[rewmie]

> Heh, I recall seeing many posts arguing against benchmarks (...)

It's one thing to argue that some real-world data might not be representative all on itself.

It's an entirely different thing to present no proof at all, and just claim "trust me, bro" on marketing brochures.