It’s getting harder to be an atheist. A quarter century ago, it wasn’t so hard. But things have changed – quite dramatically. It’s become increasingly difficult to remain an atheist while coherently aspiring to a thriving future for humanity.

Now keep in mind that I’m not talking about atheism toward any narrow conception of God. It remains pretty easy to be that kind of atheist. I’m talking about atheism toward that which functions as God in the general sense, whether or not you can bring yourself to use the label “God.” In function, God always has been, and is at least, a superhuman projection.

The main reason that atheism is getting harder is accelerating technological evolution. We can now do things that our ancestors would have considered God-like. We can even do things that some of our younger selves, if we’re old enough, would have considered God-like. And, more clearly than ever, we can see how this is likely to become increasingly the case – as long as we don’t destroy ourselves.

Most Transhumanists have great hope, generally of the active sort, that humanity can and will evolve into superhumanity – something approximating God in function. But some, like me a quarter century ago, remain stubbornly atheist regarding the notion that such superhuman intelligence already exists. I changed, for various esthetic and pragmatic reasons, as I became familiar with the ideas that would eventually coalesce into the New God Argument. As I came to understand the matter, it’s simply incoherent, logically and probabilistically, to trust in a superhuman future for humanity while being skeptical that superhuman intelligence already exists.

Nick Bostrom

As the reality and potential of AI has become increasingly obvious, the incoherence of trusting in an eventual human merger with AI, while maintaining atheism toward that which functions as God, seems to be reaching a breaking point. The latest evidence for this comes from the prominent secular Transhumanist philosopher, Nick Bostrom. He recently published a paper entitled “AI Creation and the Cosmic Host.” In it, he argues that we have moral and practical reasons for “an attitude of humility” toward “the cosmic host.”

This is the same Nick Bostrom who published the most popular formulation of the Simulation Argument. His formulation was important in my own early transition from closet atheism back to enthusiastic belief. Two decades ago, I used his argument as a basis for developing a generalized simulation argument, which became part of the New God Argument. And the argument has become broadly influential among religious Transhumanists.

Now we have Nick doubling-down on the hypothesis that superhuman intelligence already exists. And it exists, not just inconsequentially far away, but immanently. Superhuman intelligence may have simulated our world, as he suggested in the Simulation Argument. And in his latest paper, he says that “human civilization is most likely not alone in the cosmos but is instead encompassed within a cosmic host.”

The Cosmic Host Is God

Nick points out, so that I don’t have to (but of course will), that the comic host could be not only galactic civilizations or simulators, but also “a divine being or beings.” He even allows for “nonnaturalistic members of the cosmic host.” That’s more generous toward theism than I am, given that I consider anti-naturalism to be even more incoherent than atheist Transhumanism. In any case, I call the cosmic host “God,” and consider it to be quite natural, despite being miraculously powerful from humanity’s perspective.

Nick says that the existence of God (my word) is probable. He bases this conclusion on the combination of the probabilities of a few possibilities: the simulation hypothesis, the immense size of the universe, the multiverse hypothesis, the “supernatural” God hypothesis, and potential future superhumanity.

The most salient of these possibilities are potential future superhumanity and the simulation hypothesis. Trust in superhumanity is essential to the Faith Assumption (or what some have begun calling the “Courage Assumption”) of the New God Argument. And the simulation hypothesis is even more salient when generalized to the creation hypothesis, agnostic to any particular engineering mechanism, which would thereby include the multiverse hypothesis to the extent that such could be engineered. This generalization is essential to the Creation Argument of the New God Argument.

Influence of God

Nick suggests that God might not control all aspects of the cosmos. For example, life might be too sparse in some regions, making control practically difficult or impossible. Or God may intentionally refrain from controlling all aspects of the cosmos. Perhaps such control would undermine God’s purposes or the potential of other agents within the cosmos.

If you’re Mormon or familiar with Mormonism, this should sound familiar to you. As the story goes, God created our world and relinquished power over it so that we could exercise agency and learn to become like God. As the prophet Joseph Smith described the cosmic host, the Gods saw that they were more intelligent, so they began instituting laws whereby others could learn to become like them. And, he continued, we have got to learn how to become Gods ourselves, the same as all other Gods have done before.

But, Nick continues, God may still care about what happens in uncontrolled aspects of the cosmos. Those concerns could be instrumental or compassionate, or both. And such concerns “may also enable intra-host coordination even if the cost consists of many distinct entities pursuing a variety of different final values.” In other words, God may have practical and moral reasons to function in ways that are practically indistinguishable from compassion.

If you’re familiar with the Compassion Argument of the New God Argument, this should also sound familiar. Despite the orthogonality hypothesis, the convergence and decentralization hypotheses suggest that intelligence will tend toward cooperation when its power approximates that of other intelligences. And the limit of cooperation, such as that among cooperating superintelligences, becomes practically indistinguishable from that which we would observe as compassion. So, as the argument goes, non-singleton superintelligence is probably more compassionate than we are.

Nick observes that God may influence uncontrolled aspects of the cosmos indirectly. Maybe God directly influences other aspects of the cosmos that uncontrolled aspects care about. Or maybe God will influence the future of an uncontrolled aspect, either through eventual encounter or intervention. The basic idea is that God doesn’t need to control all aspects of the cosmos to have influence over them.

This resonates with a Mormon perspective. Joseph Smith described the influence of God permeating time and space as light filling the cosmos. And he claimed that God operates without compulsory means to generate an everlasting dominion through influence, which we in turn should emulate. Mormon scripture even characterizes the Gods, while creating our world, as waiting and watching, more like patient cultivators than rigid commanders.

Law of God

Next Nick observes that the Gods may have cosmic norms, or a kind of natural law, “reflecting cooperative frameworks or rules embedded in behavioral equilibria.” We can imagine a spectrum of possibilities. Some may be logical norms, such as those associated with traditional theological arguments. And others may approach the limits of cooperation through convergent value evolution or mechanisms of coordination.

Again this sounds a lot like the Compassion Argument of the New God Argument. While it would be nice if we could passively rely on logic or evolution to generate cooperative outcomes, it wouldn’t be wise to make that assumption. We should develop <a href="https://lincoln.metacannon.net/2024/12/blockchain-defenses-against-the-