There's just one thing missing...

It's $$$O(1)$$$ vs $$$O(\log n)$$$ on average. Obviously, that difference will matter on some problems. Avoiding hashed types in all cases to "be safe" will cost you eventually

That's a bit extreme though. You just need a custom hash function most of the time

For your second point, no need to cast it. You can just use cout << fixed << your_double next time.

Immaculate drawing.

Agreed, though, technically, the halting problem is decidable for physical computers because they have finite states and thus aren't actually Turing machines.

They'll just be on top of each other (or theoretically perfectly overlaid if you get +0 on the second one)

Are you using GCC C++20 on your machine?

I looked at both submissions and I must say that it does seem pretty believable, at least to me, that this is a coincidence (I just chanced upon this comment while idling around recent actions).

Better this than having a contest become unrated because of technical issues, though.