The problem is that, IMO, centralization of the internet is simply the natural course. Especially when people seem to care so little. Centralization is cheaper, easier to monetize, easier to control, easier to update and upgrade. That's why people prefer Discord to IRC, that's why people prefer Slack to SMTP or Jabber, that's why for many people these days the distinction between git and GitHub is blurry.
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.
> Are there statistics on the browser usage on HN? Just for a good laugh.
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:
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:
I wonder if mobile Safari domination is partly because iOS apps themselves use "Safari" when displaying webpages. If I tap an article using the Octal HN app, it opens in an internal "browser view" that is just an instance of Safari.
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.”
Safari on iOS is seriously the most buggy browser ever. You can mostly write something and it will work on Firefox and Chrome without issues.
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'm not sure where people get this idea. Chrome on my old Note 8 is as fast as Safari on my newer iPad, and doesn't need to reload tabs as I switch between them nearly as often. On top of that, the Safari interface is completely unintuitive ("I have to hold which button to access that? I can't just tap it?").
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.
>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.
It is wholly fair. Developers have treated Android users like second-class citizens from Day 1. Android apps come out months after their iOS counterparts, with hobbled features that never reach parity. Customers have responded with second-class compensation (favoring advertising). Additionally, what you rationalize as "better" is better thought of as a result of expertise lock-in, as developers with more (exclusive?) experience developing for iOS favor that platform. Android is fundamentally more open as a platform, and Android apps would benefit if companies actually bothered to hire and incentivize Android developers as a priority. An especially damning example: the rapid development of ARKit-based technologies after YEARS of Google making its Project Tango resources available. ARCore still lags behind ARKit even though its underlying architecture had BEEN there for developers to explore and iterate on. They simply wrote off the very concept of SLAM-based interaction until Apple said, "Let's do this." It's embarrassing.
> Customers have responded with second-class compensation (favoring advertising).
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?
>You make it sound as though developers have some grand scheme against Android.
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.
That's a really good point. Though on the other hand, any of the browsers on iOS are actually just the mobile Safari engine under the hood anyway, regardless of what the user agent says. To be rigorous, these statistics would need further decomposition.
My FF mobile addons: uBlock Origin, I don't care about cookies, ViolentMonkey (it can sync scripts), ClearURLs (with etag and generic filtering disabled!) and Cookiebro.
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.
Of course it isn't, but we're talking specifically about HN users...these statistics are in response to someone asking about browser share of HN readers, not browser share of the general population.
That’s why we need a regulator. The old AT&T monopoly was dismantled. Based on talking to IBM insiders, IBM actively neutered itself fearing a similar fate in the late 1980s and early 1990s.
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.
Google is both the most dominant it every has been and the most vulnerable it has ever been, at the same time.
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.
GPT-3 does not produce accurate answers to questions in any form. It produces realistic looking text in response to prompts. Sometimes that can look like it is answering questions but that is not the goal nor what it was built to do in any way.
Spin off the major parts of their business into genuinely separate companies - search, cloud compute, advertising, SaaS, Android and Chrome. With Google as a monolith, it's far too easy for them to bundle products and cross-subsidise unprofitable parts of their business to achieve dominance. Those independent companies can make deals that are mutually beneficial, but they'll have to answer to their shareholders and the FTC.
Yeah I'm not even sure how you would break up facebook, other than facebook/instagram . It would make you extremely unpopular if they forced FB to break up it's userbase into thirds or something as they'd all just flock to one third and it would just be the same thing again within a few months
I think a more effective solution than a breakup would be requiring Facebook to interoperate with other social networks. We could force them to implement APIs which would let users from other platforms chat with and send posts to Facebook users. Imagine a tweet or a snap showing up in your Facebook feed, and vice versa. ActivityPub[0] is already a protocol which tries to solve this, but Facebook would never freely implement it because it would breach their walled garden. We might be able to regulate them into doing it.
As other suggested, forcing interoperability would go a long way (it should be noted that e.g. facebook messenger WAS interoperable when Facebook wasn't as large, but they stopped it when they became the leader).
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....
Search could still be funded by search advertising. It just wouldn't be combined with AdWords/DoubleClick/etc. Same idea for other products that have their own ads like GMail and YouTube.
IMO one if the biggest blunders of Moz was not having had the vision to champion the open web by becoming an alternative monitization platform like Patreon. The biggest reason why the open web failed and will continue to fail is not because of aggregators like Google, Facebook, Reddit, and Twitter but rather because there's simply no way for people to vote with their wallets. Moz was uniquely positioned to drive such a change between their market share independence affording them the opportunity to create such a market without drawing immediate regulator scrutiny as all the other browser players would have; their failure to do so will ultimately seal legacy and fate.
I came across the Web Monetization standard a couple months ago when I was looking for alternative ways to fund my website [0], but I couldn't find a lot of information on actual implementations, so I decided to add it myself as an experiment.
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.
Nice to see that there are efforts but until they can roll it out as a core feature with prompts/nudges it's not going to cut it and I'm not entirely sure they have the market/mind share to pull it off these days. Pateron came out in 2013 and has strong community/network effects now, YouTube Red/Spotify also address much the same problem. The current state of affairs is like their idiotic mobile 6 full years after the iPhone came out and at a time when Android/iOS had already won and Microsoft efforts already imploded.
If they rolled out a payment option for third-party sites as a core feature with prompts and nudges, do you really think the reception would be positive? I expect the comment threads would be full of people telling Firefox to work on being the best browser they can, and not getting in people's way with a payment system that "nobody asked for".
I think that would depend a lot on how it's done. If it were simply a mediator/platform that could be opted-into by the serving site kind of like a pateron ad/button I think that could go a long way. The value-add of pateron being in intermediating the transaction and collecting/collating my top interests in a privacy conscious way.
Totally agree. This is why I included the second of the Web's original sins in lack of native payments. We really need cryptocurrencies to succeed here. However I fear that that whole sector has been totally swamped by get rich quick and libertarian wackos with no interest as something as pedestrian as Web Bucks
Putting aside the valid complaint about the noise from get-rich-quick opportunists in this field:
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.
> people prefer Discord to IRC ... prefer Slack to SMTP or Jabber
"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.
>"Centralization wins" is the wrong conclusion from this dataset. Discord & Slack have seen far more investment in UX & backend than IRC & Jabber.
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 $$$.
Meh, I think the longest lasting protocols are the decentralised ones, the good ones just can’t fail by design. SMTP is still going very strong, BitTorrent isn’t going away, bitcoin won’t die anytime soon
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.
Google will not be able to kill smtp, gmail is only one side of the email market. They’d have to offer a transactional and marketing email service and get every app/company to use that over smtp. It’s monumental and I don’t think it will change for decades.
You can damage open/federated email without having to break the lower level layers. See winmail.dat for instance[1].
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.
until we find the right way to do real time chat in a decentralised way
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.
> The problem is that, IMO, centralization of the internet is simply the natural course.
> 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.
I think of it as grease blobs on the surface of your soup. They naturally tend to merge and merge and merge until there is only one. Unfortunately you need active regulation to counteract this.
> 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.
Genuine question, do you mind saying (or pointing to something) what are the implications of browser choice of individuals reading HN?
Assuming HN readers are creating a bunch of products people use every day:
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.
I mean, it’s the same story as Apple and the Mac. Consumers ask their pro/techie friend what they use, and do that. People buy MacBook Pros, people use Chrome. When pros say “I won’t use Firefox because this little feature doesn’t work exactly like Chrome”, their mom doesn’t either.
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.
> The problem is that, IMO, centralization of the internet is simply the natural course
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.
I think this is one of those things that you don't "solve" and it's done. You might be right that things tend to centralize, that's why there is a need for some force to break things apart once the players become too powerful. Maybe it's regulation, maybe it's competition, maybe is people being loud enough (probably not...)
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 think you undervalue what you as a content provider can do with your choices. Sure if you're aiming for millions of clicks (which doesn't bring that much money these days), you can plaster your site with trackers and hostile script. But if you want to reach out to "hackers", that'll only turn away your audience. An alternate scenario to yours could be that Google's web is racing to the bottom and loosing readers (of the kind you care about at least).
This reminds me of many of the issues with laissez-faire capitalism: perfectly competitive markets tend to consolidate, as less efficient participants drop out of the market. Also, network effects. It seems like perhaps economists should engage in more proselytizing in order to reconvince the average person why monopolies are bad.
Centralization is clearly bad in the private sector. I'm curious why people here don't see centralization of government as similarly bad. With 3.5T in revenue, the US federal government is the single largest entity in existence, and it's scope is vast, covering things as disparate as mail and health care. There are plenty of examples of them abusing their position (and occasionally their people...), yet this forum rarely expresses the same level of frustration and outage found on this thread.
I'm supposed to be represented by my government, which we collectively give power to through the constitution. There's no constitution for Google. We don't talk about centralization of government because we figured it out ~250 years ago. We are just figuring out how to curtail the power of internet giants today. There is no inconsistency.
> We don't talk about centralization of government because we figured it out ~250 years ago.
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.
> Eh, the USA is working okay all things considered.
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.
Well please propose a solution. I see all kind of complaints, but never any solutions proposed other than nonsense like "we can form anarchy enclaves!" Enjoy your 2 or 3 years of existence before a much bigger despotic army rolls in and takes you over, good luck with research on fixing things like covid, ebola, and polio. We need to work in the system we have, we need protests and activistic populace who act to the benefit of the whole rather than "I got mine"
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.
Tech companies are a direct product of your government's structure. If your constitution didn't allow lobbyists from megacorps having more power over legislature than the collective mass of its citizens, these companies would not be in the position of monopoly they are in.
It’s clearly bad (and I’m a left wing liberal), I think we need to give states more rights even if it sometimes will hurt causes we like. I also think we need to do another thing that is similar but not the same which is to drive down the stakes of politics. The president needs less power, and the Supreme Court should be more predictable (for example each president nominates two new Supreme Court justices). Spreading out power is a form of decentralization.
Because there is no alternative; how would the state of Texas defend itself from an invasion from China? If we split up the USA into just the states they would all eventually be taken by a foreign power. There is big value in doing things at scale that can't be accomplished in little anarchy/federalist nodes
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.