I feel like Google’s search result quality has been steadily declining since at least five years, and from what I’m reading both on HN and on other platforms, people feel the same.
Add to that Alphabet’s notorious love for constantly killing products, and I’m wondering how long they will be able to keep this going?
With ChatGPT and other AI tools being introduced, will the need and reliance on classic search engines drop quickly or will it take years?
Where do you see this going?
When Google first introduced ads they were in those distinctive little boxes and not blended with the content. They said at the time that blending the ads with the content would destroy the legitimacy of the product and that they didn’t want to be evil.
Trouble is the ads compete with the search results. Why would a user even look at the ads if the search results were good? Why would an advertiser buy ads if they could spend resources in SEO and get an organic search spot that isn’t labeled as a self-serving ad?
SEO is a Janus-faced God. It has a negative aspect of spam and dark patterns (remember those weird pages you would hit on a typo domain? They looked like they were drawn with a crayon but were well-oiled monetization machines.). On the other hand, quality content that people want to link to and that fills gaps that people are searching for is powerful too. (Particularly when joined with darker tactics.)
You might blame Microsoft’s purchase of Powerset to get the knowledge graph technology behind Bing, Google’s panic and subsequent purchase of Freebase. But certainly it was the blending of ad content with the search results that made the organic search results wither away.
No one has been paying attention to the "experience" of the people using the overall Google suite/collection of products since.
As bad as Google search has gotten, for normal users, it's been much better than than the next best thing for a long time. This made Google complacent.
In the past, I thought that Google's main source of revenue, ads in search results, would be at risk because people were searching on places other than Google.com. Like using Siri on their iPhones, or asking their Alexa device something they might have used to use Google for, or searching Amazon.com for product oriented searches rather than Google.
But, Siri, Alexa, Amazon, etc, still fail at things that Google's search can do well. So what I expected to happen didn't pan out.
I do think that all the hoopla around ChatGPT/GPT-4/etc is a little exaggerated. But it does show that entities other than Google can, just recently, leapfrog them for some tasks. Like around taking a request, figuring out intent, and giving a reasonably good targeted answer.
So, perhaps where this is going is that the maturity of that kind of tech finally allows for competitors to take marketshare from Google. Since the starting point isn't Google having a huge lead.
They are too large to be innovative. Internally they are paralyzed by processes, management, and burned-out programmers who had ideals a few years ago and now are tired of the internal system.
But Google is a rich company that can buy innovative startups and acquires innovative technology whenever they want.
I think it will be key for their survival.
Google's problem isn't size. It's their poor product management.
Google's problem was deciding to go all-in on advertising as essentially the only revenue source for the entire company. Worse, as others have noted, this revenue source has the side effect of cheapening and distorting their main claim to fame (i.e., search).
Apple and Microsoft both have numerous revenue-generating products. Google basically just has ads (all the other stuff they have is a rounding error).
Now, Google has had other products that could have become revenue streams, but they have a terrible history of killing those products off. They've overtaken Oracle as the place where good products go to die.
If they don't get the advertising people on a leash, I don't see Google winding up anywhere good over the next decade or so.
Sundar might end up being looked back on as the Ballmer of Google - the leader who wouldn’t take risks in order to protect the golden goose (geese in the case of MSFT)
Where pivots work is something like Amazon and AWS where internally Amazon already did IT at a large scale incredibly cheaply, so they figured out how to sell it.
They didn't do that though because retail was collapsing, they just saw another opportunity.
Large corporations are similar, both in that they don't respond like deliberate individuals anymore but rather a reactive swarm, and in that they typically only pivot successfully from a position of strength/opportunity ... when facing adversity they more often fail in their pivot and die (usually get acquired).
When the iPhone was announced they were rich and in a powerful position, strong teams, brand loyalty, sales teams all over the world, very good relationship with telcos...
And yet, it was already too late.
The HN crowd is a good predictor for search popularity.
But I’m pretty sure they won’t manage to remain among the top 10 in the next 5-10 years.
YouTube is the biggest on the long videos segment and has no real competitor.
Google Maps is a great product too without big competition for end users.
Android is the main (and only?) alternative to iOS.
Then they have many descent products like Google Photos or Google Sheets. A lot of strong competition on this domain but again not being the biggest isn’t the end of the world.
Google Search will probably still keep plenty of users. Brand loyalty is strong. For example, a thousand of Norwegians bought the worst electric car of the market this year because it’s a Toyota. I’m sure such people will also happily keep using Google Search even when it’s only showing them garbage content full of ads.
And about the reliability, this car started well with its wheels falling off.
Exactly. YouTube's competition is not even close to challenging YouTube and is virtually non-existent.
Oh, you meant the search engine. Yeah, maybe.
Much of the profit margin Google makes now is from people clicking on ads that they do not know are ads. A good portion of that is from companies paying for clicks from searches for their own company or product names. Neither of these are really sustainable or models that produce consumer surplus. It's a form of navigational rent seeking.
There are a lot of claims that people only want to use Google and won't stop, yet Google is paying Apple tens of billions of dollars to remain the default search engine on the iPhone. There is a reason why. There is also a massive weakness here because:
#1 Regulators in the US and the EU may break this #2 The more Google's results are just Yellow Pages, the easier it is for Apple to swap out the results and no one notices.
I've extrapolated on this general argument for the last few years, and this was before LLMs. The decline of Google search quality has been discussed endlessly for years on HN. Now we have search engines which are suddenly a lot more capable than Google for finding answers to more complicated questions (perplexity.io etc.)
We can talk about how Google actually isn't far behind in AI or may actually be ahead. The problem is, besides increased cost of search query, hardware demand for AI/ML training and inference is going to be really tight for a while. TSMC is going to be deciding who ends up with this stuff. A LLM providing answers to complex questions with ads plastered around it and in the answer may not be the highest value thing to do with advanced semiconductor compute, even if Google did want to pay a premium over everyone else.
If you were investing in Google, the problem would be to identify how fast Google's search revenue is going to erode. People still pay for cable TV and watch it. People still use Yahoo as a search engine. You could bet that Google will have an abrupt pivot and suddenly be as well capital managed as Microsoft. They really have a poor track record at this point going back a decade. Apple had that pivot when Steve Jobs returned. I'm not sure Eric Schmidt is a Steve Jobs or that he would be returning if he was.
Revenues are a trailing indicator and can hide underlying issues.
Intel was also growing fast until it didn't. We all know why.
Especially because Google has exactly one revenue source. MSFT and other companies are at least somewhat diversified. But Google is incredible vulnerable right now.
And yes OpenAI is the trend right now but the jury is still out whether it will augment search or replace it.
Google is an advertising company that just happens to have a lot of engineers on its payroll.
Also, remember Nokia and how fast they disappeared.
Deleted Comment
Also keep in mind that ChatGPT in its current form hasn't been fully productized. Just like Google search was excellent when what it provided was in line with user needs and now only provides the absolute minimum value to get as much ad money as it can, so too are LLMs in that phase. There's nothing stopping LLM-driven services from returning the same kind of results based on whatever topic is being queried about. In fact, that's exactly what companies like MS want to do with it. None of the large companies pouring billions into this space have a goal of making society at large better informed. They want LLMs integrated into their services so they can deliver a new level of targeting for ads, reduce customer support costs, shape narratives, automate PR, and attract users to platforms that can then be marketed to their real customers (other businesses and governments mainly). Those of us envisioning idyllic futures that this tech will unlock need only look at every other technical advancement in recent history to see how it will end up.
And now, there's Bing with ChatGPT integration, and I mostly use Bing now. Results are still SEO'd to hell, so there's that. Remember the innocent old days when I (we) conscientiously clicked on some ad with deliberation, because the search results were good and we wanted to little guys to make some fraction of pennies that Google paid them? At one time there were a rash of search engine contenders, before Google there was AltaVista; a Google offshoot with an unpronounceable name, Cuil, Cuoil, was a thing, so was Yahoo! Search.
Google Search is not so much dying, as "Search" is being destroyed through ad monetization. I know (or believe) there are still pockets of the internet with useful information, but when I go to Google it seems that there's nothing except things to buy, it's a giant mall where everybody is but nobody wants to be.
In addition to that, I’ve been seeing legit scam websites rank in the first page since a year or so, and that’s not something I’ve come across only a few times.
You'll be jacked in no time
Several years ago, the first result for web topics was almost always MDN. These days it's usually one of those SEO programming content farms for beginners.
The other thing that annoys me to no end is that Google's own language settings don't work. I live in Japan, but I don't want Japanese programming search results. I have told Google to only return English language results, but it ignores that most of the time. The only way to force it to use English is that sometimes it asks if you want English results only, but you can't force it to do that to begin with.
For other kinds of queries it still works pretty well IMO.
That boggles my mind as well. Google already knows which media are actually relevant as a source. For example, if I want good reviews for electronics products, I'd go for Heise and Golem, for everything else Stiftung Warentest. Google could easily compile a list of reputable media for each country and prioritize these, but they don't - I suspect because they enjoy all that sweet sweet ad money coming in from blogspam sites and less-reputable media.
Deleted Comment