Comment by omoikane

> they could make a different test

The takeaway I got from the article is everyone can make their own test, as opposed to relying on other people's sentiments and memes about X is bad or Y is good.

Trying to emulate a non-expert experience without workarounds is not the common usage pattern since everyone familiar with their favorite tools have ways to get more value out of them, but this article presents a way of constructing an experiment (this is why I chose these queries, this is how I ranked scams, etc.), and I think people should follow this same spirit to evaluate if they are stuck in a local optimum with their current choice of tools.