Generative Video is entering mainstream

Google just launched Veo 3, a “cutting-edge generative video model” that supports Text-to-Video, generating whatever you can dream of at HD Video Quality.

It looks quite impressive, but as with most AI media generation today, it stands to see if it will still take 3,000 attempts to get a result that is actually usable. I feel like we are at still a stage where it’s cute but the actual commercial application is still limited.

Another system that has caught my attention is https://flux-ai.io/, which apparently allows you to keep context between images, so that your brand mascot doesn’t turn into the clown from IT after 10 consecutive generations. Question is, when will this come to their video generation…

(This multi-legged “YETI” is a product of the latest flux-ai.io SDK 1.6, back to the drawing board)