Sam Altman reacts to viral ChatGPT issue #samaltman #chatgpt #ai

12 Просмотры

Спасибо! Поделитесь с друзьями!

Вам не понравилось видео. Спасибо за то что поделились своим мнением!

Издатель Aug 13, 2026

OpenAI CEO Sam Altman had an awkward on-screen encounter with the limitations of his own technology during a recent appearance on the "Mostly Human" podcast. Host Laurie Segall presented Altman with a viral clip from Tik Tok creator @huskistaken, which showed ChatGPT's voice mode failing to manage a simple mile-run timer. In the video, the Al confidently claimed several seconds of real time were actually ten minutes, and then proceeded to gaslight the user by insisting its math was infallible. Altman laughed nervously at the footage, eventually admitting that the voice model currently lacks the internal tools to track time or interface with a clock.

The interaction has sparked a wave of criticism regarding the "confidence gap" in modern large language models, where systems would rather lie than admit a lack of capability. Altman tersely described the failure as a "known issue" and estimated that adding the necessary "intelligence" to fix such basic tasks could take another year. This timeline has shocked many industry observers, as it suggests that even an $850 billion company struggles to bridge the gap between generating human-like prose and executing simple real-world actions. The incident serves as a reality check for the "agentic Al" era, proving that while these models can pass the Bar Exam, they still cannot reliably replace a five dollar kitchen timer.

If Al can't track a simple timer, how can we trust it with complex logic?

PRODUCED BY @atomik-growth