"Mission critical workflows" SHOULD NOT be reliant on a LLM model.

It's really curious what people are trying to do with these models.

Anthropic has good models but they are absolutely terrible at ops, by far the worst of the big three. They really need to spend big on hiring experienced hyperscalers to actually harden their systems, because the unreliability is really getting old fast.