Ask HN: Is Google Slowly Dying?
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?
It started a long time ago.
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.
I put the inflection point on Marissa Mayer moving on from VP of Search Products and User Experience to Maps (2010).
No one has been paying attention to the "experience" of the people using the overall Google suite/collection of products since.
Whatever you think about Marissa Meyer (particularly how much attribution she deserves for Google’s success) certainly things got worse after her departure and the same is true for Matt Cutts.
>Where do you see this going?
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.
The last few months shows they have no moat. I expect a slow decline lasting for years. Most people are slow to change but many techies have already switched over 50%+ to ChatGPT. Even if Bard eventually matches it, where's the moat? MS/OpenAI are willing to have lower margins because they're starting from zero, but Google will have to monetize hard, making their product worse.
There is no margin for any company for providing solution to coding snippets. Search engines earn money through travel, shopping and digital products. All of them are highly real time, visaul with deep integration to supply chain. Hallucinating LLMs with only text are years behind providing any viable alternative
ChatGPT has plugins in closed beta that can fetch real-time information from other sites and services. Things are moving so quickly that even though this appeared to be a problem a week ago, it’s pretty much solved now.
And in the months Google has had to respond, they have released Bard which is very noticeably bad compared to the competition. If this is their response then Google is in deep trouble.
In that case llm are summarisation tool which exist in every major search engine. I dont see any use case where the search query for black dress or ticket to paris is going to change. Those require fundamental change in either computer vision or 3d vr. Listing the result in text is not a great experience. Those were the queries where search engine earns money. No one earns money in explaining quantum mechanics in rap
I don't think summarization covers it. It potentially replaces the whole research process. Let's say you need a new TV and you're not very knowledgeable, so you ask "I need a new TV, what should I get?", and it could inquire "Do you play fast video games? Then you need higher refresh rates and lower latency. Do you already have a Chrome cast or so you want a great smart tv?" etc etc. Once the conversation has progressed to the point where the requirements are set, this is the natural place to insert recommendations. Why ever go to Google?
I don't think that this company will die in a predictible future.
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.
Apple and Microsoft are bigger than Google and both have managed to show a lot of innovation in recent years.
Google's problem isn't size. It's their poor product management.
Microsoft have not shown innovation. They have bought innovative companies at regular interval including github and openAI which is op's point. Otherwise they have blatantly copied alternative and forced them to enterprise companies like Team. Only innovative product from Microsoft itself in last 20 years is VSCode. Contributed significantly less to deep learning and distributed system research compared to google, meta or amazon
Microsoft is a very talented copycat. With the added expertise (that all the other FAANG companies are sorely lacking) of how to sell to enterprise and small business users and lock them into their portfolio.
I wonder if Microsoft is a talented imitator, but I think it can be said that they have a strong base mainly for customer companies, so they can easily spread their products by bundling them even if the quality of their products is worse than their competitors.
VSC is a great product but I think it’s a bit of a stretch to call in innovative when text editors and IDEs is a saturated market and have been around for literally decades.
Language Servers could be seen as innovative I guess but they weren’t the first to attempt this sort of thing. It’s just other solutions didn’t have the weight of Microsoft to gain wider adoption.
None of this is a criticism of VSCode though. Something can be really good without being an innovation.
VSCode might be the best thing since sliced bread. But for Microsoft's bottom line, it is utterly irrelevant. At best, VSC is a marketing win with a subset of the programmer population, but it doesn't even do anything for the adoption of "real" Visual Studio licenses or anything. VSC is more of an accidential side project gone right.
People forget how bad most Windows laptops were before the Surface. Decent-looking basic and premium Windows laptops are common enough now, and Microsoft is falling behind in every way, but there was a time where everyone was paying attention to them the way people do Apple.
I'd say TypeScript was pretty innovative.
I dont think that is a product. Google have created dar more important products including kubernetes, go, flutter as well many ML architecture including transformer
2009: Go
2014: Kubernetes
2015: TensorFlow
2017: Flutter, Transformer
Those are some big gaps. It looks like Google is in the position Microsoft was in for so long right down to being dependent on a cash cow (their ad network) and lock-in (Workspace). They've rested on those laurels for so long they forgot how to innovate. The difference is Microsoft realized they were on the path to irrelevance and started working on it years ago. Google just realized there's a threat to their dominance, and that threat is moving faster than anything that challenged Microsoft. LLMs are making leaps measured in weeks and months.
Google Tables (2020) could be amazing, but who's going to risk it with their reputation for shuttering new things once the people responsible have been promoted? Think about your list: it's all open source. Nobody's trying new Google stuff unless they know it can be forked.
> Google Tables (2020) could be amazing
Google Tables was an experiment, but…
> but who’s going to risk it with their reputation for shuttering new things once the people responsible have been promoted?
…its been declared a success that will be moved to a GCP product. How often has GCP cancelled products?
I would have expected it to go to Workspace. It seems to fit in more with office suite database applications. But you're right, the GCP timeline on Wikipedia shows no shuttered products. Maybe I'll actually use that CRM template.
How much money are those things making Google compared to Search?
Exactly. And how much money typescript is making for MS. This are not products
While innovation is often good, it's a side issue.
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.
In 2022, GCP revenue amounted to $26.28 billion, accounting for 9.2 percent of Google's total revenues.
They’re here to stay - but they need a cultural shake up
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)
Google is too big to die anytime soon, but they're already dead to me. The company that seemed so cool in 2004 is long gone. They're just another Oracle or IBM or Microsoft now.
The "pivot to try to save the business" usually doesn't work.
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.
It's a bit like evolution, where species don't adapt to a new environment because their old one is killing them (in this case they just die out), they adapt because the new environment offers them a more advantageous lifestyle, and those members closer to it prosper and supplant the old guard.
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).
Nokia is not dead, but they are not an important player anymore. Not even a skeleton crew company.
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.
Maybe they won’t die, but they are losing blood fast. I feel like even if they don’t die, they will only be an obsolete corporation that managed to somehow survive like yahoo.
But I’m pretty sure they won’t manage to remain among the top 10 in the next 5-10 years.
there are few other corps with cash piles who compete for innovative startups, namely MS invested in OpenAI
Google Cloud is a solid product. It’s extremely unlikely to be bigger than Azure and AWS in the next 10 years but being the third biggest American cloud provider doesn’t sound bad at all.
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.
Saying a particular car is the "worst" when it's regarded as the most reliable brand out there, is a weird take. It's like saying a navy blue suit is tacky. Just goes against the grain.
The Toyota BZ4X is truly the worst car you can buy in its category. It’s so disappointing that it’s entertaining.
And about the reliability, this car started well with its wheels falling off.
> YouTube is the biggest on the long videos segment and has no real competitor.
Exactly. YouTube's competition is not even close to challenging YouTube and is virtually non-existent.
Google has been increasing revenue by $20-30 billion per year, with a huge bump of $70 billion in 2021. They made half a trillion dollars in the last two years!
Oh, you meant the search engine. Yeah, maybe.
Google is basically a digital Yellow Pages at this point. Sure they have been producing a lot of high quality research and concepts but so did Xerox PARC.
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.
Perplexity.ai, not .io
RIMs revenue was highest 3 years after the iPhone came out and the year Android came out.
Revenues are a trailing indicator and can hide underlying issues.
Google's annual today is 11x RIM at it's highest. That's a lot of billions, I don't care what underlying issues you have. Even if Google falls to what RIM is today, Google would still be making tens of billions.
Past performance means nothing.
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.
Past performance does mean that you have a larger bar to cross before you can claim the company is dying.
And yes OpenAI is the trend right now but the jury is still out whether it will augment search or replace it.
IBM is still “alive” in a walking dead zombie type of way.
IBM has more employees than Google. It might only make a quarter of Google's revenue, but that's still a lot of money. That's a big zombie.
> Especially because Google has exactly one revenue source.
Google is an advertising company that just happens to have a lot of engineers on its payroll.
They have posted record drops in profit in the last two or three Qs.
Also, remember Nokia and how fast they disappeared.
Google is utterly massive, and companies that get that big tend to never fully die. Like large countries and other entrenched institutions, they can decline, but it's a long way down to zero. As such, Google will likely exist in some form for the rest of our lives, just like IBM and GE will.
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.
I don't know about Google is dying but Google Search has lost me. Remember when Bing! first came out (IIRC it had the exclamation point and the advert line was something like, "Don't say you Binged it, say you Banged! it!" -- or something equally cringey only an adman could love) ... the results were not great, and someone (probably on Reddit) said that Bing was only great for porn searches. And I was aware of all kinds of optimizations went on at MSFT to make Bing competitive with Google Search and I remember thinking to myself, they will never catch up, it's too late, Google Search was light years ahead. And I think Bing was the replacement for MSN Search which was even worse. I didn't use Bing for the longest time, but as Google Search got worse and worse and everything I search for now leads to something to buy instead of information, I went to Bing to see if it's just as trashy. Lo and behold! There's a beautiful photo of something interesting in the world every day, and I love it. It's a Google Doodle tactic I guess, but they got me. All those engineering hours of algorithmic grinding and really all they needed to get me was some happy picture of the world and animals. Apparently you can win the Internet with cat pics. Steve Ballmer must be kicking himself.
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.
I feel a bit out of the loop here. Google's search engine results have only ever been getting progressively better over time from my perspective......?
Really? I’ve been unable to find quality results for quite a few years now. It’s always results from Quora, Reddit, Facebook, YouTube and other giants.
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.
Unfortunately that is the internet now. The results and pages you are getting are poor because people stopped publishing websites. Now most of the internet forums and self curated pages are gone and information is in those silos
That’s.. so sad. I think you are right.
As much as I don’t want to accept this, it seems like it’s the new reality.
However, isn't this also the consequence of various SEO tricks? A website such as YouTube can do SEO way more efficiently than someone who produces their content manually. If you have billion pages of content, you can easily make billion pages link at whatever you want... perhaps the things where the users most often click the ads, or whatever.
Basically:
Google: "good pages are those that have lots of links pointing towards them".
YouTube: "let's make all our pages link to the 'why quantum mechanics proves Trump wrong' viral video, people seem unable to resist clicking on it when they see it among the search results".
Random guy on internet: "I wonder what quantum mechanics is... let's ask Google..."
Since you're new..here's a tip Do one pushup every time there's a "How I degoogled my life" post on the front page Do one pushup every time there's a "Google is too big and should be broken up" post on the front page Do one pushup every time there's a "Google is dying" post on the front page
You'll be jacked in no time
I don't know, I disagree, that hasn't been my experience.
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.
I wish I could tell it I always want English language wikipedia results first regardless of which language I search in. I'm tired of my top result being a poorly translated version of the English wikipedia page I actually want.
Yep, that's also very frustrating. And if I wanted results in Japanese I'd type in Japanese, so it's not particularly hard for them to guess intent here.
It really depends on the domain. The biggest problem I've noticed is they don't seem able to filter out those shitty auto-generated spam sites for comparisons or top 10 lists. So if you're searching for "X vs Y" or "best X in 2023" you'll get absolute dross.
For other kinds of queries it still works pretty well IMO.
> The biggest problem I've noticed is they don't seem able to filter out those shitty auto-generated spam sites for comparisons or top 10 lists
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.
Google is already showing big brands in most cases: https://detailed.com/google-control/
Yahoo Directories was this exact idea, and it worked well, until Google's PageRank blew it away.
You still gotta do actual information analytics on the data.
That does not match my experience or the general zeitgeist here on HN. IMO search quality has been declining due to increased invasiveness of ads and the wrong incentives for the company from the perspective of users.
Welcome to HN. This is the one undisputed fact that 99% of us can agree on. Arrogant Google hasn't done anything of note for decades and search in particular is a dying shell of itself. I remember the good old days when it could find everything for me and my wife still loved me. Now I get a bunch of Pinterest links on the front page
I had to check your profile based on this comment -- looks like you're new and the HN hivemind hasn't assimilated you yet. Don't worry, you'll be in the loop soon enough if you stick around. However, be mindful that echo chambers are usually intolerant of facts (and sometimes, surprisingly, knowledge in general...this place can be a good jumping off point if you know how to do your own research, though).
I need to add reddit to all queries to get good results.
I don't think GPT is a replacement for Google search, but I do believe they need to update their ranking algorithm. Ever since day 1 people have been trying to game that algorithm and figure out how ranking works to be able to put their sites on top. This is not going to stop, in fact it's only going to be accelerated by GPT[0].
Google claims your website and content should be user-focused, and not SEO focused. But then when you make a page that's user-focused, you get ranked bellow 10 other crap pages who don't actually offer much of the info you were looking for, but they show up ahead because their owners gamed the ranking algorithm. This is in my opinion the most important thing they need to fix right now.
[0] https://seo.ai/
Bear in mind that Google doesn't make money from users who are reasonably tech savvy, have adblockers running and are also of the mindset that they would never click on an ad anyway.
Folks have been complaining about search quality for at least the past 15 years btw.
Yeah, seems like most people only use the internet to find funny pictures of cats, but does that necessitate a dumbing down of all search results?
How hard would it be for Google to better tailor it's search results based on the complexity of a user's previous searches?
The internet is already bifurcated, because humanity is bifurcated. Roll with it.
I'm already rolling with it, I don't bother searching for answers to complex problems anymore because it's a waste of time.
Many pay for gsuite and YouTube premium
Yes, you are correct but I doubt Google would be such a huge enterprise relying on these subscription-based products where there is already a lot of competition.
I feel like Google’s search result quality has been steadily declining
With ChatGPT and other AI tools being introduced
Could it be they are de-tuning search intentionally so that people are more likely to be accepting of it's replacement? And I agree with you that search result quality has been dropping, at least for my searches. I can almost instantly find any comment on HN using Google however results for technical articles seem to be harder to find or require digging dozens of search pages into Google.
No, they aren't doing that. It's just a result from the revenue source that they have which is ads.
Little by little, like by natural selection or AI training, the search got worse because revenue increased.
That makes sense especially since search has been declining for a while now.
Google search gives me links, which is usually what I want. It's pretty easy to ignore the ads and promoted content.
Google maps is quite another thing. Nearly half the screen is filled with "helpful" information that is of no interest to me. Sure, I can make some of that go away on a given view but if I zoom in from a region to a city, that side panel reappears. Making things worse, the side panel does not change size if I narrow the window. At the size I normally have for a web browser, the side panel is half the screen. Perhaps it's a personality defect, but this really annoys me and if that side panel reopens after I've closed it, I normally just switch to apple maps or openstreet maps, although I might go back to the google product if it has a more recent photograph of something I want to see.
Yesterday I had the worst googling experience of my life - I wanted to find some info about some seeds, but google was only pushing me webstores that sold the seeds, and completely ignored the language I was searching with and gave shops based on the country I was in. It was so infuriating!
It’s hard to predict. Companies with good DNA can bounce back. Take a look at Microsoft. They were written off in Google’s hay days.
Google has a good technology DNA. I think they need an inspiring leader, maybe another insider who is a strong engineer and a leader.
There’s a big difference between Microsoft and Google though - diversification. Microsoft have been truly diversified in their portfolio. When they release new products they mean it, they iterate, they put wood behind the arrow. Google’s sole interest is their ad business. If a product doesn’t increase ad revenue, it’s just a PR stunt to keep the brand relevant, it will be quietly sunset when nobody is looking. This will be their Achilles heel.
We still need search engines to crawl the web and discover new content.
If you rely on an LLM as an all-knowing Oracle then you're dependent on what data it was trained, how often it gets updated, and whether the web content you were hoping to find, or some close approximation to it, is statistically regenerated by the LLM or not.
I think LLM-augmented search such as Bing or perplexity.ai is the way forward.
Perplexity.ai is interesting as it's very capable, and despite being based on one of the GPT-N models seems to generate output quite a bit different to Bing/chat.
In a sense traditional search is just a really, really dumb LLM. They both injest a bunch of text and build an "index" based on the text.
The traditional search model spits out links and short summaries, based on crude-ish matching of the search terms, whereas the LLM spits out actual answers (derived from the content corpus) based on sophisticated understanding of the query (but also, currently a big caveat, can also just lie to you!).
But everything the LLM needs to do traditional search is already in there. The main problem to be solved is just getting it to lie less; another problem is getting very frequent content updates, like search indexes do.
Even so, LLMs are already better than traditional search engines for many queries, as search engines qua search engines.
https://imgur.com/a/12jcyKT
> In a sense traditional search is just a really, really dumb LLM. They both injest a bunch of text and build an "index" based on the text.
Not really .. the significant difference is that a search engine is backed by a web crawler that is running continuously discovering new content, vs the LLM which has a fixed training set that will only be updated infrequently (very slow and expensive to retrain). Also, once a new web page, or new version of a web page, has been indexed by the crawler then you should be able to find it if your search terms match, while with an LLM all bets are off as to whether you can coax it to generate something based on something in the training set.
Totally agree getting fast updates of new content is a problem that still needs to be solved. In terms of finding matches, I think LLMs are already better than Google, for content that is there.
(A super-traditional search that's just looking for raw text, like the HN search, could claim an advantage there, but something like Google or Bing is already far far off in the land of guessing intention rather than raw text matching, in my experience; in any case, 99% of the time raw-text matching is not what you want, anyway.)
Maybe time will prove me wrong, but my prediction is LLMs will improve on their weak-points and traditional search will mostly fade away in relevence.
> https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
The question comes up every year and the oldest one is almost a decade old. I admit we are at the juncture of a very interesting phase in the software world, and the threat is more tangible. But having seen enough hype cycles, I want to reserve my skepticism. Google is pretty fast for most of my queries. If want to check recipe for a Cheese Cake, I would much prefer to search "cheese cake" and check a link or a video vs ask ChatGPT about it.
Another reason for my skepticism is that sometimes projections can get overly-optimistic. Uber was supposed to be only a few years away from a fully autonomous fleet of cars. There obviously many problems that we are ignoring when it comes to adopting ChatGPT on a global scale (like cost of a query and wrong information). As an example, I asked ChatGPT, "how can i convert user agent string to browser and version in big query" and it told me that I can use "PARSE_USER_AGENT", a function that doesn't exists. I also have issues with how opinionated ChatGPT is and refuses to answer certain questions based on what it perceives as fair.
I don't know how future will pan out, but it isn't as crystal clear as people might think.
Google Search is far more popular than the next runner up.
Imagine a front-runner in a marathon race who doesn't care about setting records or beating a personal best, but only winning the prize of that race. He or she looks over her shoulder and can barely see the next runner in the distance. That front-runner barely has motivation to maintain pace, let alone surge ahead.
It will take a runner-up who starts to catch up to Google in popularity and seriously eat into their search business for them to do anything.
Bard seems to be pretty impressive as well and it seems to be a mix between a GPT and a classic search engine, queries seem to be able to handle information about both current and future events. It's naive to believe a giant like Alphabet with $100B in the bank won't adapt, perhaps they will incorporate Bard into Google Search or keep them separate services with part of Bard powering Google, YouTube, Gmail, Drive and other platforms.
I think search-wise Google is an interesting place. The results page (especially mobile) has transformed dramatically the last few years. Now you get a dizzying array of widgets filled with rich info that’s been surfaced from other places.
Personally I find the way it’s organized haphazard and frustrating to navigate. However the info they’re a managing to put front and center before you click a link is impressive. I think there’s and opportunity for Google to harness this power and create the best possible power search interface.
I also don’t believe chat is the instant death of search. I think many would prefer a killer dedicated search interface over chatting with their computer.
Side note: as someone who loves a good old list of links, I’ve been using kagi.com and have been happy. Big fan of paid search as a model.
I also really like Kagi, but I think the new pricing is on the expensive side for someone that does a lot of quick searches.
It depends. I see a lot of folks ready to immediately accept whatever BS comes out of ChatGPT. ChatGPT could hallucinate an entire Wikipedia clone and some people wouldn’t question it. I see this trend accelerating, especially among people who are Google Ad’s target demographic.
That’s a big risk of ChatGPT.
Another huge risk is that they could change the perception of what is right and wrong in the long term if GPT gets accepted.
Data for the last four months suggests Google search market share share is up, whereas Bing search and proxies are down [0]
We have seen noticeable and increased growth in usage of our "classic" search engine Mojeek in the same period.
[0] https://twitter.com/ColinHayhurst/status/1636720458564349952
I didn't know Statcounter was still around.
https://gs.statcounter.com/faq#methodology
5 billion isn't a small number of impressions, but I would be interested in comparisons with other sources based on more mainstream trackers.
I can resonate this. I used to be a huge Google Clown Platform fan but boy these folks displays a sense of condescending attitude to their users/customers.
Long story short. They don't give a shit of you and have zero empathy.
When the money from ads stop, it will be inevitable.
Context: https://news.ycombinator.com/item?id=35133917
Google will become good again. With TikTok threatening Youtube and ChatGPT threatening search, long-term success decisions become relevant again.
Nowadays, I think Google search is just a big brand like Coke, Nike, etc. Google search is relatively new and should be time tested this decade.
I remember when Google search worked, I remember when Google Adwords attracted customers. All memories. It was better than Altavista and Yahoo Ads. We need a new leap. Don't care if it is about AI, GAI, or next algorithm(s). Startups were lazy in this space.
They already died imho.
Lets admit it - previous research loop was going for a google search then iterating on results, which takes time and effort. LLM based research (although could be misleading) is just much faster and more suitable for natural language.
It happened.
This doesn't mean that Google will be obsolete, if it leverages its collected data so far it will be a strong competitor. But Google isn't leading at the moment.
This is a very limited view of the things people use search for. You still need links and a way to browse through the results. And for many queries you need factual up to date information. LLMs don’t do “local supermarket opening times” and “coffee near me” - or the ones that do run a search query anyway and give you back the results.
Except Google invented transformers, which is what all the major LLMs use.
If Google was dead, this stuff would not exist right now
For Google, it has had no trouble developing the technology necessary to be a leader. While it sometimes does not take advantage of the result as well as others, the ability to invent that stuff is critical. It's much easier to become good at product management than good at inventing the future. You need to do both, but I'd bet on a company good at the latter being able to become good at the former well before I'd bet on a company good at product management that can't innovate anything.
For starters, the ones good at inventing the future have less competitors in the future they create. You have some time to try to build the right product. The ones good at products often are competing against hundreds of other companies with the same good idea
The purpose of a company is not to invent technology. It’s to invent products that can make a profit. Google is very bad at producing profitable products outside of adTech.
Interesting and surprising comments. Anyone ever heard of AOL? Are they dead?
I think Google will likely go through similar path as Microsoft; such a tech powerhouse and cash cow for so long, innovation became eclipsed by profit optimization.
Microsoft turned it around a decade ago with a successful cloud pivot.
Google would have to fail fabulously for minimum of a decade before we could even begin to start claiming it’s terminal.
> Anyone ever heard of AOL? Are they dead?
I think that when people say "dead" in this context, they aren't really meaning "out of business". They're meaning "no longer relevant".
AOL, as well as IBM, are both still doing pretty good business, but in terms of relevancy (at least culturally), they are both pretty dead. Neither can move markets. They are shadows of their former selves.
Microsoft “turned it around” by extracting money out of its existing enterprise customer base. Apple had a base of fanatic customers even during its lowest point.
What does Google have?
How many times have you used google or gmail or YouTube or android today?
If Gmail is just email and I use it on my iPhone with the native mail application -- along with 5 other email accounts. If Gmail disappeared tomorrow, I would choose another email provider. There are very few "fanatics" for Android. It's the only alternative if you don't want to spend on an expensive phone. But the majority of people who can afford a more expensive phone -- buy an iPhone.
I hope you realize that Google isn't just a search engine.
They have YouTube, Chrome, Chrome Web Store, Android, Maps, Google Docs, Drive, Gmail, GCP, Firebase, etc etc and some how they are "dying".
The death of Alphabet has been greatly exaggerated on HN and in the markets. The fact is, it would take more than just ChatGPT to "kill" them.
If you said Yahoo instead, then I would agree.
If they don’t label generated text as such, the value of the search results will be compromised. All non-ads will be assumed to be generated without inspection, making ads the most “reliable” results. This should hasten shifts in the user base to non generated alternatives.
Has been dead to me for years. Only youtube is irreplaceable, but enshittening is at full swing.
I saw that word ("enshittification") in Cory Doctorow's writing:
https://doctorow.medium.com/mass-tech-worker-layoffs-and-the...
...which also had this amazing summary of what we're seeing now:
"The Big Tech firms are locked in a race to see who can eat their seed corn the fastest."
I use the subscriber search engine Kagi, and it is now better than Google imo
how is that there is a lot of fuss about bad search results, but there is no screenshots/screencasts that show the query and the result ? or the overly exaggerated comments that there is only spam in the results and nothing more, which is hardly believable.
for me the search result is getting better, sometimes im looking for generic words via right click "search google for <word>" and the results are really contextual, like google is considering the page that i'm currently looking at, wichi sometimes surprises me a lot.
Both Google and Bing's results quality, in my opinion, have declined. What they both were is now Yandex. I've started using Yandex for searches more frequently.
I already use Ecosia for most of my searches because Google just forces so much useless knowledge cards and buries search results.
Only use Google when I find no good results.
IBM has been around for 111 years. It’s been on the downtrend since the late 70s early 80s. A big company like that can take a long time to disappear
When I google I always add the word Reddit / stack overflow / the website I am likely to find what I am looking for
GOOG is an ad company and all their products are there to hoover up user data to serve users hyper-personal ads.
Have you ever seen the HBO show "Succession"? It's about a fake Rupert Murdoch character and a fake NewsCorp company and one of the plots is they get in trouble for big misconduct around abuse of workers and all sorts of bad stuff but they come up with the PR line "We're Listening." And do this big PR campaign to try and show they "get it" and they are hearing you. BUT at the same time one of their products is like an Alexa is actually listening to people's conversations creepily! Google is getting a reputation for being creepy and they _hate_ this.
[dead]