[basic.def.odr] lambda-expression example seemingly constradicts normative wording

On one hand, lambda-expressions in default arguments are not required to have the same closure type:

In each such definition, except within the default arguments and default template arguments of D, corresponding lambda-expressions shall have the same closure type (see below).

http://eel.is/c++draft/basic.def.odr#14.6

On the other hand, the example shows the opposite:

inline void g(bool cond, void (*p)() = []{}) {
  if (cond) g(false);
}

If the definition of g appears in multiple translation units, the program is ill-formed (no diagnostic required) because each such definition uses a default argument that refers to a distinct lambda-expression closure type.

http://eel.is/c++draft/basic.def.odr#16

I can't make sense of it; it seems like a direct contradiction to me.

Start by comparing the linked [basic.def.odr] paragraphs with the lambda-expression wording and the g example. Determine whether the apparent conflict is resolved by context or requires a standards-draft clarification; done means the normative text and example no longer appear contradictory.