The preprint by Fu et al. (arXiv:2605.26209, May 2026) employs a data-driven semi-empirical model that ingests observed UV luminosity functions at z~5-12 and derives star formation efficiencies via abundance matching to dark matter halo accretion rates, bypassing explicit hydrodynamics. This yields predicted stellar mass functions and clustering that align with data, but reveals star formation efficiencies spiking to 0.8-0.9 at z>9 before declining, consistent with bursty histories inferred from SED fitting. Yet the approach exposes a core tension: efficiencies near or above unity become unphysical once dust attenuation is included to match massive dusty systems at z<8, forcing a factor of 2-3 reduction only achievable via top-heavy initial mass functions. This directly challenges Lambda-CDM structure formation models that underpredict the abundance of UV-bright galaxies at z>8, as reported in multiple JWST surveys. Mainstream coverage often overlooks how IMF variations alter not just stellar masses but also ionizing photon budgets critical for reionization timelines. Related analyses, such as those in Finkelstein et al. (ApJ 2024) on evolving UV LFs and Steinhardt et al. (2023) highlighting IMF-driven mass overestimates, reinforce that fixed Salpeter IMFs produce inconsistent growth tracks when dust is self-consistently modeled. The preprint's methodology, while robust in reproducing the star-forming main sequence without tunable feedback, lacks direct constraints on stochasticity or metal yields, limiting predictions for ALMA dust continuum detections. Ultimately, these results imply that early galaxy assembly requires redshift-dependent IMF shifts, a factor frequently sidelined in favor of exotic feedback tweaks.