The business insurances are in is a business of statistics. As long as you can model things giving you an expected value and a standard deviation, you can offer an insurance policy which gives you X amount of profit with Y amount of risk, and the insurance premiums are adjusted such that the insurance's risk for negative profit is negligible, according to the model.
What does it mean for climate change? Current insurance models apparently don't work well, so they don't dare to offer policies in certain areas. But just like city planners need to adjust (build further away from shore, higher up, build in flooding protections) and home owners do (AC, think twice if you want a basement) and farmers (choice of crops, irrigation systems), so do insurances by finding better models that allow them to have better statistics.
My expectation in the long run is that insurances will be offered again, but with so high premiums for certain areas (of high risk) that it will just be too expensive to live there. Which is fine. Nobody lives on the moon either. And the public shouldn't be paying for somebody's privilege to have a nice waterfront property in a hurricane area.
TL;DR: The current public discourse about this topic conflates predictability with cost when talking about "insurability". They are very different things.