Primaries are notorious for bad polling. What it basically comes down to is that the voting model for a primary is hard to determine since it has such a low turnout. This is not a new phenomenon and pundits have routinely failed to predict primary results. Things get much, much better when you are talking about a general election.
Adams' "prediction" was more a restatement of what he wanted to see happen and once it comes to pass you seem to think it is an indicator of some keen insight rather than blind luck combining with wishful thinking. Most pundits work with anecdotal models about what they think will happen and things which break outside of that model cause problems (c.f. the mea culpa from Nate Silver). No, this election is not "different" other than having one of the worst candidates possible manage to come out on top of the Republican primaries.
As far as Adams himself, I think you are misunderstanding what an ad hominem attack is. I noted that Adams' loathsome misogyny led him to his particular prediction, which is completely accurate. If I had instead stated "Scott Adams is a bad driver because he is an MRA douche" then I would be engaging in ad hominem. I this case his misogyny is what is leading him to think that a mere woman could never convince people to vote for them, because (like children and handicapped people according to Adams) he feels that women are treated differently and no one would ever let a woman do something crazy like lead a superpower. His grotesque personal views colored his prediction to the point where there is no difference between the two.
The thing about Trump is that there is no wider demo he can appeal to. If he was capable of pivoting he would have done so already once the primary was locked up. You are reading in to him a populist touch that he simply lacks; like most "reality" stars what you see is what you get. There is no hidden depth to Donald, or we would have seen it already in his long history in the spotlight. What is most amusing is that even when smarter Republican politicians try to advise and/or course-correct for his bumbling misstatements he ends up attacking them. Trump does not have "populist skills", he just has a Mr. Angry persona and a small rump of the Republican party that think this is just what they were looking for in a candidate. The problem for Trump is that 30% of the smaller party in American politics is not enough to get the job done.
Of course, this is just looking at Trump and not at the map. The map is what matters. Presidential elections are won on a state-by-state basis and here things look even worse for Trump. In these sort of analysis you start with the previous presidential results (the prior for this experiment) and then try to determine what states this new candidate can put into play to change the outcome of the election. Here is where Trump goes from failure to embarrassing punchline.
The short version of electoral analysis is that Trump might be able to put Pennsylvania, Maine, and possibly Ohio into play, but that doesn't get him to the finish line. In contrast, Clinton is polling well enough that Arizona and Georgia are in contention. OTOH, you could also just look at the state of either campaign to see that Trump has no ground game, no organization (seriously, he has less than 100 people working on his campaign), and very little money.
If you want predictions then here is one: Clinton is going to put Texas into play this election cycle. Maybe not win it, but come close enough that the RNC and Trump will have to start running ads.