Here is a screenshot of my Cloudfront dashboard showing the breakdown of user agents into particular browsers when an article of mine reached the top of HN about two months ago:
Looking at the number of distinct IP addresses in Athena for that day, about 20,200 of those hits were distinct viewers:
This is not definitive, but it's something I guess. Here are a few observations based on that data:
1. On the desktop, it looks like Chrome and Firefox are at about parity for HN users. I kind of doubt this actually, and think it's more likely that Chrome still beats Firefox by a sizable margin. But even if this is off, it looks like HN users actually do use Firefox significantly more than the general population, which means the evangelization here isn't just a vocal minority among HN users.
2. On desktop Safari is a distant third. It would appear more people browse HN in any mobile category than browse HN on desktop using Safari.
3. On mobile, mobile Safari dominates. Then mobile Chrome has about half as many users as mobile Safari, and mobile Firefox again half as many as mobile Chrome.
EDIT: I pulled an Athena report of the referrer breakdown as well, if anyone's curious:
As a user Safari is fantastic and basically the whole reason I even own an iPhone. It’s just stupidly faster than Chrome on any Android flagship.
A world where there’s “choice” of browsers on every platform in practice just cements Chrome as the web because if you didn’t have to develop for Safari nobody would. On every other platform it’s “have bug-for-bug compatibility with Chrome or die.”
Not only is safari full of bugs, the bugs won't be fixed, and if they do, it will take years.
Current serious bugs: 1. Add a site to the home screen, go to another app for 20 seconds and then switch back: congratulations, it's frozen. 2. You can't even stop scrolling on the body element, so tens of thousands of developers resorts to ugly hacks like eg. https://github.com/willmcpo/body-scroll-lock that kinda works in some situations. 3. The lack of features. Who doesn't as a developer feel bad for the person who spends forever trying every single browser in apples app store in an attempt to find a browser that supports push notifications. Of course none of them do, it's all Apple's shit.
You really can't have much experience with web development if you don't know how terrible iOS is for web as a result of Safari.
Heck, Safari isn't even behaving close to Safari. You have safari on desktop, safari on iOS and safari standalone on iOS. All of them behaves completely different and has different bugs and feature support.
I wouldn't even bother owning an iOS device if it weren't for the third-party app selection (specifically for art production). Which is the most infuriating part: devs could do all of this stuff on Android, they just refuse to.
That's not fair. There are good reasons Android doesn't have art production apps. The biggest is that Android users don't pay for apps like that and devs like eating, but iOS has several other advantages like the Pencil, the better graphics APIs, the better graphics performance, etc.
This is more the result of Android being the budget option.
You make it sound as though developers have some grand scheme against Android. Devs do prioritize iOS, but if you want it to change you (and Google) should understand the reasons for it. ARKit vs ARCore is a great example.
Tango launched on a single phone that no one bought, it required specialized hardware, and it had serious performance issues. Google forged ahead with the hardware requirements for years, in which the market for Tango apps was zero. They gave it limited support and it was clearly not a priority. It was only after Apple launched ARKit all at once for every modern iDevice, in a big presentation to make it clear this would be a major iOS going forward, that Google killed off Tango and screwed over all the devs who had bought in (with a tweet, for the extra insult to injury). They launched ARCore instead, which still didn't run on the majority of Android phones and still didn't track as well as ARKit.
ARCore isn't the only time Google has behaved this way, it happens with nearly all of their efforts. Do you see why developers don't engage with these kinds of practices?
No, it's simply the same sort of tribalism that drives so much of American culture. The black-and-white, good-vs-evil format warring is nothing new; as always, it's driven by a kernel of legitimate difference and a whole lot of snowballing bias. What needs to happen is for developers to pull their heads out of their asses, realize just how enormous the Android market is, and supply experiences that justify paying for software (especially in a climate where most users don't see anywhere near the value they give in personal data returned in software utility).
>Tango launched on a single phone that no one bought
That doesn't matter. Tango was the cutting edge of mobile XR, which is a space that still won't be mature and profitable for years from now. The point would have been to get a jumpstart on developing port-able technologies and UX norms. Can you imagine how much more solid app development would have been if someone had showed up in 2003 and said, "In 10 years, multi-touch display smartphones will become the norm. Here's a dev kit that approximates what will be possible."? That's Tango. I don't know how you can argue that opportunities weren't missed or pushed back years because of this platform bias. Now the space is even contracting somewhat because Apple marketed ARKit (and forced Google to market ARCore) as a consumer-ready platform, and companies are realizing that it isn't (control isn't figured out, UX isn't figured out, applications aren't figured out; we're JUST getting the basic technology layer above ARKit/Core figured out). If we'd taken Tango seriously, we'd be so much farther ahead.
This goes for so much in the iOS/Android dynamic. It's dumb.
I really like addons and that's why I don't use chrome on mobile.
Surprisingly I also started using Firefox Focus. At first it looked like nonsense to me, no addons. But then I found out it sometimes makes sense to start from scratch, focus on some topic and delete all tabs with one tap when done. And if I find what I searched for or want to read it later, I share the page to Saved messages on my Telegram. This workflow works surprisingly well for me.
But then, with Bush Sr and then Clinton, the rules changed; Microsoft started playing the political game, and essentially avoided a similar fate.
There is no antitrust enforcement anymore. Breaking a company once indeed only buys a little time. But getting regulation workiNg again WILL work, and will likely include breaking them.
If only the US would vote in politicians that cared about competition. I can’t find any amongst either presidential or congressional candidates, though.
From what we've seen with GPT-3, we can train one machine learning model by ingesting the entire scraped web and a whole lot of (online/offline?) books, and it can often produce correct answers to questions. The accuracy of those answers, presuming the questions are not subjective ones, is only going to increase. Google's status, as a search engine, in 2025 is not something I would bet money on.
The iOS/Safari thing is another issue, which perhaps other anti-trust regulators will address in Google's favor. Assuming they do not, consider that if Google stops paying Apple many billions of dollars a year and another company (or Apple) replaces the default iOS search, Google's search ad business will be severely and irreversibly damaged. Removing iOS defaults would not only remove a large chunk of Google's US mobile search traffic, it would be the more affluent chunk. Everyone who knows about marketplace dynamics knows how damaging that would be to Google if that inventory ended up with Microsoft or Facebook.
Even without regulatory intervention (which is almost 100% guaranteed at this point, barring some act of god, though I don't know how large that intervention will be) Google's days atop the mountain are certainly drawing to a close. Perhaps the ground has already given way, they know it, and the maniacal stupidity of trying to coral the web into some AMP graveyard is a reaction of desperation.
To address the original blog post, I think the author is way off the mark on video. Youtube is big, but far from the king. Google has repeatedly flopped on the social side, and isn't going to be allowed to buy TikTok. They don't appear to be taking any ground paid streaming services like Netflix or Amazon. They are big, but no where near as important relative to their dominance in search.
Hopefully those same anti-trust regulators also have statistics for browser market shares :)
What other alternatives are there for Youtube, that are even in the same region as Youtube?
Additionally, separate to instagram / facebook / whatsapp ; possibly also facebook social / facebook marketplace / facebook events ; They should have a level of access that others have. Imagine Yahoo users could subscribe to Facebook events without ever setting foot in facebook.com ... I know it sounds like science fiction, but it's possible. like email....
Or you could break up search / browser / advertising / mobile; search and browser would need to invent other ways to monetize themselves.
It's actually pretty cool, and earnings from web monetized users now cover ~3% of the operation of the site during the first month of having it implemented.
You basically just have to add a meta tag to the site that directs to a payment pointer (you also have to sign up with one of the wallets for the payment pointer).
Definitely not getting rich from the implementation, and the earnings are much lower than if I just ran ads, but it is a nice feeling to have the users of the site directly contribute to supporting the product. Plus it is such a low barrier to set it up that there is very little downside to having it added to the site.
Implementing simple "Web Bucks" is not pedestrian at all. There are real reasons why all Paypals suck, and you can't simply ignore them and hope to build an open cash alternative unbothered.
"Centralization wins" is the wrong conclusion from this dataset. Discord & Slack have seen far more investment in UX & backend than IRC & Jabber. Because chat is addictive, relatively easy to implement, and easy for users to access via website or app, there's huge incentive roll your own chat service, and then incentive to differentiate from innumerable competitors. Many users are on half a dozen chat sites.
Email is going strong, even tho it's seen rather little UX investment. Decentralization is one feature keeping it afloat in a world of uncountable chat alternatives.
I'd still argue that it correlates with centralization. We could argue about causation but it still goes hand in hand IMO.
A company is much less likely to invest resources into improving IRC or Jabber because if you made huge breaking changes you'd end up with an incompatible, custom fork of the protocol (negating the benefit of using a standard protocol, since you'd effectively lose the interop with third party client) and even if you kept the standards open and federation possible you're basically hurting your own business since you make it easy for people to use your work without giving you a way to monetize it.
If your objective is to be profitable it makes a lot more sense to just create your own centralized, proprietary protocol and clients that you control end-to-end. Standardizing and federating your protocol is a massive amount of headache in the long run for very little benefit in terms of $$$.
Except most emails end up at Gmail SMTP servers (@gmail.com and hosted G Suite).
slack/discord will probably disappear and be replaced by another closed system until we find the right way to do real time chat in a decentralised way.
I don't know if it's still in the pipeline but I remember that Google proposed an "AMP for email" not so long ago, with basically self-updating email contents.
Open email can be severely damaged if good old SMTP is only used to tunnel proprietary formats, especially if those formats are effectively just metadata that's used to fetch the actual content of the email from 3rd parties.
[1] https://en.wikipedia.org/wiki/Transport_Neutral_Encapsulatio...
Most companies use mail internally or to interact with customers usually using some type of external mailing service.....
If SMTP is only used to tunnel proprietary content à la winmail.dat, it doesn't really matter that it's still technically open.
Isn't that what IRC was supposed to be? But as others have pointed out, it's trivial to roll your own chat software and monetize it. You can't do that with a decentralized standard.
> I suspect that if you broke those companies up you'd only win in the short term, then it would consolidate again.
The natural course of humans is to walk around on the ground too, yet we fly. Regardless of whether it takes energy to do, it's something we should obviously do anyway because the alternative is worse. Sorry for the bad analogy but "it's natural" and "hard" isn't a valid argument against doing something about something.
I don't really buy your story anyway though. It assumes it was a fair fight and I simply don't think that's true in the case of the web.
Genuine question, do you mind saying (or pointing to something) what are the implications of browser choice of individuals reading HN?
I think at some point Chrome became significantly better than Firefox and did a better job listening to users, so power-users (developers) started switching, once they switched, they primarily developed for the browser they use themselves (it's just more convenient when you get a link to look at to open the inspector in the browser that is set as system standard), so support improved further on that platform, so market share would constantly rise. This in turn leads to more devs reducing their focus on Firefox, so users turn away when a site looks or functions worse.
Now, Firefox has caught up on a bunch of things, yet Google's market share is huge both in browsing and advertising. It would be good for more developers to actively support the strongest alternative platform, because this in turn might affect adoption of that browser, even if it's just small numbers (every site that doesn't work might lead to a user switching to Chrome). It would keep them at least a little more safe from a monopoly that might start imposing policies that are bad for their business, the same way app stores are causing issues because they are the only way to access mobile phones (outside of web app - which is the point of the whole discussion)
Edit: note there's probably many reasons - this is just one way of looking at it.
Easier to monetize means you can fund all the work required to achieve superior user experience, and user experience is the most important differentiator. People have repeatedly shown that UX is more important than security, privacy, cost, freedom, openness... anything. If it "just works" and can be adopted with little or no friction, it wins.
Yes, central structures allow capturing of monetary value, which provide economic incentive to people and that is a powerful motivator for attracting good UX/UI/Designers to make the product improve mass appeal
For whatever reason, good UI/UX hasn't been the FOSS's strong suite. Until that changes (and there's also no reason why it can't change) it will continue to lose ground on anything that is user-facing. Anything that requires expertise however, FOSS will continue to shine.
Just like democracy, you are never "done". "Oh great we are in a democracy nothing can remove that from us!" nope, gotta fight to keep it that way...
I know that your parent explicitly mentioned the US, but let's not pretend either that the US found the, or even a, right answer to structuring government, or that no other country has found their own answers.
There's not one right answer here.
For many people, it probably is—the key to its longevity (and it's very young on the global stage) is working well enough for enough people enough of the time—but there are plenty of people for whom it's not working—me and plenty of other people in the US, but also the global victims of the international policies of the US. (Before the whataboutism comes in, this is not to say that other countries don't also have dissatisfied residents or bad international policies. I was responding specifically to a comment whose wording could be read as saying that the US, rather than any other country one might want to discuss, had found the right answer to government.)
> There's not one right answer here.
I agree, and that's why I was urging against wording that seemed to imply that the US's answer was the answer to 'correct' government.
Personally I'd like to live under a much more European model of governance, but that, too, isn't perfect. (And, of course, one way to do that is to move to Europe. But this thread started not from a debate about the best place to live, but from my response to a post whose wording, "we figured it out ~250 years ago", could be read as suggesting that the US had found the right solution to government.)
But it is meaningful to disagree with "we figured it out ~250 years ago" (a quote from the post https://news.ycombinator.com/item?id=24178311 to which I was originally responding), which says that we already know the answer, even if I don't know the answer (and even if, as I suspect, there isn't just one answer—if there is even one).
> good luck with research on fixing things like covid, ebola, and polio.
Polio is gone. Ebola, to the extent that it is gone, is because of a massive worldwide effort, not because of any one government—it literally, scientifically, could not have been eradicated by one government. It's not clear to me that the US's record on dealing with COVID is such that it can afford to sneer at anyone else's ability to deal with it, "anarchy enclave" or otherwise.
I suspect that if you broke those companies up you'd only win in the short term, then it would consolidate again. You'd need a constant pressure to prevent this situation and given that the USA would be in the best position to do that I'm not holding by breath.
The internet started decentralized because there was no money to make and we were still figuring things out. As soon as it became mainstream it started to coalesce and we're just now reaching the final stage of this evolution. Even here on HN you have threads full of people coming up with excuses for why they won't run Firefox, and that's a community that should understand the implication of those things. Good luck convincing the internet population at large.
Are there statistics on the browser usage on HN? Just for a good laugh.