The world without causality is quite easy to understand: all we can really ask of reality is that whatever happens in the future must be statistically consistent with what has happened. That is, any future explanation of any phenomena must also apply to any past phenomena which qualify for the explanation. All that is required to satisfy this kind of explanation is the logical procedure required to obtain the result; if that is the case, "causality" is not an issue at all.
This Humean approach to causation has one serious flaw. It is just as valid to say that any past phenomena must apply to any future explanation, or for that matter, any phenomena that is in anyway associated with the past and future events. For example, George Washington's picture being on the dollar bill can be considered as an effect of this man's influence on American history. However, if we remove causative accounts as valid and look to statistical accounts, then one could say equally that George Washington's influence on American history is because his picture happens to be on dollar bills in the present time. Of course, we could not talk in terms of cause and effect at all and just say that George Washington as a historical figure is also statistically linked to his picture being on dollar bills, but then why should these be linked? What if I happen to draw a mustache on George Washington's dollar bill picture today, is that also statistically linked with George Washington? Why? What if I choose not to draw the mustache, does that mean that this potential event is not statistically linked to George Washington, but just my thought of the action is statistically linked?
Obviously, any attempt to explain why unrelated and supposedly related events are statistically linked is deemed a non-issue by an anti-explanationist account of causation. We are simply asked not to ask why there are statistically links, they just exist. All of our decisions, then, are really without cause, they are all statistically linked or they are not statistically linked and that's the end of that. Which, of course, raises a paradox for a statistical argument. If the statistical argument was originally intended to explain why it is that events appear to occur as they do (i.e., because a statistical relationship exists), and we find that there is no reason for the statistical relationship (i.e., no cause), then, in fact, there are no true statistical relationships since there is no explanation of cause which ties one event to another. So, the paradox is that a statistical account cannot explain anything since there is nothing to explain! All a statistical account of causation can provide is 'patterns' of occurrence that are only observed patterns having no reality whatsoever. A pattern is only a real pattern if there is an algorithm that exists which proves the pattern is to be treated as a discernable whole. If after finding there is no cause or explanation for the pattern that can possibly exist, then all we have is the illusion of an algorithm where none actually exists. Hence, patterns don't actually exist by this account, and therefore any appearance of a pattern is completely random in nature (assuming that we can accept that random patterns don't also conform to a pattern that actually exists).
Beyond just this kind of paradox is the whole ridiculous concept that we don't actually cause events to occur but they are all just statistically linked for no cause as part of a random mesh. Even our thoughts are part of this random nexus in such a view.
The degree that some people will go in promoting their ridiculous models of reality having no empirical support is literally staggering.