dessa
Block's Dessa voice model is trained only on speech
The speech-to-speech model listens and talks at once and, Block says, runs on a MacBook Pro with an M4 Max, but early access goes only to safety researchers.
SLOP TV News Desk
filmflow
FilmFlow turns a text prompt into a Blender camera path
The Shanghai Jiao Tong University-led system ships with CinemaBench, 150 Blender scenes and 750 camera instructions, but no code is released yet.
collage
Ex-UTA agent Dante DeSario launches Collage as AI studio broker
The Beverly Hills firm closed an undisclosed seed round, and Deadline lists eight AI-native studios, from Asteria to Secret Level, as launch clients.
AI likenesses
AI likeness deals: who signs them and what the rules allow
Sunny Hostin and the cast of Higgsfield's Cully Hill Boys licensed their likenesses; Catherine Zeta-Jones prefers real performers and Nicolas Cage refused an AI waiver.
Hy-Nomas
Tencent's Hy-Nomas world model ships with no published scores
The Apache-2.0 weights, code and a 280-case benchmark went up on October 23 with no paper and no results, and one configuration targets a single RTX 4090.
nvidia
Nvidia's $4,999 DGX Spark runs some AI video models, not all
NVIDIA's six OEM partners start selling the 64GB box on October 23, and two units linked by a QSFP cable pool to 128GB for models up to 200 billion parameters.
comfyui
ComfyUI adds VEED nodes to lip-sync a still photo
The three cloud nodes also re-sync a speaker's mouth in footage you already shot and cut a subject out with an alpha matte.
verus-insight
Kling and Peking University publish a detector that shows its work
A Peking University and Kling team's 8B model lifts detection accuracy, but its own benchmark shows it still misses most marked artifacts.
seedance
ByteDance's Seedance 2.5 goes 4K for film and television
Volcano Engine says extra training improves facial expressions and small text, and a 480p Draft mode lets teams storyboard before the final 4K render.
bona-film-group
China's first AI feature film opens to weak ticket sales
Bona Film Group's fully AI-generated feature sold 80,000 yuan in advance tickets against a forecast of under 150,000 yuan and a production cost over 10 million yuan.
google
AI video's preview trap keeps retiring models mid-project
The deprecation table names gemini-omni-1.1-flash as the replacement, and the same date retires Gemini Omni Flash's own preview build.
OpenArt
OpenArt says annual revenue nears $200M with 8 million users
The AI creative platform hired Salesforce's former global brand chief, Scott Larson, and says headcount grew to 80 from 20 in a year.
ai& video
Japan's ai& launches ai& video with 12-second generation
ai& video turns Japanese text, photos and existing clips into new footage, and enterprises can train a private model on assets they own.
Kling 4.0
Kling 4.0's API opens to developers in early access
Kling's own documentation calls the release early access while its pricing table and developer page still say the 4.0 API is coming soon.
Higgsfield
Higgsfield adds Nvidia and TCV to its Series B extension
Nvidia and TCV's money builds on an August round that valued Higgsfield at $5.4 billion, and the size of the new investment was not disclosed.
SlopTV Games
SlopTV Games review: Cold-Gas Teslas and a 1984 NFL rebuild
A browser arcade with 35 titles, no install and no sign-up, and four in-house games worth the trip: an eight-stage Cold-Gas Tesla shooter, a top-down cocaine co-op thriller, a six-fighter political parody, and a 4.2 GB rebuild of ESPN NFL 2K5 set in the 1984 season.
ai-political-ads
Midterm AI ad spending jumps to $137 million from $80 million
The Wesleyan Media Project counted 220 AI-made ads through Oct. 11, but only 27 percent carried a disclaimer and only 30 states require one.
matrix-game
Matrix-Game 4.0 renders game worlds at 60 FPS on one RTX 5090
Skywork AI's open-weights release splits an executable game engine from a causal video model and ships three playable scenes under Apache-2.0.
Pictory
Pictory 3 adds an AI video agent that drafts the first cut
You brief the agent, answer five setup questions and review a script and scenes before anything renders, Pictory's own pages say.
Kling 4.0
Omni reference: what AI video models now take as input
Kling 4.0 takes up to 15 reference assets in one generation, while MiniMax H3 and the open FastH3 Omni cap a request at 12 files.