Can I get some help with the Amber rules (Amadio & Cardelli '93) for #subtyping?
I haven't seen mentions that it's possible to derive `μx.x <: t <: μx.x` *for every `t`*.
Here's my proof of the 1st (the 2nd is the same):
```
μx.x <: t ⊢ μx.x <: t
--------------------------
μx.x <: t ⊢ x [μx.x/x] <: t
------------------------------
⬝ ⊢ μx.x <: t
```
What's wrong?
The Amber rules do induction afaik, is the problem that `x` is not under a `+` or `→`?