OpenAI chief executive Sam Altman has predicted that the next major step toward superhuman artificial intelligence will come when AI systems achieve what he called “infinite, perfect memory.” Speaking on the Big Technology Podcast, Altman said the company is working toward this goal by 2026, envisioning AI that can remember every detail of a user’s life. He contrasted this with human assistants, who cannot recall every word, document, or interaction, describing current AI memory as still “very crude” and “very early.” Altman’s remarks follow reports that he declared a “code red” at OpenAI after Google launched its Gemini 3 model in November, which the company described as marking a “new era of intelligence.” Despite acknowledging Gemini 3’s strong benchmark performance, Altman said OpenAI’s response was standard practice when facing competition. He noted that Gemini 3 had not yet had the impact initially feared but did expose weaknesses OpenAI is now addressing. Altman outlined OpenAI’s strategy to stay ahead in the AI race: build the best models, create strong products around them, and ensure sufficient infrastructure for large-scale deployment.