Following this puts new work involving it at the top of your briefing, with a note saying why it is there. Links are taken from the source record, never inferred.
A production A/B experiment on a real ride-hailing platform with statistically significant improvements in key metrics, but the work is a preprint, lacks peer review, reports only summary results without detailed effect sizes or confidence intervals, and is described as an engineering system deployment rather than a clinical or definitive health/safety study.
A novel methodological proposal with algorithmic validation on semi-synthetic and industrial observational data, but without peer review, no confirmatory RCT of the full method, and results presented as descriptive frontier consistency rather than independent statistical significance.
This is a computer science methods paper presenting a novel algorithmic approach to improve LLM reasoning in game theory; it demonstrates a proof-of-concept with a specific metric reduction but lacks clinical or real-world application and is not peer reviewed.