<OpenAI>

A personal post written after an attack at Altman's home, using the incident to reflect on family safety, public rhetoric, and the heightened anxiety around AI. It also lays out his current beliefs about broad access to AI, democratic control, safety, adaptability, and the need to argue about the future without intimidation or violence.

The post discusses upcoming changes to Sora based on user feedback, focusing on enhancing rightsholder control over character generation and exploring revenue-sharing models for video content. The aim is to foster interactive engagement while ensuring creators can manage their intellectual property effectively. The author emphasizes a commitment to rapid iteration and responsiveness to user needs.

The post introduces Sora, a new app designed to enhance creativity through video creation and sharing. It emphasizes the potential for a significant boost in artistic expression while acknowledging concerns about social media's negative impacts. The team outlines principles aimed at ensuring user satisfaction, encouraging content control, prioritizing creation, and supporting users in achieving their long-term goals.

The post highlights the collaborative dynamic between Jakub, the chief scientist, and Szymon, emphasizing their exceptional problem-solving abilities and complementary skills. It reflects on the unique partnership they share, likening it to successful research collaborations of the past, and celebrates their ongoing contributions to OpenAI.

The post reflects on the journey of OpenAI over the past nine years, particularly focusing on the transformative impact of ChatGPT's launch. It discusses the challenges and growth experienced within the company, the importance of effective governance, and the gratitude felt towards colleagues and supporters. The author emphasizes the potential of AI to benefit society and the commitment to advancing safety and alignment in future developments.

The post highlights two key aspects of OpenAI's recent announcement: the commitment to providing advanced AI tools for free or at low cost, and the introduction of a new voice and video interface that enhances user interaction with AI. The author expresses excitement about the potential for these innovations to transform how people engage with technology.

The post discusses the launch of DALL•E 2, an AI tool that generates and edits images from natural language prompts. It highlights the tool's potential to enhance creativity, its implications for labor markets, and the evolving nature of AI in relation to cognitive and creative tasks. The author emphasizes the importance of responsible deployment and acknowledges both the benefits and risks associated with such technology.

The post discusses OpenAI's recent achievement in reinforcement learning, where a team of agents trained using Proximal Policy Optimization successfully played Dota and defeated semi-professional players. This demonstrates the potential of deep reinforcement learning to tackle complex real-world problems through self-training in simulated environments, paving the way for advancements in machine learning and general intelligence.