OpenAI has officially released Dreaming V3, a major upgrade to ChatGPT's memory system that utilizes a background synthesis process to automatically organize and update user preferences, project contexts, and time-sensitive data. Unlike the previous static memory which required manual 'remember' prompts, this architecture proactively archives stale information—such as past travel dates—and prioritizes current project constraints. The system boasts a 5x improvement in compute efficiency, allowing OpenAI to expand advanced memory features to free-tier users for the first time.
The AI industry is experiencing a 'vibe shift' from reactive chatbots to proactive, agentic systems. Major players like Microsoft and Anthropic are pivoting toward 'persistent agents' that live within operating systems and enterprise stacks. These agents, such as the newly unveiled Microsoft Scout, move beyond simple Q&A to observe user behavior and execute multi-step workflows like meeting scheduling and task tracking without explicit human prompts.
Anthropic has reportedly reached a valuation of $965 billion, surpassing OpenAI's $852 billion to become the world's most valuable AI startup following its landmark IPO filing this week.
Which new OpenAI architecture allows ChatGPT to automatically synthesize and update user context in the background?