Here's a distinction I find clarifying: you can predict technology, but you can't predict science.
Future scientific knowledge — the next genuine discovery — is unpredictable almost by definition. If you could predict it, you'd already have it. But future technology is a different animal. Once the underlying science exists, the engineering that follows from it is surprisingly forecastable. If something is scientifically possible, it's usually just a matter of time before someone builds it. Leonardo da Vinci is the proof: he sketched flying machines and tanks centuries before the materials and methods caught up. The ideas were sound; the world just hadn't filled in the engineering yet.
What's changed is the clock speed. Progress builds on itself — each new layer of ideas and tools makes the next layer arrive faster, so the curve isn't linear, it's multiplicative. Nearly every technological trend, when you actually plot it, bends exponential. That's why da Vinci's ideas took five hundred years to become real, while an idea like molecular nanotechnology — at least as far out, conceptually — might take only a few decades. The science is the bottleneck. Once it's through, the technology comes fast, and faster every year.