I have a self imposed rule about inventions, that probably applies to predictions as well. It's based on the idea that a patentable invention consists of more t

I have a self imposed rule about inventions, that probably applies to predictions as well. It's based on the idea that a patentable invention consists of more than just an idea blurted out by some manager (I was the manager when I arrived at this), but needs to be novel, non-obvious, useful, and reductible to practice. These ideas are usefully applied as a self discipline, completely aside of the idiosyncrasies of the existing patent system. Non-obvious means that a generally smart person could not have come up with the same idea by combining existing knowledge. Reductible to practice means you have to provide a recipe for actually making it. For all practical purposes, you don't know if it can be reduced to practice unless you've actually made the effort and discovered all of the gotchas. In addition, a prediction needs a timeline. "An asteroid will strike the earth" is an empty prediction without quantifying the time scale in terms of something like a probability per year or an expected duration. This is why I dismiss most futuristic predictions as not qualifying as predictions or inventions. I think my rule also works for disposing with giving managers credit for "ideas" just for showing up at brainstorming sessions, at the expense of the people who are actually doing the work.