Therefore naturalism cannot tell the difference between good and bad meals”.

Is this sensible in any way? No. The fact that X can produce Y or Z obviously does not mean that you can’t tell when X will produce Y instead of Z, or why. That’s sort of the entire point of Richard’s article: For example, when he says, “As long as atoms are arranged so as to compute information (as computers now indisputably do, and human brains have been well enough demonstrated to do), and as long as that system has a way to check its outputs against reality (like, say, the basic human senses; as well as success or failure at avoiding danger and eating, and other like necessities), then suitably arranged systems of atoms can indeed tell the difference between true and false propositions.”

Thus not only can naturalism explain how people can differentiate between true and false beliefs, but it itself is capable of doing so.

You say that Richard’s article is a smokescreen, but it facially addressed this point.

Worse, how the heck does theism help? Theism very demonstrably does a very bad job of differentiating between true and false beliefs. Theists can look at the exact same text in the same language and derive vastly different information.

The problem with the theistic explanation for reason is that, as with all theistic arguments, it makes predictions that are falsified, and that theists virtually never defend because their goal is to defend theism not explain phenomena. If our reason is the product of a god, it should be consistent and accurate. Our intuitions should be perfect, we should have no psychological biases, and whatever built-in information we have about the world should be perfectly accurate to all contexts.

Figuring out why we are capable of achieving the reason we can is a complex process. Theism is a non-explanation.

I’m not sure what you are missing here. The syllogism you present is a non sequitur (“Therefore naturalism cannot explain” does not follow from the given premises preceding). That is my point.

Perhaps you mean to clarify that their position is inductive (probabilistic), not deductive, i.e. the falsely affirmed consequent is not that all causally formed beliefs are false, but that a causally formed belief is unlikely to be true.

In other words, in a deductive fallacy, it would be “if there is a false belief, then it was caused; therefore, if a belief was caused, it is a false belief,” but an inductive version of that fallacy would be “if there is a belief unlikely to be true, then it was caused; therefore if a belief was caused, it is unlikely to be true.” Which is still illogical.

As to how we tell the difference between a causal process that is reliable and a causal process that is not, this is what the entire article you are commenting on is about: how we tell the difference (and not just we, but blind natural selection even) is by observating the congruence of the output of the process and reality (by observational testing or simple death-or-survival outcomes). The role of causation alone has no bearing on this distinction. The likelihood of a causally produced belief can only be ascertained by what is causing it (such as, a rational or a nonrational process), not merely that it is caused.