All the important "meat and potatoes" information is in the current closing price at a given time, so basically no matter how much knowledge you gain about the current price of Bitcoin now you still won't be able to predict the future prices and that's exactly what the efficient market hypothesis states and that is also where the term the market is in a random walk all the time comes about, random meaning that there is no pattern that can be recognized easily. At the end of the day the best a trader can do is to guess and guess "emotionally"; but it's these emotions of greed and fear that will eventually cost traders' big time in the long run, so a "good" trader isn't the one who has a breadth of knowledge about the market's historical past, but one who can "control" their emotions effectively. Thus when it comes to AI, AI is non-emotional and only computes, and so AI is somehow able to find some form of trend from very complex data in ways which we humans don't seem to be very good at.