!full! | Filma25

In parallel, generative AI research in computer vision (Ho et al., 2022; Brooks et al., 2024) has demonstrated text-to-video synthesis with increasing temporal coherence. While early models (Runway Gen-1, Pika Labs) produced short, surreal clips, newer systems (OpenAI’s Sora, Google’s Lumiere) achieve minute-long sequences with causal continuity. For Filma25, these models are not auxiliary tools but central production engines.

Filma25, post-cinema, generative AI, algorithmic authorship, decentralized film production, DAO, dynamic narrative. 1. Introduction Cinema has always been a technology-driven art form. From the Lumières’ cinématographe to digital intermediate workflows, each major shift in production and distribution has redefined what “film” means. However, the mid-2020s present a unique inflection point. Generative artificial intelligence (GenAI) now produces moving images from text prompts; blockchain technologies enable decentralized funding through tokenized collectives; and streaming platforms have habituated audiences to algorithmic personalization. Within this context, the term Filma25 emerges from online creator communities, experimental film forums, and speculative design discourse as a shorthand for a new mode of filmmaking—one that is neither purely human-authored nor industrial, neither fixed-length nor theater-bound. filma25

de Filippi, P., & Loveluck, B. (2020). The invisible politics of blockchain governance. Internet Policy Review , 9(2). In parallel, generative AI research in computer vision