Even a million tasks a day is less than 12 a second. Most queues are going to have surges since that's part of the point of a queue, but it's still a few orders of magnitude away from what should overwhelm a database.
Just use a dedicated tool. It is not that hard. If you want higher level abstraction, you have a whole spectrum of next gen queues, like Temporal, Trigger.dev, Inngest, Defer, etc.
Why use a dedicated tool if you have something in your stack that can solve the problem already?
The less separate pieces of infrastructure you can run, the less likely something will break that you don't know how to easily fix.
The article touched on this in the list of things to avoid when it said "I estimate each line of terraform to be an order of magnitude more risk/maintenance/faff than each line of Python" and "The need for expertise in anything beyond Python + Postgres"