钛媒体· TechPulse·· 2026-10-20AI score40
HiDream.ai bets on model-workflow synergy as AI video generation competition shifts to delivery capability
AI视频生成交付能力成焦点,智象元一押注模型与工作流协同
AI briefing
HiDream.ai co-founder and COO Wang Ke said during the Yunqi Conference that AI video generation is moving from "usable" to "easy to use," and that over the next 1–2 years the core of competition will shift from model capability contests to workflow closed-loop capability. Its video model HiDream-O1-Video-1.0 adds a front-end intent understanding module and adopts a self-developed native omni-modal UiT architecture; the film and television creation collaboration agent Zhenzan supports script analysis, storyboard generation, and character design, and a 5-person team completed 130 minutes of content for the 24-episode short drama "Yanmo's Beloved Concubine" in 15 days.
Open original tmtpost.com
Video GenerationComfyUI & WorkflowsPerspectives & InterviewsPhenomena and Trends#Commentary#Trends#Video generation#ComfyUI & workflows