Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.
Alibaba released Wan 2.7 on April 22 with first-and-last-frame control, 9-grid multi-angle reference input, and natural-language instruction editing — open weights are expected this week or next. First/last-frame alone rewrites iterative shot design: define your bookend frames, model fills the motion. Full 9:16 vertical support (Wan 2.2's biggest gap), 1080P up to 15 seconds, native audio included.
Lightricks' LTX 2.3 (22B parameter DiT) handles text-to-video, image-to-video, and native audio sync in one pipeline — no separate lip-sync node. Output: 4K at 50 FPS. Official workflow is live in the ComfyUI template browser now; requires v0.16+ and 24GB VRAM minimum. Step-by-step tutorials are up at apatero.com and docs.comfy.org.
Sora web/app access ended April 26; API goes dark September 24. Lifetime revenue: $2.1M against $15M/day compute — the math never worked. Kling 3.0 now tops the arena leaderboard at 2127, ahead of Seedance 2.0 Fast (1993) and Happy Horse (1962). Kling weekly actives jumped to 2.6M in the week following Sora's announcement. Any remaining Sora dependencies need migration before September.
Seedance 2.0 caps at 2K vs. Kling's 4K, but ByteDance engineered it for throughput: director-style controls, multimodal reference inputs, and joint audio-video generation that syncs dialogue, ambient sound, and music in a single pass with no post-processing. For operators running 20–30 clips per week for social or ad creative, Seedance is the current fastest path from brief to deliverable.
TikTok's 2026 algorithm now requires 70% completion rate (was 50% in 2024) and actively detects and deprioritizes fully AI-generated video in favor of human-voice content. Saves and shares now outweigh likes significantly. The surviving format: 60–180 second videos with human voice over AI-assisted scripting and editing. Creator Search Insights is live, and search-intent-matched content is outperforming pure FYP optimization in multiple niches.
YouTube Shorts RPM by niche: finance $0.15–$0.45/1K views, AI tools/tutorials at $15–22 CPM with 18x YoY growth. Long-form finance runs $10–35 RPM. Shorts now drive 18% of total YouTube creator earnings (up from 4% in 2024) — the operator model is using viral Shorts to funnel long-form subscriptions. B2B SaaS automation is the sleeper: high CPM, lower competition than direct finance.
A May 2026 Medium case study documents $34K earned from a single faceless AI YouTube channel at under $3 production cost per video. Clixie's scaling study corroborates the math: per-minute production cost drops from ~$3,000 traditional to ~$30 with AI — 99% reduction. Established faceless operators report $5K–$30K+/month at scale, with 6–12 months typical runway to monetization threshold.
After mass monetization suspensions hit faceless AI channels in early 2026, YouTube clarified: faceless channels are not banned. The enforcement target is low-effort, mass-produced content with no human editorial layer. The formula for staying safe: human oversight + AI disclosure label + production quality above the slop floor. With Shorts at 18% of total creator earnings, YouTube has a clear financial incentive to protect legitimate operators in the ecosystem.
