The whole problem is the adversarial environment. It's orders of magnitude harder to create genuinely useful content than it is to create crap or copy content.

The whole problem is the adversarial environment. It's orders of magnitude harder to create genuinely useful content than it is to create crap or copy content. Meanwhile, the difference between useful and lazy content requires a holistic view and even embodied (IRL) understanding. For example, it's so much more work to physically review a set of items and write that up than it is to just read and summarize other people's reviews (bad content). It's also hard to tell the difference between these without physically trying the item in real life (bad ranking). So the ranker can't tell what's crap and there's more crap than good stuff. If there was no money to be made by ranking highly on search engines, I think the promise the article's talking about might have been fulfilled.