First and foremost, we can only make predictions based upon the data we have. The data we have is mostly based upon people going for voluntary testing. Given the variability in symptoms that motivate testing, varying levels of awareness to potential exposure, varying willingness to get tested, and the availability of testing, the number of known exposures will make it look like we're in an exponential growth phase even when things are tapering off. It is hard to blame public health organizations for accepting this data at face value since the potential consequences of not doing so are extremely bleak.
The other consideration is that people will modify their behaviour based upon perceived risk, whether or not restrictions are imposed. People will tend to comply with restrictions, go about life as normal, follow some sort of middle path, or impose tighter restrictions upon themselves. It takes a truly special person to put themselves into a position of greater risk. Even then, there is a good chance that they are doing little more than translating one high risk circumstance into another (more or less balancing out the growth). We are facing a situation where scientists can make predictions based upon what is known, e.g. the outcome of restrictions, or making predictions based upon anticipated behaviour. Given there is not enough data to model anticipated behaviour, the natural response is to rely upon models that use expected behaviour (e.g. compliance with restrictions or no response). Since the average actual response will prefer over compliance, more bleak predictions are made.
Finally, everyone has a different understanding of life now as compared to life before the pandemic. I remember when the coronavirus first caught my attention: it was when major U.S. universities started shutting down. I remember when I first started taking it seriously: it was when my provincial government issued a shelter in place style order. Since that time, I have paid attention to what is happening and thinking about how I should respond. Sometimes it is in an acute manner. Sometimes is in a cursory manner. Either way, I am more likely to notice and respond preemptively to outbreaks. I suspect that many, if not most, people are the same even if their only actions are stocking up to prepare for the worse. Again, this will affect outcomes.
But before, here at least, it was repeatedly underestimating the in the aftermath clearly visible exponential growth, exactly as the r factor is calculated? Which scientist/MD doesn't understand there is a saturation, in the extreme even schoolers understand you cannot have more than 100% infected.
Yes, at some point it is becoming logistic, but the previous waves were not about getting anywhere there, but going from 0.1% to 1% or more, and there it was truly exponential growth and overwhelming the health care system, what the start of that logistic curve is about..?!