I’ve been watching the recent buzz around “testable” multiverse models - specifically the idea that we can prove the existence of an inflationary multiverse by showing our universe is “near-optimal” for life.. Déjà vu.
Back in 2006, I wrote about Lee Smolin’s Cosmological Natural Selection (CNS) (original post here). At the time, I loved CNS because it offered a beautiful, self-organizing mechanism for complexity. I think it gave some indication of how a complex universe could come into being by giving our universe a pathway to evolve its properties.
However, even then, I couldn’t buy the “testability” Smolin claimed. He argued that if our universe wasn’t perfectly optimized for black hole production, his theory was dead. My pushback was simple: Evolution doesn’t strive for perfection. It’s a messy, “good enough” process of random drift and local maxima. Finding a universe that isn’t quite at the peak doesn’t falsify an evolutionary model; it just describes a mutation that hasn’t been weeded out yet.
The recent New Scientist piece on the inflationary multiverse (April 2026) leans into a similar “test.” It suggests that if we model the constants of physics and find ourselves sitting at the peak of a “life-friendliness” curve, we’ve found evidence for a multiverse.
But this “near-optimal” test still feels like a bit of a shell game. Just as with Smolin’s black holes, if we eventually find our universe isn’t optimal, the theory won’t be discarded. We’ll likely just see more explanations about why “near-optimal” is the best we could expect given the chaos of inflation. The real pull of Smolin’s idea was the mechanism - the idea of a universe that evolves. Modern inflation gives us a “Swiss cheese” of infinite random bubbles. Both are sprawling, fascinating explanations for complexity, but trying to “prove” them by claiming we live in the “best of all possible worlds” seems to me like a scientific dead end.
Related Reading:
#cosmology #sci