The Future of AI

Charles Platt observed that all forecasts for transformative AI are about 30 years away from the forecasting date. If you don't have any idea and just want a signal that AI is far but not impossible, 30 years is a great guess. But what kind of distance from the present sounds reasonable in quotes? What numbers will correctly signal support for science and innovation without making it sound like a rosy-eyed optimist who expects miracles?