The article also covers various AI industry updates, including YouTube's new content usage options for creators, Meta's upgraded smart glasses, and DeepMind's video-generating AI, Veo 2. Additionally, it mentions the death of an OpenAI whistleblower, Grammarly's acquisition of Coda, and Cohere's partnership with Palantir. The piece concludes with insights into Anthropic's Clio analytics system and Pika's video generation model, alongside the Future of Life Institute's AI Safety Index, which evaluates AI companies' safety practices, with Meta receiving the lowest grade.
Key takeaways:
```html
- OpenAI's Advanced Voice Mode with Vision aims to enhance ChatGPT's contextual awareness but struggles with reliability, leading to noticeable errors and hallucinations.
- YouTube is allowing creators to control how their content is used for AI model training, giving them the option to permit or deny specific companies.
- DeepMind has launched Veo 2, a new video-generating AI capable of creating high-resolution clips, as part of its effort to compete with OpenAI in video generation.
- The Future of Life Institute released an "AI Safety Index" to assess AI companies' safety practices, with Meta receiving the lowest grade and Anthropic scoring the highest, though still only a C.