Google, in Rare Stumble, Posts 23% Decline in Profit

Also the 855+ has better 3D results at times. None of this is here to imply the A1x chips aren’t crushing their android counterparts at single threaded performance. They are likely wayyyy faster, but extra cores, sustained performance, 3D performance, and LTE performance all factor into the over performance of a device.

In 2019, Google Web Search optimised for ~10 billion users.

Expanding self-selected elite groups rarely sustains intelligence and skill levels.

This isn't a swipe at general intelligence, it's just a flat fact of mass media, one that's been widely recognised (though often lamented on a moral basis) for well over a century.

I've labled this "the tyranny of the minimum viable user".

The other factor is that manipulating the medium has become increasingly attractive. That's not just old-school SEO (black or white), but propaganda, psyops, and similar activities. Google are very much part of this battleground, and seem to have taken considerable time coming up to speed with even grasping this. I suspect they're now aware, if not sufficiently competent.

A friend, Woozle, has described this as a fundamental epistemic paradox as systems grow -- what was previously useful for a few becomes increasingly attractive to manipulators. Though Shoshana Zuboff's Three Laws predate their observations by about 30 years.

Also given the fact that I worked for an SEO company I can definitely tell you that’s bot the money-making enterprise anymore since it mostly consisted of opportunists who happened to know about internet before anyone else before.

That is the whole point, old Google was easy to replicate and therefore didn't cement their market position, if they still had that then why not just switch to Bing or DDG? New Google requires thousands of top level experts to tune and test things until it gives sensible results to stupid queries and is therefore impossible to compete with.

I long for the return of an Altavista-like "dumb" search engine with proper "hard" boolean support. My brain is hardwired to think that way and I strongly prefer it. I understand that this wouldn't be useful for everyone, but I don't care. How on earth does it make sense to have only one kind of search engine for all of humanity.

I agree 100%. Sometimes though it feels like the Web itself is deteriorating as the bulk of new Internet users increasingly only care about a handful of social media platforms. Where's the "sources of truth" about website quality when we need them? We used to have DMOZ/ODP providing something like that, it had a lot of problems but by and large it worked. But now DMOZ is dead, and there's nothing as of yet to replace it.

To be honest, I don't understand how everyone can remember the "quality" of their Google searches from 7 years ago. I can't even remember the quality from a year ago. The things I find most frustrating with search (such as maternity blogs being the first options when looking for medical information related to a newborn or tutorial sites being at the top of the results rather than the official language docs) rank highly because those are the things people are clicking on.

As an example, I think the crown on HN is more likely to read the official documentation rather than go to a site such as TutorialsPoint or StackOverflow. On the other hand, the average programmer most likely just wants to go to TutorialsPoint or StackOverflow and copy-paste the solution they are looking for. And I think there are many more average programmers than there are visitors to HN. From Google's point of view, they are giving the user what they want.

All that being said, it is easy to criticize Google, but dealing with search at the scale of the internet is non-trivial. It isn't clear to me that Google's algorithms from 7 years ago would do just as good of a job today as they did 7 years ago. There is a lot more junk and abuse online, people are figuring out ways to hack their search ranking or spamming useless articles.