[tablespoon]

>> Google's search got gamed by Pepper SEO, the US is not suddenly more more interested in a underwear brand than the food.

> I don't think it's "think more about" but "want to see that result for this query", those are slightly different things.

That can be true, but I think it's the kind of explanation that has to be proven beyond a reasonable doubt to believed. Unless you can do that, I think the assumption when a brand overtakes the generic term should be it's due to a SEO gaming.

So Amazon may truly be a legitimately more popular search than the amazon (because it's an online store that's so popular it's arguably monopolistic), but Pepper (apparently a niche bra for "small-chested women") is almost certainly not legitimately more popular than pepper.

[tazjin]

Intuitively I found it not that hard to believe that the number of people who just google "pepper" to find something about the fruit or spice is fairly small.

However, a Google Trends query[0] (limited to the USA) indicates that the growth in "pepper" queries is pretty constant (the bra brand apparently launched in 2017). That makes intentional SEO gaming more likely again.

[0]: https://tgsa.tazj.in/img/tazuploads/24/0

[tablespoon]

>> That can be true, but I think it's the kind of explanation that has to be proven beyond a reasonable doubt to believed. Unless you can do that, I think the assumption when a brand overtakes the generic term should be it's due to a SEO gaming.

> Intuitively I found it not that hard to believe that the number of people who just google "pepper" to find something about the fruit or spice is fairly small.

But that's not anywhere "proven beyond a reasonable doubt," rather it's the very low standard of "I can imagine it maybe being true." You provided good evidence for SEO gaming, but I don't think that was strictly necessary given where I think the burden of proof lies.

In general, I see a lot of thinking that takes "whatever a FAANG does" as an oracle for truth, so spends a lot of effort to justify their results as correct. I think that's wrong (at both the factual and moral levels), and also creates a tolerance for regression, so it should be disputed.