What if the cosmological constant isn't fine-tuned, but a self-correcting feedback loop?
TL;DR: The cosmological constant (Λ) is off by a factor of 10¹²⁰ from quantum field theory predictions, the worst fine-tuning problem in physics. I show mathematically that if structure formation sources Λ, but a larger Λ suppresses structure formation (creating a negative feedback loop), there's a unique stable fixed point near the observed value. This result is mechanism-independent and holds regardless of what drives the feedback. I then test one specific mechanism against DESI DR2 data. One reading is ruled out at 4.5σ; the other survives and makes predictions testable with next-generation surveys. I'm reporting both results and looking for serious critique or interest in these.
The problem
The 10¹²⁰ discrepancy between the QFT vacuum prediction and the observed Λ is usually treated as either an unexplained fine-tuning or an anthropic selection effect. This three-paper series (see links below) takes a different approach.
Paper I proves that any cosmology where structure formation sources Λ, but a larger Λ suppresses structure, has a unique stable attractor. The result is mechanism-independent and robust: a 100× error in the coupling shifts Λ by only 1.3–35×, versus the 10¹²⁰ range of the original problem. Five coupling mechanisms and four published mass functions all converge to the same fixed point. All inputs from Planck 2018 and Sorini-Peacock-Lombriser (2024, MNRAS).
Paper II tests one specific mechanism against DESI DR2 BAO data. It makes a zero-free-parameter prediction: w = −1.09 ± 0.04. DESI rules this out at 4.5σ. I'm reporting that honestly. The falsification kills one reading of the mechanism but the attractor survives, and the remaining reading predicts a σ₈–Λ joint constraint testable with upcoming surveys.
Paper III connects the attractor to quantum gravity. The dimensionless Λ/k² is O(1) at both the Planck scale (Reuter fixed point, ≈ 0.19) and the Hubble scale (Friedmann, ≈ 2.07). The 10¹²⁰ is just the squared ratio of the two scales. Tested across 11 truncations from 6 independent research groups. This result is conditional, it requires a rank condition ("Computation B") that hasn't been performed yet, and I say so explicitly.
Full Python code for all three papers included under CC BY 4.0. Every parameter comes from published sources, nothing fitted or calibrated.
Why I'm posting: this hasn't been peer-reviewed, and was borne out of a long period of daydreaming, my newborn baby crying and interest in this field, which is exactly why I'm here.
I'm a dude with an engineering background with no current university affiliation, looking for serious critique, pointers to related work I've missed, and if anyone with arxiv endorsement finds it credible, a conversation about getting it onto arXiv.
Paper I - The Cosmological Constant as a Feedback Attractor https://doi.org/10.5281/zenodo.20156389
Paper II - Testing a Connected-Singularity Mechanism for Gravitational Feedback Cosmology
https://doi.org/10.5281/zenodo.20222173
Paper III - Two-Boundary Determination of the Cosmological Constant from Asymptotic Safety and Gravitational Feedback https://doi.org/10.5281/zenodo.20222351