Guessing the future is easier than writing it. Steve Job did both. He knew what people wanted and built it for them.
Most people are either one of the other: analyst/forecaster or developer. The analysts' information generally direct the developers what to do, mostly because the developers just want to do the work. They want to think in code. But you can't waste a developer's time building something outdated.
Research and development flock together, ideally as one, where forward thinking meets predictive doing.