Сезон 2, эпизод 6

Can free, locally generated AI video actually look good? I direct, MiniMax H3 shoots, and Astra edits. We're comparing two 95-second cuts of Bridesmaids & Bourbon: footage generated on my PC and a hosted API version with a reported footage cost of $5.75. In Season 2, Episode 6 of The Automated Author, Elizabeth Ann West and Stacey Anderson show what's working with local AI video, ComfyUI, Codex, and DaVinci Resolve—and where human taste still makes the difference. We look at the shots, flashbacks, pacing, and context that change an edit; how reference packets keep production moving; and what 5-, 10-, and 15-second generations mean for render time. We also share our current computer setups and the studio workflow we're building. Here, “free” means no per-generation API charge for the locally rendered footage. Hardware, electricity, AI subscriptions, and separately created audio/music still have costs. The comparisons and render-time estimates are from our experiments, not a universal hardware benchmark. ⏱️ CHAPTERS 00:00 Welcome: the AI video bill 01:20 Why I've changed my mind about local models 03:00 Subscription costs vs. owning the hardware 04:31 The video opportunity for authors 06:30 The Living Writing System intensive 09:31 Simpler systems and room for AI to work 10:50 Local models, experimentation, and learning to direct 12:31 Drama Is Mine and Bridesmaids & Bourbon 13:24 MDMZ's MiniMax H3 and ComfyUI walkthrough 14:30 Letting AI operate ComfyUI and DaVinci Resolve 16:31 Local footage: what to watch for 17:12 Watch the locally generated cut 18:49 480p footage, ElevenLabs audio, and Suno music 20:38 Why editing context matters 21:48 When AI starts suggesting your replies 25:30 The API version and the $5.75 footage cost 26:40 Watch the API-generated cut 28:20 Comparing the edits and directing revisions 29:25 Cinematic read-alongs and flexible captions 31:00 Getting set up—and finding your files 32:30 Reference packets and render times 35:09 My HP Omen local video setup 38:05 The 5-, 10-, and 15-second render-time math 41:11 Overnight rendering and recovering interrupted jobs 42:09 Stacey's setup and GPU upgrade 45:14 Sharing production between two computers 46:00 Hardware budgets and new storytelling formats 47:11 Leasing and cloud computing options 49:13 Experiments, Drama Is Mine, and what's next 🎬 WATCH THE CINEMATIC READ-ALONGS Drama Is Mine:    / @dramaismine   Start Bridesmaids & Bourbon, Chapter 1 (WIDE):    • WIDE Cinematic Read Along Bridesmaids and ...   ▶️ PREVIOUS EPISODE I Turned My Novel Into AI Video Without Touching the Editor:    • I Turned My Novel Into AI Video Without To...   🧪 EXPLORE THE AI VIDEO EXPERIMENTS https://ai.elizabethannwest.com/exper... 🎓 FUTURE FICTION ACADEMY Classes and intensives: https://futurefictionacademy.com Newsletter: https://futurefictionacademy.com/subs... 📚 READ MY BOOKS https://books.elizabethannwest.com 💬 Which matters most to you right now: lower generation costs, faster rendering, or better creative control? Tell us what you're trying in the comments. Subscribe for practical AI workflows, publishing experiments, and the next chapter of storytelling. #MiniMaxH3 #AIVideo #TheAutomatedAuthor