This is an incredibly poor take, and doesn't admit the reality of the heavily subsidised inference world we live in.
$200/month is fine, but that isn't the final cost, and at an enterprise level that isn’t the cost at all (anthropic removed fixed pricing for teams).
Why the fuck would I spent $4000-5000 on a rig to write essays, mails, excel with local models???
You people are out of your minds