If you take a look at the poking holes video, the complaints can all be lumped into complaints that his argument is an incomplete bayesian analysis of a cost-benefit basis for decision making. There are certain assumptions which can be made to make the argument completely rigorous and the table completely accurate, they are implied by the broad language used in his definition of "True" and "False", the unspecified actions A and B, and the outcomes chosen for the various possibilities.
The long and short of it is that there is a lot of unspecified semantics which causes unnecessary disagreement when the actual arguments should revolve around likelihoods on impact assessments.
That is to say, some people will simply attach zero [prior] probability onto the outcome of successful action and true climate change and will subsequently be unmoved by the argument.