Sam Altman believes the next major breakthrough in artificial intelligence will not come from better reasoning alone. Instead, the OpenAI CEO says long-term, near-“infinite” memory will define the future of AI systems.
For years, AI development has focused on improving how models reason and solve problems.
Altman argues that this approach may be missing what users actually value most. According to him, AI becomes truly useful when it can remember people over time.
Altman shared these views during a podcast conversation with Alex Kantrowitz. The discussion covered AI memory, infrastructure growth, AI hardware, and the long-term pursuit of artificial general intelligence. Altman said the ability for AI to remember user preferences, habits, and past conversations creates a more meaningful experience.
He described memory as the feature that allows AI to feel personal rather than transactional. In his view, intelligence without memory limits how helpful an AI system can become.
While acknowledging major improvements in ChatGPT’s reasoning abilities, Altman pointed to memory as the real turning point. He said persistent memory helps AI systems grow alongside users instead of starting fresh each time. This shift, he suggested, could matter more than marginal gains in raw intelligence.
Altman’s comments signal a broader strategic direction for OpenAI. Rather than focusing only on making models smarter, the company appears to be prioritizing continuity and personalization. If successful, this approach could reshape how people interact with AI on a daily basis
