
Making babies with AI might be the strangest thing I've ever done in my 40 year…
Gemini Veo 3's talking baby trend isn't as easy as it looks
A recent video uploaded by tech enthusiast LanceUlanoff highlights the ongoing challenges in AI video generation, specifically with the Veo 3 model. The video showcases a series of AI-generated clips featuring a baby astronaut in various space settings, including a selfie on Mars and an encounter with an alien. According to Ulanoff, while the initial clip of the baby astronaut perfectly matched his creative prompt, subsequent generations demonstrated significant inconsistencies. "The first clip followed my prompt perfect. The rest...?" read the on-screen text in the video, underscoring the unpredictable nature of current AI tools beyond initial successful outputs. This demonstration provides a practical look into the complexities faced by creators utilizing advanced AI models for visual content. It emphasizes that despite rapid advancements, achieving consistent and precise results remains a hurdle for AI video generation platforms like Veo 3, prompting discussions on the need for further development in AI's ability to maintain narrative and visual coherence across multiple generations.
Gemini Veo 3's talking baby trend isn't as easy as it looks