> It's not that clear how much that helps.

It's clear that it helps enough to subsidize the entire rest of the Google operation (enough to run a more computationally and sophisticated business like search). AdTech is all about targeting by extracting context and minimally about tracking events.

Leveraging search to power ad targeting is the most efficient means. I'm not sure how this is a topic of debate other than to split hairs on what measuring stick to use.

Translate at least seems to have some people ignorant of the alternatives convinced that they have a clearly better product — “no competitors that even come close in terms of quality” when they objectively don’t. https://news.ycombinator.com/item?id=23975001

The claimed “objectively” better translator is DeepL.

You need to read like 8 paragraphs deep to get to that and there is no real explanation as to why it is objectively better - just a link to an article where some journalists apparently thought it was better.

https://www.deepl.com/en/translator

The bottom of the page you linked shows ten articles, six of which praise DeepL as better than the competition and four which praise it without comparison. The least positive is probably Wired’s: “WIRED's quick test shows that DeepL's results are indeed in no way inferior to those of the high-ranking competitors and, in many cases, even surpass them. The translated texts often read much more fluently; where Google Translate forms completely meaningless word chains, DeepL can at least guess a connection.” — which is a long-winded way of saying that DeepL’s worst translations are as bad as Google Translate’s, but many of them are better — making DeepL strictly better overall. Of course, any individual review must be made by subjective humans, but you can find as many corroborating comparisons as you like and every review I’ve seen has either described DeepL as better or at worst as bad as Google Translate, and the former is exactly my personal experience after using each for years. We can also compare the results of sending your comment on a round trip through the top language on each:

DeepL German: “The allegedly "objectively" better translator is DeepL.”

Google Translate German: “The alleged "objectively" better translator is DeepL”

DeepL Spanish: “The allegedly "objectively" better translator is DeepL.”

Google Translate Spanish: “The best translator "objectively" that is claimed is DeepL.”

As a native English speaker, the DeepL results seem clearly clearer and cleaner to me. If broad consensus isn’t evidence to you, you could also compare quantity of languages instead of quality of translations, in which case you’d find that Google Translate is missing languages that Microsoft Translate has. Either way, Google certainly doesn’t have “a pretty clearly better product.”

I experimented with testing with this line of Japanese.

フクロネコの仲間の中でもっともふ地いことで知られています

DeepL: It is known to be the most ground-breaking of all the bovine cats

Google: Known for being the most remote of all quolls