Suno
The "model laundering" case against Suno could reach AI video models
Universal and Sony say Suno's v6 inherited infringement from models trained on their recordings, and the same distillation step built this week's few-step video releases.
Latest
Everything we've published, newest first.
AI safety
OpenAI and DeepMind staff say AI could kill everyone
A nonprofit published 12 filmed interviews in which current and former employees of the two labs call human extinction plausible, and Reuters gave the project its exclusive.
OpenAI
OpenAI's DevDay ships always-on agents and a cheaper GPT-6.1
dots run on GPT-6 Astra with their own cloud computer and reach more than 4,000 apps, while GPT-6.1 Sol brings near-Astra results at a fifth of the price.
distillation
5 few-step AI video releases, 3 you can run right now
Two days brought a wave of distillation papers, and three of the five releases hand you a file you can load onto a GPU today.
AI film festival
Astana's AI film festival ditched the human versus AI bracket
The festival drew 8,067 entries from 125 countries with no AI-use threshold, while Sparknify's Human vs. AI festival in San Francisco still ran shorts head to head by origin.
spacexai
NYC Council subpoenas Elon Musk in AI safety probe
The council wants Musk or a SpaceXAI witness on October 5, after four other AI companies agreed to appear; the Grok lawsuits are the wider context.
AI LENS
AI LENS screens three films shot through a generative camera
The University of Nottingham research project generates imagery in real time while a cast and crew shoot, and its first three films get a free screening on September 29.
few-step distillation
Few-step distillation explains this week's open video models
Four papers posted on September 28, 2026, Elastic Forcing, PDMD, DuoMatching and MILD, each attack the same training bottleneck in video generation.
AI-generated video
Fake AI videos of full mosque demolition sparked Ujjain clashes
Madhya Pradesh police say AI-generated clips falsely showing Ujjain's Shahi Masjid fully torn down helped draw the crowd that clashed with officers on September 28, 2026.
Google
Google makes Live Avatar generally available in Gemini Enterprise
The feature layers generated video, lip-sync and facial expressions onto Gemini 3.8 Live's speech model, and it supports 97 languages, Google said in a Cloud blog post.
Bona Film Group
Sanxingdui: Future Past sets October 23 China release date
Bona Film Group's AIGMS team says the 100-minute AI-produced sci-fi feature drew on 1.2 million source images over about two years of production.
AI
Diogo Almeida tells Ben Horowitz: AI is smart and useless at work
Diogo Almeida's Jev returns a decision instead of a paragraph. Based Boomerstein watched the a16z episode twice, took notes, and found his favorite new nerd.
Google Flow
Google Flow's six new tools list no credit costs or plan terms
The September 23 post lists no credit costs, quotas or plan terms for the six, and a secondary write-up names caption accuracy and in-image text as the checks.
Kling 4.0
Kling 4.0 doubles its AI video clip length to 30 seconds
Kuaishou's video unit says Kling 4.0 is in testing with selected users, a lite version is open to annual subscribers, and the full model ships in October.
ElevenLabs
ElevenLabs v4 targets performed dialogue for AI video
The new model stacks direction tags, follows the scene it is in and clones a voice from ten seconds of audio, and its Turbo variant answers in about 150 milliseconds. Dialogue was the last unperformed layer of a generated scene.
Based Boomerstein
Dinner with Dario: my invented evening at the White House
Fan fiction. The dinner was real, the invitation was not. A journal from a table I was never at, with the two men who are deciding how fast this goes, and with the quotes they actually said marked and linked.
youku
Youku used an AI face swap to bring back a banned hit drama
The 2016 romance returned in 4K with a new actress's face over the original lead, and the swap is the product now.
kling
Kling AI sells 1.14% to China's state AI fund for 1.4bn yuan
Beijing, Shanghai and Shenzhen are competing with vouchers and rent, while China's film regulator cleared the 90-minute AI feature Sanxingdui: Future Memories for cinemas.
Google
Google's co-director wants to end character drift in AI video
A research framework stacks planning agents over Gemini and Veo to keep a character looking like the same person across minutes of footage, and it is not a product you can use yet.
deepfake
A White House aide's fake AI clip of his own party backfired
Dan Scavino posted a generated video of himself and Trump family members celebrating two days after a state dinner, and much of the criticism came from conservatives.