I think a lot of people overestimate how easy it is to make a search engine, and underestimate how hard it is to sell search to advertisers.

I think this is closer to a 50B problem than a 1B problem. Not impossible, but I don’t know anyone who has 50B that they would want to spend fighting uphill to kill google.

I find this really hard to believe.

Say, you manage to recruit the top-30 engineers who know all the in-and-outs of the google search engine. Add to that, their recent replacement of trad AI methods with full NN I believe, and I can't realistically believe it's not more than a 1B problem.

One of the reasons Google is going horizontal because their search engine can't be "magical" forever, this applies to their ad platform too.

Here are just a few of the things you aren’t considering:

* how many people worldwide keep their google places information up-to-date and not their local government office or their website.

* how advertisers can get value for their search ad campaigns; how to convince them of this

* how much google ad/targeting intelligence comes from off-site (eg Adsense, partner programs, xml ads)

* how many websites and website owners focus obsessively on their google presence making sure that even where Google’s tech gets it wrong they can get it corrected.

* how valuable the feedback loop is: google ads literally tells google what pages and terms are useful, what people like, and what to do more of at scale

* giving away free email and chat to build a whitelist of “good shares”

And that doesn’t cover video or the inverted index you need to provide governments access to emails and search terms. And so on.

These are largely not-software problems.

Apple has all that data, if not better, on a much more valuable audience at that, and can market themselves as "privacy friendly" because only they will leverage said data and won't let anyone in on their platform, and naturally you will trust Apple.

Hell they can even throw a spin on how it's about "privacy" and not about taking another way to bank on their walled garden monopoly and the Apple crowd here will gobble it up with "we are finally free from Google spying, praise Tim and his holy father Steve in heaven".

I've been arguing app stores are a monopoly for years now and the amount of cultist here that were like "Apple is doing it to protect me from bad software" and "I would not use my iPhone if there was an option to install a different store" was staggering.

Sure. Let's say Apple users gobble it up.

Apple's shareholders won't.

They'll want to know where their 10bn (ish) USD went that Apple currently gets from Google.

Is Apple going to increase their device sales substantially because of an Apple search engine? I think everyone who is going to buy Apple has bought Apple, so I think not.

On the other hand, if Apple sells ads on Apple search, and they charge the same fees Google does, they'll still need 50bn in turnover to drop Google; Advertisers aren't going to double their marketing budget, they'll split it based on the ROI, and unless Apple ads convert as good or better than Google, this is a net loss.

Remaking Google requires engineering talent and lots of clickstream data (ie. Hundreds of Billions of search queries and information about what search results were clicked by which users).

Nobody else has this, and until they do, they won't be able to recreate Google.

Is it not like statistical certainty: i.e. you need a million or so queries to cover the range of your possible outcomes, and the hundreds of billions just takes you from 0.999999 certainty to 0.9999999999999999999?

You do NOT get added benefit from big data by taking it from millions to hundreds of billions. It becomes just a lesson in scale, without bringing any added benefit of insight.

With ML methods, they do benefit from taking data from millions to billions.

You can see that yourself with Google by searching for something obscure (eg. "how to unblock a drain jammed with cat fur") and getting three friends to click the 2nd search result. Come back in a week, and suddenly the 2nd result is now the first result! (ignore the videos and answer box)

It turns out just three new data points is enough to influence ranking for that query. It even affected related queries if you change the word "jammed" for "gunked".

It's not a tech problem.

Google is valued like "real estate", it's the best address. And then you need the advertising network for it be worthwhile.

Apple, like Microsoft, can replace google as the default search, but that alone is not a big chunk of the market. It might not be worthwhile.

Otherwise, of course it can be done, and for even less than that.

If some company can sell something to people, even while perfectly fine analogues exist, and then charge double price and have people scream in delight about it - that's Apple.