Extreme AI zoom, new DeepSeek, full body avatars, Kling 2.1, new text-to-speech, robot kickboxing
Recent AI advancements include Tencent's realistic character animator, a new open-source text-to-speech generator, humanoid robot tournaments, and detailed 3D model generators, showcasing significant progress in AI capabilities.
MAIN POINTS FROM TRANSCRIPT
- Tencent's Hunyan video avatar offers realistic lip-sync animation for characters using a single image and audio.
- New open-source text-to-speech generator claims superiority over 11 Labs.
- First humanoid robot fight tournament successfully concluded in Hongjo, China.
- Deepseek's upgraded model rivals top models like Gemini 2.5 Pro.
TAKEAWAYS
- Tencent's AI can animate multiple characters with realistic emotions and movements.
- The Hunyan video avatar supports various media types, including drawings and 3D characters.
- AI advancements include converting PDFs into scientific posters.
- All models and resources from Tencent's project are available on HuggingFace and GitHub.