Re: oh there are differences
There are differences between Y2K and Climate Change, and I suggest that one of those differences has trapped your thinking.
Nassim Taleb of Black Swan fame essentially cries foul on making predictions or even trying for repeatable tests within extremely complex systems where you cannot possibly know all the inputs/outputs, particularly over time. He calls this Extremistan. Within bounded systems where you can understand what is going, predictions are safe. He calls this Mediocristan.
Your testable hypothesis position is fine in Medocristan, and for structured experiments, but falls over in Extremistan.
An example of extremistan is the famous Turkey problem. If you are a Turkey bred for thanksgiving then for most of the year you are happy, well fed, and protected. The farmer who raises you is your friend and carer. Then on Thanksgiving he chops your head off. Repeatable observations and prediction confirm all is well - until they don't.
Climate Change is Extremistan. Y2K was Mediocristan (in a symptomatic sense anyway, not necessarily in the outcome after the deadline - but we could tell that the clocks were going to stop working, and when and how they were going to stop working. Essentially we knew that thanksgiving was coming and that on that day something bad was going to happen, but not what that would be. We don't have that luxury with Climate Change).
In Extremistan the best way to proceed is avoid messing with the system you do not understand, and also to try to not leave yourself vulnerable to whatever strange things will eventually befall you - design for robustness or even better, antifragility (a subsequent Taleb concept). Therefore I believe the best approach to dealing with Climate Change (which I accept as occurring, and that acceptance is supported by vast reams of published and peer reviewed evidence, even though systemically predictable outcomes and effective modelling cannot be reliably derived from that) would be to try and rein in the potential causes of divergence - in this case excessive fossil fuel consumption etc. How? Renewables and ETS would be a good step, carbon sequestration would also be useful.
Nuclear power would also be an option - but not a preferable one because it has its own known raft of unpleasant issues. It also comes with a lot of fear, uncertainty, and doubt in the public mind, and strong ideological opposition from the more extreme Green elements. It should not be a first choice. It could be a part of an integrated solution or a stopgap. It may yet become a last minute panic option.
Another issue with asking for provable results on Climate Change is that by the time you can really, categorically state we have an issue, the momentum will be very difficult to reverse, and may lead to attempts at direct intervention (eg. particulates) which risk having their own unknown effects. Thus a wait and see approach is terribly risky, whereas a sensible and staged shift to renewables and lower impact manufacturing etc will not only help to address climate change risks by seeking to be as close to the non-industrial status quo as we viably can, but will probably lead to a more robust and competitive energy and manufacturing market in the long run.
I do agree though, that to accept you need to act on Climate Change does mean we take the risk of wasting a lot of effort and time we didn't need to waste. But given the foreshadowing we have, the costs of inaction are truly terrifying. Really. Death on a biblical scale. WW3 level horror.
Lastly, and back on the Y2K issue - I am also very annoyed at people that think it was a hoax. I was a programmer and I worked on stopping that issue - and I know firsthand that we would have had some sort of trouble if no action had been taken, likely big trouble. But because it was dealt with, many people just choose to think it was never an issue at all.
People respect and recognize a disaster repaired, they don't respect and recognize a disaster averted (except in cases where the dodge was narrow and so usually obvious).