I've always chosen frameworks for their abstractions and their design decisions, rarely their performance. Great research by the author, I feel better about that decision now.

Frameworks do matter in the sense that a good, well-organized, well-thought out framework will always outperform a shabby one. And the throughput matters. But having said that - it's a fine balance.

We moved one of our customers from a vanilla wordpress backend (gotta be the shittiest code organization I've ever seen) to a custom Phoenix/Elixir based backend that handles over 500M+ requests at 1/3rd the cost of competitors and 40% cheaper than their old Wordpress backend. So yeah, it really does affect the bottom line.

“ No database, no ORM, no serialization. Pure framework overhead.”

And many more examples. Great, another AI written article.

Would like to see the prompt that generated this article because it's narrowly looking at underlying base framework performance as its only metric, yet buries that lede very well (at least in prose - obvious enough from the data of course).

Of course the conclusion is wrong though - looking at the data I would conclude that it's base framework performance that "doesn't matter" as a metric rather than framework choice itself.