They can even add mind reading capabilities, until they aren't the default search engine on the biggest browsers and mobile platforms in the world(FF, Chrome, Android, Safari, iOS), their market share ain't growing significantly. Defaults matter most, and Windows-Edge isn't big enough to drive this.
> The “Google is going downhill” narrative only exists in the tech bubble. No one actually thinks this way in the real world
Maybe. But the opinion that Google is returning less relevant results is sooner or later going to go mainstream unless they start returning relevant results.
We're users too, but we notice things first. Just because the other users haven't noticed them doesn't mean that they won't in the near future.
I personally don’t even agree that Google returns irrelevant results.
I am a DuckDuckGo user and I still use the !G flag when the results I’m getting aren’t good. Every time I go over to Google the results improve slightly.
ChatGPT is a game changer when it comes to getting a system to understand the underlying meaning of my query. I used to give the example of searching for Frank Lloyd Wright’s height - doing so on a search engine would give you results about the buildings themselves. You’d have to dig for the simple information of how tall the dude was.
Alas, Google actually fixed this and quickly displays the man’s actual height now.
Microsoft and Google are both going to integrate AI into their search engines and they’ll both improve drastically as a result. We’re only scratching the surface on the subject.
It’s not so much that Google returns irrelevant results as it is that Google returns results that are optimized for taking as long as possible to glean the information you’re looking for.
Recipes are the classic example. Because Google upranks sites that people spend more time on after a click, most recipe searches yield pages that start with the history of flour milling and what milk is, before eventually getting to the desired recipe 5000 words later.
I don’t agree that Google is optimizing for time wasting, either. The faster they get you to click an ad link the faster they get paid.
The phenomenon with recipes that you mention has nothing to do with Google: it has more to do with copyright. You can’t copyright a recipe, but you can copyright written commentary about cooking a recipe.
The recipe blogs also need enough content to place ads around.
From my memory working on web indexing at Google roughly 15 years ago, there was a step function for detecting "short clicks" that indicated a page was irrelevant for that query. As long as the user spent above some minimum before returning to Google, increased time on the site didn't indicate higher quality.
So, the real optimum is ranking for one or more popular searches (regardless of relevance), looking vaguely interesting, and just confusing enough that it takes a couple of seconds to realize this isn't the page you wanted. Either that, or being exactly what someone wants for one of the popular searches.
Though, things may have changed, and my memory definitely has faded.
Also, plenty of SEOs may have misunderstood short clicks as a quality signal, so even if maximizing time on site doesn't help, I wouldn't be surprised if SEOs still recommend it.
> ChatGPT is a game changer when it comes to getting a system to understand the underlying meaning of my query. I used to give the example of searching for Frank Lloyd Wright’s height - doing so on a search engine would give you results about the buildings themselves. You’d have to dig for the simple information of how tall the dude was.
> Alas, Google actually fixed this and quickly displays the man’s actual height now.
That's really the only thing I want: A search engine that understands my queries. I don't want to chat with a computer.
Altavista et al got spammed into unusable mess and their use collapsed almost overnight when PageRank (temporarily as we've learned) managed to filter out the spam.
Of course nowadays Google's tentacles are so deeply everywhere that they will churn on with subpar technology just like e.g. Oracle, SAP, IBM, Microsoft etc etc. The oligopoly is probably here to stay.
It's so easy to try out another search engine, and people do all the time. It's also easy to set any search engine as the default, unless you use iOS. It's also super easy to use different search engines in parallel. I often search HN for a different perspective and for live search. Sometimes also github. And of course amazon, whenever I want to order something.
> PageRank (temporarily as we've learned) managed to filter out the spam.
It was more temporary (shorter time period) than you probably realize.
I was a developer on the Google web indexing system. PageRank was only one of many signals that went into ranking, and PageRank was gamed by link farms pretty early. There's a whole team dedicated to fighting webspam, and their spam signals are some of the other inputs to the scorer.
By Google's golden years, there was a whole lot of secret sauce going into fighting webspam, and PageRank had been pretty thoroughly gamed.
PageRank was important, but I think its importance is generally over-estimated.
Also, about 15 years ago, Google was testing another link-graph-based metric that was more efficient to calculate and less susceptible to linkfarms. I imagine PageRank (as described in the patent/Wikipedia) has been replaced, though perhaps the new algorithm is still referred to as PageRank. (IIRC, the proposed replacement was being called PageRankNG while in testing.)
> Maybe. But the opinion that Google is returning less relevant results is sooner or later going to go mainstream unless they start returning relevant results.
Less relevant compared to what?
Consumers don’t see this as Google going downhill, they see it as the internet going downhill.
And I agree. It’s not as though I can jump over to DuckDuckGo or Bing and my search quality radically increases. Usually the opposite.
ChatGPT is helpful because it gives a very convincing feeling of being right. Whether or not it’s actually right is another question.
Compared to what the user wanted when they ran the search in the first place.
> It’s not as though I can jump over to DuckDuckGo or Bing and my search quality radically increases. Usually the opposite.
Of course it's a personal anecdote but as a DDG user since 2014 I routinely add !g to my search when the results don't satisfy me (which happen every once in a while, say once a month). This used to help, as most of the time Google had the answer I was looking for and could not find otherwise, but I don't think it happened even once over the past two years: Google is now at least as bad as DDG when it used to be strongly superior.
> And I agree. It’s not as though I can jump over to DuckDuckGo or Bing and my search quality radically increases. Usually the opposite.
I acknowledge that it is a paid service, but you haven't tried Kagi yet you should give it a go! Default Search quality is very good, with no ads, tracking, or telemetry, and you can use Lenses and customizations to get even better results. You can even block certain sites from appearing in search (or simply de-rank them), which has dramatically reduced Pinterest and other SEO spam for me.
> Consumers don’t see this as Google going downhill, they see it as the internet going downhill.
And I think they're right. More and more information on the internet is siloed into apps or behind paywalls. Knowledgeable people who used to write technical articles now make youtube videos (far less searchable). And the good stuff on the internet is diluted with low quality content written by someone who has to churn out an article every 15 minutes that is hard to filter out, even as a human. (I wasted 3 hours yesterday because a website gave me the wrong formula for something)
It doesn't matter who agrees Google is delivering less relevant results until someone is delivering significantly more relevant results. As of now, that does not yet exist. It should only be a matter of time, but the only safe assumption is that smart, >$1T companies don't give up on their core businesses too easily.
It’s going to be a freaking cliff when it happens - the moment Tim Apple thinks that Google as a default isn’t good enough for the iPhone he’ll have it switch to some LLM powered by DuckDuckBing or something, and Google’s search traffic will plunge.
The decline of a monopoly is like going bankrupt. It happens slowly first. Then all at once.
Their position is entrenched enough that they'd be able to exist just on brand fumes for a good number of years yet but once the rot sets in there's no coming back.
Also, average people’s tech habits don’t change overnight.
Tech people will change browsers, search engines, and SaaS tools without much thought. The average person learns something and sticks with it for a very long time. Bing is playing a long game.
There are a lot of examples of things people in tech bubbles think that have little relevance in the real world: That Google can’t be trusted, that everyone wants laptops with upgradable RAM and phones with swappable batteries, that desktop websites should take priority over mobile, that offline apps are preferred over cloud apps. Spend some time in product management and you realize that tech people’s demands are unique.
Even the claimed demands of tech bubble people often fail to materialize into sales. Just look at all the demand for mini iPhones in tech bubbles compared to how little they actually sold.
This used to be the case when using a computer before the mobile walled gardens took hold.
Non tech savvy people don't even realize they are using Chrome in android or Safari on an iphone. They will never change they just use the default search bar that is on their device.
I used DDG as my standard but I find that I still need to go to google for all the conveniences that they have built into their search functionality. I have to give them credit they have figured out somethings to keep you binded to their search.
While I don't think it's enough to drive people to a different product, the fact that Google search is noticeably worse than in prior years has definitely gained mainstream attention. The amount of ads and SEO junk has caused frustration, but the alternatives aren't well known for most people to bother to switch.
Honestly my fairly tech illiterate mother is constantly complaining about how google is full of ads and irrelevant information. I'm not sure that people don't notice. It's just they know there's nothing better.
I believe there's a sweet spot before companies grow too much where there can still be some ethics behind some business decisions, after that ethics become a byproduct influenced by marketing strategies and public sentiment weighted against market competition and many other factors, basically it transforms into an entity that tries to make as much profit as possible at all costs.
Since they’ve added it, my searches in the mobile app fail about 50 percent of the time because the connection isn’t good enough. I don’t know what they’ve changed but it has some weird connectivity requirements. I only use it for the rewards and now Google is going to match it(at least I got a postcard with an online offer)
Too late, they blew their one shot. Ain't no one who tried it when it was new and hyped and was told that they couldn't use it since they weren't using Edge setting reminders to go back occasionally and check whether Microsoft has finally pulled their head out of their ass, it just instantly ceased to exist for them.
I tried it a few times but was mostly disappointed with the results. I built a superior solution (for my usage anyway) in like 50 lines of Python.
Just uses duckduckgo for search, extracts text from top 8 results, shoves into gpt-3.
(I think I got a massive cost reduction by filtering paragraphs with embedding API, but that's just an optimization.)
Phind is better than what I built, though. I don't actually use it that much (I'm not sure why—habit? Impatience?), but every time I do I'm surprised with the accuracy.
I think a big part of it is which pages to select for sources. Google Bard failed hard here in my testing, gave nonsense answers because its sources were AI generated SEO spam!
Yup, I too use Phind sometimes, and it's amazing. Plus if I want to chat with approximately any model known to man, there's chat.lmsys.org and labs.perplexity.ai. Microsoft thoroughly shit the bed by taking a narrow opportunity to get people to move to Bing as an opportunity to instead go "hey, let's blow this on trying to force people to use Edge!"
I hadn't had much luck with any of these. Phind is indeed better for programming but it still has the same trouble with the obscure stuff.
It seems that the main problem right now is that they only look at the first few results. Basically they all do this:
1. Ask model for a search query.
2. Run search and feed the results into the model.
3. Ask model for an answer.
The problem is that for very specific questions the search results won't contain the required information. So it has to say something irrelevant or hallucinate. Especially for negative queries like "a foo library that doesn't depend on std" or similar it really struggles because it can't effectively filter out libraries that fail the requirements and can't keep searching until it finds one.
Basically they can be fast if a decent query can have the answer within the first page of results. But fail otherwise. In many cases you are better off just reading through search results on your own.
Market share should be relative, so if there was a stark decline in search bing would have declined close to as much as Google. Unless this is some other measure erroneously labeled as market share
.
Sure
I've just seen 'market' share be used to spin things.
Like hyping how your market share went UP. But you lost customers, sales were actually DOWN. Just the other guy lost more. So your market share went up.
I've been in companies that had bad sales, and they super hyped how market share had gone up like we were winning. But the entire market was shrinking enough that total sales had gone down.
googles search was so much better than the competition for so long it would take a big drop in quality for people to get fed up and look elsewhere (en mass)
Tbh I find it annoying everytime it processes for answers when I look for results (I'm using Bing as default engine because I'm stuck on windows 11 and I want to keep it minimally bloated as possible relying only on Edge), waste of cpu resources on my browser I can lookup for answers on my own.
I tried it a few times. Was useless. I asked stuff like "how to do X in Y" and it proceeded to give some generic answer about what X was and then something like "how to this in Y is specific to Y". Well duh.
I guess it can be interesting if you're after some Wikipedia-like information, but in those cases I just go to Wikipedia or similar.
Clearly I lack imagination, so, how to use this in a useful way?
I have Bing enabled on the family computer by default on Windows for the Family Safety features. I tried searching for "rainbow friends official" yesterday to understand what it is (yes I'm that out of touch).
I got tons of what looked like knock offs and spam, with big bright video previews and no official looking site.
I switched to Google and found the official site immediately and realised it's a Roblox game, so the Bing results weren't wrong at all but they sure looked wrong. There's just something about Bing that looks bad, I don't know what it is.
>I got tons of what looked like knock offs and spam, with big bright video previews and no official looking site.
Same for me on most searches. Bing is far too cluttered and busy, all over the place, making your eyes constantly hunt to find what you're looking for.
Microsoft, if you're reading this, simpler is almost always better, no need to make the results page look like an Arch user's terminal session.
This is what Microsoft will always lack: taste and style. Every once in a while they accidentally (-or not) come up with something that is decent but it doesn’t last, next version is atrocious in some way or another.
- Im sure there’s talented professional behind anything that is good and the accident is their decisions surviving the corporate gauntlet of enshitification or expnanding the scope of the project such that it becomes an unpleasant mess, probably obnoxious too.
>This is what Microsoft will always lack: taste and style.
They alwawys lacked an authoritarian UX and design chief, like their own Steve Jobs and Johnny Ive who, for better or worse, would be the only one with the final verdict on what the corporate design of all products should look like and what is allowed to ship.
Microsoft is too fragmented, everyone of its products is basically a totally separate company with its own design language.
I think the absolute best pitch for LLMs is a natural language interface to things like PDFs. The vision of being able to "talk to" a book or paper rather than having to scroll and scroll is compelling. The same obviously also would apply to the Internet. For some reason, they just can't seem to pull that off. Asking ChatGPT to summarize an article is just a disaster.
Instead of doing search yourself, you ask bing chat to tell you the answer.
It searches the web and writes you a cited essay based on the content of the search results. I read the essay it provides to determine which links to follow instead of scrolling a page of links. I’ve found it really useful.
Good for investigating generalities more than specifics, and there is a slight learning curve. It likes short direct questions.
With the added value of having all seo, spam and ads removed, I find Bing chat better than Google search for generalities or recommendations. No more searching for a manual - just ask how to do something and it will find a way. It would be even better if you could configure the search engine. But for shopping it's obviously not very useful.
"But but but journalists from The Verge told me it's going to take over the world" /s
It's a brand issue, people doesn't associate Microsoft with great products, you see it with their gaming division as well, despite insane marketing, they fail to match Sony both in HW and services (Sony has like 3x more, both, console sales and service subscribers than MS)
We need less of both, Google and Microsoft both want to be in a monopoly position, which is bad
From what I remember of the Xbox 360 days it was seen as a great console vs. the PS3, and the Xbox Live service was seen as far superior, and was popular.
X360 released 1 year before the PS3, online gaming was booming, same era as WoW and the PSP, released 1 year earlier, was just as good and was the one to popularize online live services, it released with a web browser and online store! and sold as much as the X360, then the PS3 came 1 year later, after the X360 and catched up
X360: 84m
PS3 87m
X1: 58m PS4:117m
Xbox (X/S): 25m
PS5: 50m (ongoing)
Stats and numbers doesn't lie, people didn't stick with what Microsoft had to offer, not good enough
Good point - so I guess it could be anything from 0% to 40%, although I guess my broader point is that the measure of absolute percentage points could hide significant ‘actual’ growth.
it was the dumbest question ever, too! that might have made it worse. "What happens to the virus plot line in the movie Office Space?" something along those lines.
I have similar experience with ChatGPT. Sometimes I'm hesitating to use it because I feel that it is better to not get information then get information which may or may not be true.
That Microsoft only made Bing Chat available from Edge shows that they still don't understand the web. Everyone uses Chrome today, because they started using Google first, from whatever browser they were using.
It’s available on safari on iOS and Mac. I think it’s limited to 3 queries without signing in. I think originally it was only bing app or edge app but no longer the case.
It was available there first. It was a value-add for using Edge over Chrome, and it was somewhat effective for about 3 months like that. I started using Edge to have free access to that and I enjoyed it at the time.
That event caused me to take another look at Edge, and I now prefer it to Chrome and use it as my daily driver at work. I am split between Edge and Brave at home.
Honestly, Microsoft has done a good job with a browser (finally). It only took them 25 years.
My parents don't even know about chatgpt, the current gen of ai, &c. and they're not even retired yet. It's a bubble, most peeople don't know about it and most of the people who know about it heard of it through TV/Web news with "AI will replace your job in 6 months" kind of BS.
idk why bing would become super popular all of a sudden when a good 90% of the web users are tech illiterate. In most people's head internet = google = chrome (and apple = safari because it just works)
Even in my tech literate circles the new flavor of AI is barely used
Your parents aren't exactly a predictor of current or future markets. Most people under 30 with the financial ability to buy a subscription have heard of ChatGPT. That's the only thing that counts in market terms.
Agreed that origin commenter's two parents are not a great enough sample size to gauge the popularity of generative AI implementations such as ChatGPT.
The data point is the article and reality. My comment was just part of the answer, aka real world AI usage is nowhere close to what people want to believe here.
One commenter says everyone under 30 has a paid chatgpt account, yet chatgpt as a self reported user base of 185 millions, with only 100 millions unique active user per week, a far cry from every people under 30 with a spare 20$.
Even if every single one of these people installed bing on all their devices it wouldn't make a dent... chrome is used by more than 3b people.
I took a bet last year on this very forum about how chatgpt wouldn't become anywhere to what the HN prophets promised, I can take another bet for 2025 lmao
GP's parents aren't a good datapoint because LLMs are still in the early adopter phase. Which tracks with LLMs not making a big difference in Bing's market share (yet).
I have a lot of friends who have incredibly varied backgrounds and this rings true for me.
A friend who grew up in the worst rural schools, still lives in mobile home 2.5 hours from the nearest major city, driving 35 minutes through the country to their restaurant job in their grandmothers car, while taking care of the 3-year old kid they had when they were 18 while co-parenting with their divorced husband, while dealing with felony drug charges, and makes <$1,500/month has absolutely grokked the realities of ChatGPT and Midjourney without me mentioning anything about it to them. Instagram and TikTok expose people to these trends and mobile apps with free accounts open up immediate exploration to the masses.
Two friends who are 35 year old mostly-unemployed interior designers with an anthropology degree and no college degree are using it for drafting quotes.
Every lawyer, MBB consultant, and computer programmer I know either uses them heavily or are aware of colleagues who do.
I tutor a ton of students at a mid-tier state school, every single student I know in high school or university are using ChatGPT for their schoolwork.
And the 5-7 year olds I know have much worse typing skills than I had at that age, because they use Siri to do 90% of the things on their phone/tablets. It annoys the shit out of me because it looks so painful and the noise of them talking to their phone is obnoxious, but I can absolutely see the paradigm changing to where the youngest generation might not even use keyboards for their first choice as a computing interface. Powerful LLM's will greatly improve their interaction with Siri.
Ok, let's do the math. How many pinterest and quora users are paying for a subscription? 0? How many chrome users? 0? How many ChatGPT users? 100m. That's far more impressive than having 3b users who can't or don't want to pay anything.
Are pinterest, quora and google relying on the ads business model? Sure. OpenAI? Not so much. They have the users actually paying for the service.
Does any of that have anything to do with bing market share ? (also no, not 100% of chatgpt users are paid users, far from it)
Impressive or not I don't really care, all I'm saying is that even if they all left chrome to go to bing (and they didn't) it wouldn't have been much more than a 1% market share movement, which is what we've seen, hence there is no reason to be surprised.
People generally overestimate the short term effects of new technologies but underestimate the long term effects.
I've heard the exact same things about smartphones in the past and now even my small village of 500 inhabitants is communicating with an app. It took a while though and suddenly, it was everywhere.
The previous generations of AI (translation, text & image recognition...) are already widespread and there's no reason to think it won't be the same here in the future.
My grandparents barely ever figured out how to use a computer enough to use e-mail, despite email being quite popular for the last 25 years of their lives. It was always hit-or-miss whether they'd be able to open up the email client.
I wouldn't use people who are 50+ as bellwethers to determine if something is a bubble vs. a new paradigm.
LLM's will absolutely be the next transformative "interface" that we use to interact with our computing devices. They're just too powerful to ignore. Even the current generation would be amazing for my MacBook. Recently I had an issue at work where I took a new job that uses MacBooks instead of Windows and couldn't get Slack notifications to pop up reliably. I swore all my notification settings were correct both in the app and in system settings. Turns out that my laptop had "Do Not Disturb" mode on due to seeing that my work phone had it on. If Siri had a robust LLM that could connect to system and app settings via API, a high-quality LLM like ChatGPT absolutely would have automatically figured it out and been able to both tell me why the notifications weren't popping up and ask me if I'd like it to go ahead and fix the settings for me.
LLM's will simplify complex computing systems for so many users. Yes, it will suck for tons of systems which implement it poorly or aren't conducive to speech/textual communication. But that's mostly a "skill-issue" for development teams implementing LLM's in apps/OS's, not a fundamental issue with the technology itself.
LLM's may be approaching a temporary bubble, but we won't be in that "bubble" until after Windows, Mac, and Android build them in as first-class interfaces for one option for primary interaction with your devices.
We already have touch GUI, mouse+keyboard GUI, and CLI interfaces, but will now get typing/speech-based LLM interfaces as a new paradigm.
Non-technical people overwhelmingly buy Windows PCs.
Windows PCs don't come with Chrome or Firefox or Brave. Non-technical people use the default IE preset to Bing.
Windows makes it a pain for non-technical people to switch browsers or search engines.
Even on a slanted playing field, Bing relevance is so bad that even non-technical people manage to climb over the hurdles Microsoft places in front of them to escape their defaults.
That’s global which is hardly relevant because for most people outside of highly developed countries MacBooks are prohibitively expensive so aren’t even an option.
Apple seems to be at 28% in the US (a few percent below HP and above Dell). I’d guess that if we removed corporate sales it would be considerably higher.
I‘m sorry, I was not aware there were arbitrary restraints on the subject. Given that no one specified any country before it became disadvantageous to Apple‘s market share. I am more than happy to discuss the USofA‘s side of things.
> I’d guess that if we removed corporate sales it would be considerably higher.
I highly doubt removing corporate sales would be an appreciable benefit to Apple. If anything it may be a detriment. Consumers in the US (and globally) cannot afford nor do they prefer Macbooks. Just because they are common in Silicon Valley does not mean they are common elsewhere.
Additionally, what is your data source? For US-focused statistics I have found this[0], which is quoting Statistica, which says Apple is a firm 3rd at 24%, 3% below Dell and a whopping 11% below HP. However it is from 2021, so more recent statistics are always a benefit.
Where does the article specify this comes from exclusively corporate sales? Admittedly I haven’t read every word of the article, it‘s rather verbose, but given another scan I cannot find any support for that claim. Nor in this US-centric article: https://techaeris.com/2021/11/08/the-most-popular-laptop-bra...
Additionally, do you have any counter evidence that asserts that Apple‘s Macbook is the most common consumer laptop?
What is the percentage of Google searches actually performed by humans? Maybe if we all switched there would still be so much bot traffic the market share would barely moved? Then again I haven't started using it and I don't know anyone who has so at least in my bubble it's not surprising
Following the link to the StatCounter[0] source, it seems bing chat isn't taken into account in the first place, nor do they have an acurate view of the search activity.
To add to that since it's an analytics plugin, it could also affect how those pages are ranked in Google and Bing in not so subtle ways.
All in all, that feels like very little facts to come up with any opinion on the subject matter.
> We have no way of measuring the number of queries performed in bing chat. However, we also don't measure the number of queries to regular search engines like bing or google either. Instead we track search engine referrals.
Ai search is unrealiable and of low quality. It limits discoverability and reduces exposure to novel content and it is no use to businesses. Sadly google is also experimenting with ai and as such their results are of lower quality, but still far better than bing.
I used chatgpt to try to fix some issues with certbot, docker, and nodejs, I just gave up and went straight to forums, etc... I found out what I was looking for by interacting and putting together the answer out of those interactions. Of course, chatGPT is not "intelligent" to comprehend but that's how "open" AI and tech bros are selling it. I didn't even flinch when Microsoft announced it was going to implement AI in its search engine. Habits are difficult to change.
I'm still using Search engines (Brave most of the time and google when I can't find what I'm looking for), tech forums, StackOverflow (Still a tech gold mine), etc... and lastly chatgpt if I have time to write a really good prompt.
The bubble is still there, few are paying attention outside it, let alone using it to be "rich". Heck, Crypto Bros are now selling courses on how to monetize chatGPT, and what's worse, people are buying it! It's the same cycle over and over again.
Yes, LLMs brought decades of R&D to the forefront of society and put that piece of tech on people's hands yet that doesn't mean everyone will adopt it just because. Our phones are engineering marvels yet we use them for mundane things most of the time.
128 comments
[ 26.1 ms ] story [ 5517 ms ] threadThe model is too reliant on search results which are often wrong, but it treats them like gospel.
Generative AI features can enhance search somewhat but aren’t a replacement. It’s not enough of a driver on its own to take market share from Google.
Maybe. But the opinion that Google is returning less relevant results is sooner or later going to go mainstream unless they start returning relevant results.
We're users too, but we notice things first. Just because the other users haven't noticed them doesn't mean that they won't in the near future.
I am a DuckDuckGo user and I still use the !G flag when the results I’m getting aren’t good. Every time I go over to Google the results improve slightly.
ChatGPT is a game changer when it comes to getting a system to understand the underlying meaning of my query. I used to give the example of searching for Frank Lloyd Wright’s height - doing so on a search engine would give you results about the buildings themselves. You’d have to dig for the simple information of how tall the dude was.
Alas, Google actually fixed this and quickly displays the man’s actual height now.
Microsoft and Google are both going to integrate AI into their search engines and they’ll both improve drastically as a result. We’re only scratching the surface on the subject.
Recipes are the classic example. Because Google upranks sites that people spend more time on after a click, most recipe searches yield pages that start with the history of flour milling and what milk is, before eventually getting to the desired recipe 5000 words later.
The phenomenon with recipes that you mention has nothing to do with Google: it has more to do with copyright. You can’t copyright a recipe, but you can copyright written commentary about cooking a recipe.
The recipe blogs also need enough content to place ads around.
So, the real optimum is ranking for one or more popular searches (regardless of relevance), looking vaguely interesting, and just confusing enough that it takes a couple of seconds to realize this isn't the page you wanted. Either that, or being exactly what someone wants for one of the popular searches.
Though, things may have changed, and my memory definitely has faded.
Also, plenty of SEOs may have misunderstood short clicks as a quality signal, so even if maximizing time on site doesn't help, I wouldn't be surprised if SEOs still recommend it.
> Alas, Google actually fixed this and quickly displays the man’s actual height now.
That's really the only thing I want: A search engine that understands my queries. I don't want to chat with a computer.
Altavista et al got spammed into unusable mess and their use collapsed almost overnight when PageRank (temporarily as we've learned) managed to filter out the spam.
Of course nowadays Google's tentacles are so deeply everywhere that they will churn on with subpar technology just like e.g. Oracle, SAP, IBM, Microsoft etc etc. The oligopoly is probably here to stay.
It was more temporary (shorter time period) than you probably realize.
I was a developer on the Google web indexing system. PageRank was only one of many signals that went into ranking, and PageRank was gamed by link farms pretty early. There's a whole team dedicated to fighting webspam, and their spam signals are some of the other inputs to the scorer.
By Google's golden years, there was a whole lot of secret sauce going into fighting webspam, and PageRank had been pretty thoroughly gamed.
PageRank was important, but I think its importance is generally over-estimated.
Also, about 15 years ago, Google was testing another link-graph-based metric that was more efficient to calculate and less susceptible to linkfarms. I imagine PageRank (as described in the patent/Wikipedia) has been replaced, though perhaps the new algorithm is still referred to as PageRank. (IIRC, the proposed replacement was being called PageRankNG while in testing.)
Less relevant compared to what?
Consumers don’t see this as Google going downhill, they see it as the internet going downhill.
And I agree. It’s not as though I can jump over to DuckDuckGo or Bing and my search quality radically increases. Usually the opposite.
ChatGPT is helpful because it gives a very convincing feeling of being right. Whether or not it’s actually right is another question.
Compared to what the user wanted when they ran the search in the first place.
> It’s not as though I can jump over to DuckDuckGo or Bing and my search quality radically increases. Usually the opposite.
Of course it's a personal anecdote but as a DDG user since 2014 I routinely add !g to my search when the results don't satisfy me (which happen every once in a while, say once a month). This used to help, as most of the time Google had the answer I was looking for and could not find otherwise, but I don't think it happened even once over the past two years: Google is now at least as bad as DDG when it used to be strongly superior.
I acknowledge that it is a paid service, but you haven't tried Kagi yet you should give it a go! Default Search quality is very good, with no ads, tracking, or telemetry, and you can use Lenses and customizations to get even better results. You can even block certain sites from appearing in search (or simply de-rank them), which has dramatically reduced Pinterest and other SEO spam for me.
And I think they're right. More and more information on the internet is siloed into apps or behind paywalls. Knowledgeable people who used to write technical articles now make youtube videos (far less searchable). And the good stuff on the internet is diluted with low quality content written by someone who has to churn out an article every 15 minutes that is hard to filter out, even as a human. (I wasted 3 hours yesterday because a website gave me the wrong formula for something)
They can answer a lot of questions in a lot of ways that Google just cannot.
Their position is entrenched enough that they'd be able to exist just on brand fumes for a good number of years yet but once the rot sets in there's no coming back.
However, I don't understand how statcounter got the data about usage it the first place. Their methodology might be flawed.
Tech people will change browsers, search engines, and SaaS tools without much thought. The average person learns something and sticks with it for a very long time. Bing is playing a long game.
There are a lot of examples of things people in tech bubbles think that have little relevance in the real world: That Google can’t be trusted, that everyone wants laptops with upgradable RAM and phones with swappable batteries, that desktop websites should take priority over mobile, that offline apps are preferred over cloud apps. Spend some time in product management and you realize that tech people’s demands are unique.
Even the claimed demands of tech bubble people often fail to materialize into sales. Just look at all the demand for mini iPhones in tech bubbles compared to how little they actually sold.
Non tech savvy people don't even realize they are using Chrome in android or Safari on an iphone. They will never change they just use the default search bar that is on their device.
An artist I follow on Youtube put out this (comical) video regarding it - https://www.youtube.com/watch?v=jrFv1O4dbqY
Maybe not having tons of distractions in the form of advertisements on the search homepage allows people to achieve more.
Just uses duckduckgo for search, extracts text from top 8 results, shoves into gpt-3.
(I think I got a massive cost reduction by filtering paragraphs with embedding API, but that's just an optimization.)
Phind is better than what I built, though. I don't actually use it that much (I'm not sure why—habit? Impatience?), but every time I do I'm surprised with the accuracy.
I think a big part of it is which pages to select for sources. Google Bard failed hard here in my testing, gave nonsense answers because its sources were AI generated SEO spam!
It seems that the main problem right now is that they only look at the first few results. Basically they all do this:
1. Ask model for a search query.
2. Run search and feed the results into the model.
3. Ask model for an answer.
The problem is that for very specific questions the search results won't contain the required information. So it has to say something irrelevant or hallucinate. Especially for negative queries like "a foo library that doesn't depend on std" or similar it really struggles because it can't effectively filter out libraries that fail the requirements and can't keep searching until it finds one.
Basically they can be fast if a decent query can have the answer within the first page of results. But fail otherwise. In many cases you are better off just reading through search results on your own.
I'm using Google less and less, but not using Bing. It is more like Google is not as useful anymore.
So would like to see a trend of usage rates, to see if Bing market share didn't go up much, did overall Google usage go down?
Market share might have only gone up 1%, but did the overall market shrink.
Like hyping how your market share went UP. But you lost customers, sales were actually DOWN. Just the other guy lost more. So your market share went up.
I've been in companies that had bad sales, and they super hyped how market share had gone up like we were winning. But the entire market was shrinking enough that total sales had gone down.
I guess it can be interesting if you're after some Wikipedia-like information, but in those cases I just go to Wikipedia or similar.
Clearly I lack imagination, so, how to use this in a useful way?
I got tons of what looked like knock offs and spam, with big bright video previews and no official looking site.
I switched to Google and found the official site immediately and realised it's a Roblox game, so the Bing results weren't wrong at all but they sure looked wrong. There's just something about Bing that looks bad, I don't know what it is.
Same for me on most searches. Bing is far too cluttered and busy, all over the place, making your eyes constantly hunt to find what you're looking for.
Microsoft, if you're reading this, simpler is almost always better, no need to make the results page look like an Arch user's terminal session.
- Im sure there’s talented professional behind anything that is good and the accident is their decisions surviving the corporate gauntlet of enshitification or expnanding the scope of the project such that it becomes an unpleasant mess, probably obnoxious too.
They alwawys lacked an authoritarian UX and design chief, like their own Steve Jobs and Johnny Ive who, for better or worse, would be the only one with the final verdict on what the corporate design of all products should look like and what is allowed to ship.
Microsoft is too fragmented, everyone of its products is basically a totally separate company with its own design language.
It searches the web and writes you a cited essay based on the content of the search results. I read the essay it provides to determine which links to follow instead of scrolling a page of links. I’ve found it really useful.
Good for investigating generalities more than specifics, and there is a slight learning curve. It likes short direct questions.
e. Punctuation
I guess that's the main issue then for me. I'm seldom after generalities, and very often after specifics.
It's a brand issue, people doesn't associate Microsoft with great products, you see it with their gaming division as well, despite insane marketing, they fail to match Sony both in HW and services (Sony has like 3x more, both, console sales and service subscribers than MS)
We need less of both, Google and Microsoft both want to be in a monopoly position, which is bad
X360: 84m PS3 87m
X1: 58m PS4:117m
Xbox (X/S): 25m PS5: 50m (ongoing)
Stats and numbers doesn't lie, people didn't stick with what Microsoft had to offer, not good enough
it is "less than 1%"
Nope Perplexity gave me a totally wrong answer.
I cannot tell you how upsetting a user experience this was to me. Thinking about canceling my Pro sub off of one wrong answer. Can’t trust it.
I have similar experience with ChatGPT. Sometimes I'm hesitating to use it because I feel that it is better to not get information then get information which may or may not be true.
That event caused me to take another look at Edge, and I now prefer it to Chrome and use it as my daily driver at work. I am split between Edge and Brave at home.
Honestly, Microsoft has done a good job with a browser (finally). It only took them 25 years.
idk why bing would become super popular all of a sudden when a good 90% of the web users are tech illiterate. In most people's head internet = google = chrome (and apple = safari because it just works)
Even in my tech literate circles the new flavor of AI is barely used
One commenter says everyone under 30 has a paid chatgpt account, yet chatgpt as a self reported user base of 185 millions, with only 100 millions unique active user per week, a far cry from every people under 30 with a spare 20$.
Even if every single one of these people installed bing on all their devices it wouldn't make a dent... chrome is used by more than 3b people.
I took a bet last year on this very forum about how chatgpt wouldn't become anywhere to what the HN prophets promised, I can take another bet for 2025 lmao
https://www.demandsage.com/chatgpt-statistics/
https://www.statista.com/statistics/543218/worldwide-interne...
You formulated this as if you are sure of it and it is a fact. Could you link a source?
A friend who grew up in the worst rural schools, still lives in mobile home 2.5 hours from the nearest major city, driving 35 minutes through the country to their restaurant job in their grandmothers car, while taking care of the 3-year old kid they had when they were 18 while co-parenting with their divorced husband, while dealing with felony drug charges, and makes <$1,500/month has absolutely grokked the realities of ChatGPT and Midjourney without me mentioning anything about it to them. Instagram and TikTok expose people to these trends and mobile apps with free accounts open up immediate exploration to the masses.
Two friends who are 35 year old mostly-unemployed interior designers with an anthropology degree and no college degree are using it for drafting quotes.
Every lawyer, MBB consultant, and computer programmer I know either uses them heavily or are aware of colleagues who do.
I tutor a ton of students at a mid-tier state school, every single student I know in high school or university are using ChatGPT for their schoolwork.
And the 5-7 year olds I know have much worse typing skills than I had at that age, because they use Siri to do 90% of the things on their phone/tablets. It annoys the shit out of me because it looks so painful and the noise of them talking to their phone is obnoxious, but I can absolutely see the paradigm changing to where the youngest generation might not even use keyboards for their first choice as a computing interface. Powerful LLM's will greatly improve their interaction with Siri.
Sure but they have two phones, two laptops and a desktop, all of them running chrome with absolutely no reason to switch
Hence I'm not surprised adding a gimmick to bing didn't create any market share movement
100m active chatgpt users
3b chrome users
Do the maths, even pinterest and quora have 2-4 times more active users
Are pinterest, quora and google relying on the ads business model? Sure. OpenAI? Not so much. They have the users actually paying for the service.
Impressive or not I don't really care, all I'm saying is that even if they all left chrome to go to bing (and they didn't) it wouldn't have been much more than a 1% market share movement, which is what we've seen, hence there is no reason to be surprised.
100m people are paying for ChatGPT?
I've heard the exact same things about smartphones in the past and now even my small village of 500 inhabitants is communicating with an app. It took a while though and suddenly, it was everywhere.
The previous generations of AI (translation, text & image recognition...) are already widespread and there's no reason to think it won't be the same here in the future.
I wouldn't use people who are 50+ as bellwethers to determine if something is a bubble vs. a new paradigm.
LLM's will absolutely be the next transformative "interface" that we use to interact with our computing devices. They're just too powerful to ignore. Even the current generation would be amazing for my MacBook. Recently I had an issue at work where I took a new job that uses MacBooks instead of Windows and couldn't get Slack notifications to pop up reliably. I swore all my notification settings were correct both in the app and in system settings. Turns out that my laptop had "Do Not Disturb" mode on due to seeing that my work phone had it on. If Siri had a robust LLM that could connect to system and app settings via API, a high-quality LLM like ChatGPT absolutely would have automatically figured it out and been able to both tell me why the notifications weren't popping up and ask me if I'd like it to go ahead and fix the settings for me.
LLM's will simplify complex computing systems for so many users. Yes, it will suck for tons of systems which implement it poorly or aren't conducive to speech/textual communication. But that's mostly a "skill-issue" for development teams implementing LLM's in apps/OS's, not a fundamental issue with the technology itself.
LLM's may be approaching a temporary bubble, but we won't be in that "bubble" until after Windows, Mac, and Android build them in as first-class interfaces for one option for primary interaction with your devices.
We already have touch GUI, mouse+keyboard GUI, and CLI interfaces, but will now get typing/speech-based LLM interfaces as a new paradigm.
Windows PCs don't come with Chrome or Firefox or Brave. Non-technical people use the default IE preset to Bing.
Windows makes it a pain for non-technical people to switch browsers or search engines.
Even on a slanted playing field, Bing relevance is so bad that even non-technical people manage to climb over the hurdles Microsoft places in front of them to escape their defaults.
https://fortunly.com/articles/lap-top-market-share/
Apple seems to be at 28% in the US (a few percent below HP and above Dell). I’d guess that if we removed corporate sales it would be considerably higher.
I‘m sorry, I was not aware there were arbitrary restraints on the subject. Given that no one specified any country before it became disadvantageous to Apple‘s market share. I am more than happy to discuss the USofA‘s side of things.
> I’d guess that if we removed corporate sales it would be considerably higher.
I highly doubt removing corporate sales would be an appreciable benefit to Apple. If anything it may be a detriment. Consumers in the US (and globally) cannot afford nor do they prefer Macbooks. Just because they are common in Silicon Valley does not mean they are common elsewhere.
Additionally, what is your data source? For US-focused statistics I have found this[0], which is quoting Statistica, which says Apple is a firm 3rd at 24%, 3% below Dell and a whopping 11% below HP. However it is from 2021, so more recent statistics are always a benefit.
[0] - https://techaeris.com/2021/11/08/the-most-popular-laptop-bra...
Additionally, do you have any counter evidence that asserts that Apple‘s Macbook is the most common consumer laptop?
To add to that since it's an analytics plugin, it could also affect how those pages are ranked in Google and Bing in not so subtle ways.
All in all, that feels like very little facts to come up with any opinion on the subject matter.
[0] https://gs.statcounter.com/faq#methodology
> Is bing chat included in search engine share?
> We have no way of measuring the number of queries performed in bing chat. However, we also don't measure the number of queries to regular search engines like bing or google either. Instead we track search engine referrals.
I'm still using Search engines (Brave most of the time and google when I can't find what I'm looking for), tech forums, StackOverflow (Still a tech gold mine), etc... and lastly chatgpt if I have time to write a really good prompt.
The bubble is still there, few are paying attention outside it, let alone using it to be "rich". Heck, Crypto Bros are now selling courses on how to monetize chatGPT, and what's worse, people are buying it! It's the same cycle over and over again.
Yes, LLMs brought decades of R&D to the forefront of society and put that piece of tech on people's hands yet that doesn't mean everyone will adopt it just because. Our phones are engineering marvels yet we use them for mundane things most of the time.