AI News, 15 November 2026

Sunday 15 November

China’s first fully AI‑generated feature film flops at the box office

Source-page capture accompanying China’s first fully AI‑generated feature film flops at the box office

AI-generated summary China’s first fully AI-generated feature film, Sanxingdui: Future Past, earned approximately 1.8 million yuan during its first three weeks in theatres, according to box-office tracker Beacon Pro. Social media critics and the state-run People’s Daily newspaper noted that the 100-minute movie suffered from poor visual quality, stiff facial expressions, and a disjointed plot.

Also covered bySouth China Morning Post

NVIDIA backs four Korean AI startups at Seoul Forum

SI Agent News. SI is the new AI. Two AI podcast hosts on a newspaper front page

AI-generated summary NVIDIA announced investments in four Korean artificial intelligence startups during the MSIT-NVIDIA AI Startup Partnership Forum held in Seoul on November 9, 2026. The chip giant confirmed it had already invested in Point2 Technology, while additional investments were planned for Motif Technologies, RLWRLD, and Upstage.

Earlier:

cinecode Skill: Claude Code Films From Code, 2.5K Stars

cinecode: A Skill That Lets Claude Code Make Whole Films From Code

AI-generated summary The sarpstar/cinecode repository gained about 2,500 GitHub stars in its first ten days according to explainx.ai. The project functions as a skill for coding agents rather than a video model, allowing users to make films with code.

Earlier:

LLM Judge Audits Literature-Mined Chemistry Data

LLM Judge Audits Literature-Mined Chemistry Data

AI-generated summary Researchers at RWTH Aachen University and Forschungszentrum Jülich posted a preprint on arXiv introducing an automated verification architecture that audits literature-mined experimental records without ground-truth labels. The preprint describes a two-phase, citation-grounded pipeline applied to homogeneous catalytic poly(ethylene terephthalate) depolymerisation to establish an auditable provenance framework for chemical research data curation.

AWS adds Nemotron 3.5 Super VL to SageMaker JumpStart

Conceptual illustration of a document, image and video frames flowing through a compute module to a cloud endpoint.

AI-generated summary AWS has added the NVIDIA Nemotron 3.5 Super VL multimodal model to Amazon SageMaker JumpStart for managed deployment in SageMaker Studio or with the SageMaker Python SDK. The 120-billion-parameter hybrid mixture-of-experts model processes text, images, and video, while AWS testing reported specific throughput and latency metrics for different GPU configurations.

Earlier:

DeepSeek V3 architecture implemented in 190 lines of JAX

AI-generated summary Vaibhaav Tiwari has released an open-source implementation of the DeepSeek V3 architecture using 190 lines of JAX code. The project was trained on a MacBook Air using public-domain detective fiction books and runs without heavy frameworks.

Also covered byTheorem WeeklyEl Solitario
Earlier: