While we still wait for OpenAI’s Sora, others are stepping into the video generation ring such as KLING and now Luma Labs. Their Dream Machine launched this week and offers, free, realistic videos, from text and image prompts. They state on their website: “It is a highly scalable and efficient transformer model trained directly on videos making it capable of generating physically accurate, consistent and eventful shots. Dream Machine is our first step towards building a universal imagination engine and it is available to everyone now!”
Luma Dream Machine generates 5 second cinematic videos with lifelike motion and maintains a degree of visual and physical coherence even in complex scenes but does struggle with animals, humans, character movement and text. The model’s simple interface makes its accessible to a wide range of creators, although more control will be needed to shape the sometimes-unpredictable results. Social media platforms have been flooded with videos showcasing Dream Machine including many classic memes converted into videos.
You can see our newsletter image here, animated to show the viewer flying through the futuristic scene. Using your preferred image generation tool to prompt the video creation seems to get the best results.
Takeaways: Luma Labs AI has made the model available for free, offering users up to 30 video generations per month.