The price of IBM changes from second to second, but there's no way that actual news about IBM is coming out that fast. The information available about IBM changes a lot more slowly than its share price. If I am to believe that the market is setting the price at exactly what IBM is worth, how can the "true price" of IBM be changing so quickly?
You‘re right. The "true price" should only jump when news arrives but in practice, market participants need time to arrive at a new equilibrium, i.e. the market needs some time until it is clear how the news affects the price. Therefore, you see more adjustments taking place and more trading after arriving news.
The description above gives some reasoning why prices changes in a more noisy way and don’t simply jump. After all, it’s just a theory and does not translate to real life with 100%. Indeed, you could view the stock price as a noisy proxy for the true price. This is in particular true for high-frequency data where there is a lot of market microstructure noise.
It depends on which form of EMH you're considering as to provide a rationale.
Strong-form EMH could or would assert there's private information potentially changing hands that agents are acting on to cause the price movement.
Semi-strong form EMH would suggest all available public info (eg, the news you reference WRT IBM) is immediately incorporated into share price, nullifying any ability to extract a profit. The reality, as the other user noted in his/her response, is that this information is often incorporated imperfectly in practice, whether that be at a lag or in an overeaction at its release.