> Can it be used to make predictions about the future, for example?

Yes: most invested dollars will be unable to beat the market, on a risk adjusted basis and net of costs.

Aggregate predictions, yes. For example, the long-term, risk-adjusted expected return of all liquid securities should be the same if we believe EMH. Therefore by holding a diversified sample of securities in a market, your expected return should be the same as that of the market as a whole (though actual return won't be). Any expected difference would theoretically be due to a difference in risk.

Now, it may be that the risk is one that other people care about but you don't. Some investors might be more sensitive to short-term volatility than others, for example. But a weak EMH can at least give a framework for thinking about these kinds of decisions. (Like whether to value tilt, for example. Or whether to invest in active or passive funds.)

It also gives you a framework to think about in which cases it is more or less likely to hold. Small, illiquid markets are the least likely to be efficient, in my opinion and experience. My only actively managed investment is in such a market.