In all honesty, the generated code mentioned in the blog post is more of an industry standard at this point rather than anything else - which is a bit of a sad state of affairs.

It's just that previously snippets like this were copied from Stack Overflow answer section or some random blog post, whereas now the LLMs which have been fed that content simply regurgitate it.

Probably neither the first nor the last time an "auto-complete" will be producing straightforward insecure code.