Putting more energy into a meta-stable chaotic system will greatly increase the number of extremes across the board. Models and forecasts are currently mostly focused on how the various averages are trending, which is fine and useful.
It's most correct to plan around every location getting more wet, more dry, more hot, more cold, more storms and more droughts.
It's not only about the climate but also the rate of change of climate. When the climate changes very quickly, ecosystems cannot adapt and they die.
Remember that humans are actually very adaptable, our range of habitats extends from the arctic to the tropics, desert to rainforest. Most life is not so tolerant.
Indigenous peoples have been living in all corners of the globe for thousands of years. The range of our habitat has stayed roughly the same since antiquity, but the population that any given land can support has been vastly boosted by technology.
Both of these things are true, depending on what is meant by our range of habitat. A small number of people can survive just about anywhere just as you describe, but current population densities are not sustainable in many places without cheap abundant energy.
(I’m not actually concerned with us running out of cheap abundant energy; PV is already fantastic in this regard, and in a hypothetical collapse of all dispatchable power sources that included never being able to make more batteries, we would be severely inconvenienced rather than utterly finished).
A chaotic system is generally defined to mean one in which small changes to the inputs lead to large variations in the outputs. So your reply noting that it’s only a 1% change doesn’t contradict the previous statements.
It depends on the system, I’m not familiar enough with climate science to state anything about the particular case at hand, but chaotic systems can be deterministic.
It's most correct to plan around every location getting more wet, more dry, more hot, more cold, more storms and more droughts.