Look, I mention unsafePerformIO in the post in passing as an example of a thing that could hypothetically wreck my codebase, with or without an effect system. The prime example is actually supposed to be lazy evaluation, since that’s the most realistic one, but I padded it up to three for decency and added a footnote about Safe Haskell. I can remove the mention of unsafePerformIO and it won’t change the meaning of the tangent, let alone that of the post.
If I’m to believe Reddit comments, this one point is easily the most important part of what I wrote, and, in fact, I must not care about type safety whatsoever if I dare utter such an outrageous claim.
The title is the exact same case. I never say that my solution represents every possible use case, or that I’m totally correct on all of this, or that what I wrote is final. The very reason I’m making this post is because I want people to tell me if my thought process is correct. The title is my conclusion as a sum total of all of the observations I make. I shouldn’t have to put “I think” before every single thing I write in my blog.
The problem has barely anything to do with statements themselves, it’s with people isolating them and giving them the most uncharitable readings imaginable. And the lack of insightful responses (except for @tomjaguarpaw’s unyielding fervor) is not leaving much room for me to speculate as to why this is the case.