Our mark is 銀幕, ginmaku — the silver screen, and the same two characters in Japanese, Chinese and Korean.

I came to Tokyo at twenty-three for a summer job at a Shibuya mini-theatre and the job never quite ended. Fifteen years of festivals later I still think the only review that counts is the one the room gives you: where the laugh lands, the moment three hundred people stop moving at once, whether anybody stays for the credits.

Cinepuuulse follows Asian cinema — Japan, Korea, China, Taiwan, Hong Kong, Vietnam, Thailand — across about a hundred and twenty sources. A good number of them are Spanish, French and Portuguese, because those countries programme this work seriously and write about it better than most of the English-language trade. Genre and art-house get the same seat here. I am not ranking them, and nothing is scored out of ten.

Each story is a headline, a short excerpt, one still and a link to the outlet that did the reporting. We never republish the piece, and the still belongs to its studio or its photographer. A Korean trade paper appears in Korean, a French review in French. We point at the work; we do not put words in anyone's mouth.

— Daren Wise, Tokyo

← Cinepuuulse Archive

VP Land

Utopai Launches PAI Production Platform and Utopai X Video Model, Linking Script Breakdowns to Generated Shots

10.26

Utopai Studios has launched PAI, a production intelligence platform for film and television that breaks a screenplay into scenes and shots, alongside Utopai X, a new video generation model that…

AI in film Generative AI

Read at VP Land ↗

One letter a week, starting soon.

Latest

Utopai Launches PAI Production Platform and Utopai X Video Model, Linking Script Breakdowns to Generated Shots

Related

'Cocaine Cowboys' 20th Anniversary Re-Release Adds AI-Generated Reenactments and a 4,000-Shot Visual Overhaul VP Land

AI Feature 'Sanxingdui: Future Past' Debuts to Weak Ticket Sales in China as Low-Budget 'Niu Lai' Nears $10M VP Land

fal Launches FLUX 3 Max, a Post-Trained Video Model Benchmarked at About 12 Seconds per 10-Second Clip VP Land

Hy-Nomas Is Tencent's New Open-Source Video World Model You Steer With Text and Camera Paths VP Land

BytePlus Introduces Seedance 2.5 4K With Lower-Cost 480p Draft Previews VP Land

OmniHuman 2.0 Is BytePlus' New Avatar Model for Five-Minute, Audio-Driven Performances VP Land