Ever feel like your AI videos are stuck in a rut? You either get a cool scene with no talking, or a talking head that can’t do much else. It’s the classic ‘show’ vs. ‘tell’ problem.
Well, I just stumbled across an incredible post from Tianyu Xu, and he’s figured out how to do BOTH with Veo 3. I think this is a total game-changer for anyone in education or storytelling.
He’s been testing Veo 3 and found it’s the ideal model for creating clips that can ‘show and tell’ at the same time. To get a high success rate, he follows 5 golden rules.
Here are Tianyu Xu’s rules for mastering Veo 3 effects:
- 📌 One Job at a Time: Don’t overload a single prompt by asking the AI to multitask. Keep it focused.
- ✍️ Use Plain English: No need for complicated code or jargon. He says to describe the actions and effects simply and clearly.
- 🖼️ Focus on the Image: This was a huge insight for me! Tianyu found that putting extra effort into describing the still image part of the prompt leads to a better video result.
- ✂️ Embrace Imperfection: Don’t chase a flawless final clip. He recommends using the partial results you get, as they’re often more than good enough.
- ⚙️ Use the Right Tool: He points out that while Veo 3 is a ‘show and tell’ champion, it has its limits. For overly wild or crazy effects, other models might be a better fit.
According to his tests, these rules get the job done most of the time. It’s an amazing leap forward for AI video! 🤯
This is just my breakdown of his awesome work. To get the full scoop and see more tips, go check out the original post from Tianyu Xu!