Amit Singhal, who was Head of Search at Google until 2016, has always emphasized that Google will not use artificial intelligence for ranking search results. The reason he gave was that AI algorithms work like a black box and that it is infeasible to improve them incrementally.
Then in 2016, John Giannandrea, an AI expert, took over. Since then, Google has increasingly relied on AI algorithms, which seem to work well enough for main-stream search queries. For highly specific search queries made by power users, however, these algorithms often fail to deliver useful results. My guess is that it is technically very difficult to adapt these new AI algorithms so that they also work well for that type of search queries.
While the old guard in Google's leadership had a genuine interest in developing a technically superior product, the current leaders are primarily concerned with making money. A well-functioning ranking algorithm is only one small part of the whole. As long as the search engine works well enough for the (money-making) main-stream searches, no one in Google's leadership perceives a problem.
Naturally, this would be a good time for a competitor to capture market share. Problem is, the infrastructure behind a search engine like Google is gigantic. A competitor would first have to cover all of the basic features that Google users are used to before they would be able to compete on better ranking algorithms.
This heavily aligns with my experiences over the past few years.
Where Google once used to return exact or close results for very specific or niche searches, I feel like it struggles to even land in the same ballpark of results. I've been asking myself whether the results have always been this bad more and more over the last year and a half thinking I've been taking crazy pills. Unfortunately, the current competitors still perform even worse for the same queries so it's not like there's enough of a reason to default elsewhere (yet, I hope).
In the past, Google felt more like a sophisticated professional tool. If you made the effort to click through the results beyond the first page, refine the search query increasingly and, on occasion, enclose individual terms in quotes, you sooner or later landed a hit.
Nowadays, the query you enter into the search field is answered by an AI algorithm that works more along the lines of "I know best what you're looking for". For users who are not very knowledgeable about search engines and tend to search rather superficially, this apparently works quite well. For professionals who want to dig out the really interesting hits in the deeper part of the web, an AI-powered search engine of this sort is quite frustrating.
Google has a verbatim search, click on Tools/All Results/Verbatim, but it is no longer really verbatim. It often offers results that have only one of your keywords. DuckDuckGo seems to completely ignore any quotes I put around keywords.
> Naturally, this would be a good time for a competitor to capture market share. Problem is, the infrastructure behind a search engine like Google is gigantic. A competitor would first have to cover all of the basic features that Google users are used to before they would be able to compete on better ranking algorithms.
That is why the basic rule of dethroning an incumbent is to not go into competition heads-on. If someone is successful in shaking Google, then it won't be a search engine company. That company will build some different product first not competing with Google at all, become a leader in that and then, if business model supports, build a better search engine. It is a long process, but that is how it usually works. Google is a classic example of this way of getting at Microsoft. After search, it tried to eat Microsoft's lunch by developing products to compete with MS-Office, and also ventured into mobile and laptop OS.
To this point - it is already happening to certain extent.
Amazon's share of product search is more than that of Google's [1]. And earlier Facebook started getting large chunk of ad money, which would have otherwise gone to Google.
I wonder if those newer algorithms are the reason I sometimes get search results that seem to have nothing at all to do with what I asked for.
Of course the older search algorithms are also AI algorithms; just not the black-box machine learning algorithms that are so popular in recent years.
My theory is that Google's algorithms care too much about what the average person wants, and not enough about what I want. I'm not the average person. And I thought Google would know enough about what I want by now.
What I really want is more personalised search where I control how the algorithm is tweaked. What would also be useful, and I really can't believe Google hasn't done this yet, is voting on search results. Those content farms would quickly be voted into oblivion.
Well, we know Google "salts" ML processing to get the results they want for censorship of all types. That seems likely to screw up objectively valid or honest search results pretty easily.
The real question should be is altavista still working?
I'm too scared to look.
Independent webmaster days used to be the best times... Now hackers and corporations run the show. Many are complaining that SEO is getting trumped by outright payola... cough
Things like Net Neutrality and paid promotion have overrun (free and independent thought) search results now and clogged them up with sales funnels to the point that we're just going to get more and more irrelevant results until people are going to have to turn to libraries again to find meaningful and succinct results to questions. AI is in it's infancy and usually geared towards profit for google more so than for human good from what I can observe, because especially during pandemic times, they have pretty big bills to cover.
The underlying problem is the monetization of information. As long as we keep driving this trend truth and accuracy of information will suffer deeply. It's best to keep monetization to less critical resources like entertainment and physical products, they can afford to be sensationalized and monetized more than things presented as science, facts, and credible news.
This trend ruins the value the Internet once had and also steers intellectual value/power back to universities interestingly though, where high tuition often provides learners to more carefully planned presentations and usually more emphasis on accuracy. The future will be expensive for all of us.
Say I search for a chicken piccata recipe. Which pays Google more:
- a static blog from 2003 without ads but an excellent recipe
- a YouTube video or promoted article with ads enabled who paid Google to feature them on this query and who will allow other advertisers to market to me on their site
You might say “well they already paid Google so Google is making money either way,” but people will only continue putting money into Google ads if there is a good return for them, ie more views or more money. So Google has material interest in returning results to advertisers, more so than they do on pointing me to a better chicken piccata recipe.
I think it's more complicated than that. Google's engineers are under pressure to constantly "improve" things, or in other words, to push out new ranking algorithms. Once you go down the AI path, it's probably hard to go back.
Then in 2016, John Giannandrea, an AI expert, took over. Since then, Google has increasingly relied on AI algorithms, which seem to work well enough for main-stream search queries. For highly specific search queries made by power users, however, these algorithms often fail to deliver useful results. My guess is that it is technically very difficult to adapt these new AI algorithms so that they also work well for that type of search queries.
While the old guard in Google's leadership had a genuine interest in developing a technically superior product, the current leaders are primarily concerned with making money. A well-functioning ranking algorithm is only one small part of the whole. As long as the search engine works well enough for the (money-making) main-stream searches, no one in Google's leadership perceives a problem.
Naturally, this would be a good time for a competitor to capture market share. Problem is, the infrastructure behind a search engine like Google is gigantic. A competitor would first have to cover all of the basic features that Google users are used to before they would be able to compete on better ranking algorithms.