AutoClipJensen × Dwarkesh: podcast clips using creator captions
View original source · The source belongs to its original creator. This page demonstrates an existing public output and does not grant redistribution rights.
This roughly 77-second automatic output keeps a continuous explanation of AI's five-layer stack. It comes from a 1-hour-43-minute interview and includes a platform cover and post copy.
Source and featured output
The source is Dwarkesh Patel's Jensen Huang – Will Nvidia’s moat persist?, roughly 1 hour 43 minutes, in English, with creator-provided captions. This output uses the Xiaohongshu destination and interview layout, starts around 01:07:04 in the source and lasts about 77 seconds. It features energy as the foundation of the AI stack.
Recorded run and conditions
The October 1, 2026 run used an Apple Silicon Mac, one low-priority queue and qwen-plus for analysis. It recorded 21 candidates, 10 rendered clips and about 7.5 minutes. About ¥0.09 is an estimate from recorded text tokens and the unit prices used then. It excludes cloud transcription, image generation, publishing services, hardware and time; it is not a provider bill.
What comes with the video
The output also includes a cover and post copy. The Chinese cover and post title describe the five-layer AI stack and energy as its base. Review the English meaning, subtitles, framing and publishing rights before using the kit.
Try the workflow with your footage
Import an interview with creator captions or a matching SRT, configure an analysis model, choose one destination and produce. Review the opening and ending of each clip, then export the video, cover and copy. Counts, runtime and estimated fees depend on the input and model.