Sora Series Core Authors Luhman Brothers Resign: Starting point: November 2020, *Diffusion Models for Handwriting Generation*, co-authored by the two, no institution affiliation, Gmail email → independent researchers. Representative work: January 2021, *Knowledge Distillation in Iterative Generative Models for Improved Sampling Speed* (Eric as first author), multi-step denoising distilled into single-step, predating Salimans & Ho's Progressive Distillation by a year, cited 445 times, referenced by subsequent distillation works like Diffusion Duality. Joined OpenAI in 2021 as Research Engineer; continued co-authoring in 2022–23 on Patching / hierarchical VAE / latent space Deep VAE (ImageNet-256 FID 9.34). Sora Report (Feb 2024): two of 13 authors, report cited 1,100+ times; Sora public version (Dec 2024): contributors; Sora 2 (Sept 30, 2025): ranked 6th and 7th among 26 researchers. Troy’s public profile explicitly states "worked on Sora and Sora 2, including new audio capabilities"—Sora 2 introduced synchronized audio-visual generation for the first time, which is the core argument for the "most important contribution" in the chart. Oct 29, 2025: Will DePue announces return to OpenAI, forming a three-person ASI team with @troyluhman and @eric_luhman1. April 26, 2026: Sora product discontinued (API shutdown on Sept 24). Scholar metrics (Troy): 2,235 citations, h-index 6, 749 new citations in 2026. Eric has no Scholar profile; GitHub ericl122333’s PatchDiffusion has 115★. Data source: DINQ