TechCrunch Summary (2023-06-09 05AM)

Source: Runway’s Gen-2 shows the limitations of today’s text-to-video tech

Summary
Google-backed AI startup Runway has released Gen-2, a commercially available model capable of generating videos from text prompts or existing images. However, one limitation is the low frame rate of the videos the model generates, which makes it unsuitable for editors hoping to avoid post-production work.

Fact

🎥 Runway, a Google-backed AI startup, has released Gen-2, a commercially available model that can generate videos from text prompts.

🤔 One of Gen-2’s limitations is the low frame rate of its generated videos, making it unattractive for editors hoping to avoid post-production work.

😕 Gen-2-produced videos often appear grainy or fuzzy, with artifacts such as pixelation around objects, while people’s arms and legs meld together and objects disappear, with their reflections warped.

❌ As with many generative models, Gen-2 struggles to understand nuances and often ignores particular descriptors in prompts while focusing on others seemingly at random.

Leave a Reply Cancel reply