Google has introduced Veo 3, a neural network designed to generate videos with audio, including accented speech and background noise.

Key highlights:
– Videos are marked with an invisible watermark using SynthID, as a safeguard against deepfakes;
– Veo 3 is currently available only in the U.S. as part of the Flow package for filmmakers;
– The package also includes Imagen 4 (for image generation) and the Gemini language model;
– Price: $125 for 3 months of access.

What are your thoughts on this? Would be happy to read them in the comments below!