China just beat OpenAI to release an insane AI Video model KLING #sora #aivideogenerator

Chinese technology firm Kuaishou has debuted KLING, an innovative video generation model that appears to match OpenAI's recently released Sora system.

According to Kuaishou, KLING can produce videos lasting up to two minutes at 1080p resolution and 30 frames per second. Notably, it can also model intricate motion sequences that adhere to the laws of physics.

One demonstration video depicts a two-minute simulated train ride showcasing various landscapes visible through the window, generated from the prompt "Train ride with different landscapes seen through the window." OpenAI introduced their video model Sora in mid-February, capable of generating relatively consistent one-minute videos.

Another example highlights KLING's longer video capabilities by portraying a boy riding a bicycle through a garden as the seasons transition. The landscapes evolve appropriately with the changing seasons, while the boy on the bike maintains a consistent appearance, potentially indicating a technique for achieving greater video lengths. However, it would be more impressive if the boy continuously circled the same garden.

A separate video showcases KLING's ability to depict a boy eating a cheeseburger at a fast-food restaurant, with the burger visibly shrinking after he takes the first bite, demonstrating attention to detail.

Currently, KLING is available as a public demo in China. Kuaishou, a Beijing-based tech company primarily known for its social media apps, is now entering the race for large-scale generative AI models with the introduction of KLING.

Follow @thealgorithmicmuse for staying updated with AI tools and the latest breakthroughs in the world of AI, robotics & engineering

🎵: Paris | Else
🎥: @bdsqlsz on X
#texttovideo