ChatGPT Still Struggles With Telling Time, Sam Altman Says Fix Could Take A Year

ChatGPT has come a long way, but it’s still not perfect and one of its quirks has recently gone viral. A demo video making the rounds online shows the chatbot getting something surprisingly basic wrong, tracking elapsed time. The clip caught enough attention that even OpenAI CEO Sam Altman weighed in, explaining why the AI still struggles with something as simple as timing. Keep reading to understand what’s really going on behind the scenes.What Sam Altman Actually SaidResponding to the viral example during a podcast conversation, Altman called it a “known problem.” According to Altman, fixing this isn’t going to happen overnight. In fact, the OpenAI CEO hinted it could take up to a year before things improve meaningfully.He explained that the voice-based

version used in the demo doesn’t actually “run” a timer. It simply talks like it does. That means when you ask it to measure time, it responds conversationally rather than performing a real-world function. Altman did mention that future updates could bring more practical capabilities to these voice models, but for now, limitations remain.

Why ChatGPT Gets Time WrongThe issue isn’t exactly a bug, it’s more about how these systems are built. Large language models like ChatGPT don’t interact with real-world tools unless specifically integrated. They don’t check a clock or start a stopwatch. Instead, they predict what a correct answer should sound like based on patterns they’ve learned. So when asked to track time, the model essentially “guesses” a response that feels right, even if it isn’t actually measuring anything.The Bigger Concern Around AI ConfidenceWhat made the viral demo more interesting and slightly concerning was how the popular AI chatbot reacted when questioned. Even after being challenged, it continued to insist its answer was correct. This isn’t unique to this case. AI systems sometimes present incorrect information with surprising confidence. It’s a pattern that has raised eyebrows, especially in fields where accuracy really matters, like healthcare, finance, or legal advice.