Compressed Feasibility: A Technical Update to the Ambient Non-Consensual Synthesis Threat Model

Abstract In January 2026, we published a convergence threat analysis demonstrating that ambientnon-consensual intimate image synthesis via AR wearables was architecturally feasible throughcloud-assisted pipelines and would achieve edge-only feasibility within 12 to 24 months [1]. Thistechnical update reports that multiple independent developments in the eight weeks followingpublication have compressed our timeline estimates significantly. Specifically, we identify threecapabilities released between February and March 2026 that collectively reduce barriers across thecritical pipeline stages: person segmentation (MatAnyone2, 140MB model achieving state-of-the-artvideo matting), inference acceleration (Diagonal Distillation, achieving 270x speedup over baselinevideo generation), and mobile 3D rendering (MobileGS, achieving 120+ FPS Gaussian splatting on aSnapdragon 8 Gen 3 phone at 4.8MB model size). We revise our feasibility classes accordingly, notingthat the "edge full video synthesis" class we originally projected at 12 months may now be achievablewithin 6 to 9 months under moderate assumptions. We discuss implications for the policyrecommendations in our original analysis and argue that the structural mismatch between capabilitymaturation and regulatory response has widened.