There it is. I'd be lying if I said I had my hopes up for a minute, but it irks me to no end that this is being framed as an assault against Google when it's really just Chrome wearing Sam...
OpenAI's browser is built atop Chromium, Google's own open-source browser code, two of the sources said.
There it is. I'd be lying if I said I had my hopes up for a minute, but it irks me to no end that this is being framed as an assault against Google when it's really just Chrome wearing Sam Altman's jacket. Nobody truly makes their own browser anymore. If anyone might have a chance of rapidly starting up a new rendering engine project that could compete, it might be OpenAI... I mean, it's not one of their core competencies or whatever, but they do have a ton of incredible engineering talent and (crucially) the raw AI resources to possibly spin up a vibe coded something that might be interesting.
I'm curious to see what "agentic" stuff they bake into it, but I'll be sticking with Firefox.
Just once I want one of these companies to build on Firefox. I use it every day and it works great. No clue why they'd be so adverse to building on top of it.
Just once I want one of these companies to build on Firefox. I use it every day and it works great. No clue why they'd be so adverse to building on top of it.
I share your view, and maybe this was a rhetorical question. But one practical reason is that Firefox's Gecko rendering engine is licensed under more restrictive terms than Chromium/Blink. The...
Meanwhile if you are OpenAI and build something on Chromium, you're free to fork any component of the codebase and keep all of your modifications closed source. This is more convenient from a process perspective, and doesn't bear the risk of competitors reusing your changes.
That's definitely not the only reason though; I'm not an expert but I think at this point Chromium also has a much larger developer community and presents more of a paved path for customization.
I believe there are also technical reasons. Gecko, on desktop anyway, is relatively tightly integrated with the rest of Firefox where chromium and blink is much more modular. As far as my...
I believe there are also technical reasons. Gecko, on desktop anyway, is relatively tightly integrated with the rest of Firefox where chromium and blink is much more modular. As far as my understanding goes, the Chromium setup makes it relatively easy to have a browser with a completely different GUI but still be able to apply security patches to the core. Where with Firefox that isn't nearly as easy to do.
Mozilla did put in the work to separate gecko in the android ecosystem a few years ago when they rebuilt Firefox for Android. But that work hasn't been done on desktop as far as I know.
Disclaimer, probably overly broad from my side and I haven't kept up recently.
I know you didn’t ask or suggest otherwise, but if you have macOS you can also use Orion browser by the Kagi people, which is Safari based, which allows for extensions (including ublock origin)....
I know you didn’t ask or suggest otherwise, but if you have macOS you can also use Orion browser by the Kagi people, which is Safari based, which allows for extensions (including ublock origin). Works on iOS too.
The second comment echoes my thoughts as well. Only 3 large search indexers exist: Google, Bing, and Yandex. All alternate search engines must use some combination of the above for a usable search...
The second comment echoes my thoughts as well.
Only 3 large search indexers exist: Google, Bing, and Yandex. All alternate search engines must use some combination of the above for a usable search engine with good results.
This puts alternate search engines in a precarious position. Placed in an undesirable negotiating position, any change from these major indexers could massively impact their service. Or worse, potentially destroy it!
About two years ago, Kagi experienced a crisis after Microsoft raised the price of their indexer API by 225% (from 1.25 cents to 2.8 cents per search!). As a result, their once unlimited premium plan was downgraded to a mere 700 searches per month. This stagnated their previously upward trajectory, leading users to unsubscribe, slowing down new sign-ups, and flatlining growth.
During this crisis, Kagi was forced to reduce their dependence on Bing and insulate themselves from unilateral API changes. Like what the other commenter said, if you don't like the entrenched Google/Microsoft duopoly, your only other option is Yandex.
This setback was not easily resolved. It took half a year to find alternatives and reduce their reliance on Microsoft. Only after they felt comfortable reinstating their unlimited search plan, did their trajectory finally improve.
Two months ago, something devastating happened in the search API sphere:
Here's the fundamental problem: had Kagi still relied on Microsoft and not insulated themselves from these unilateral API changes, they would have been completely screwed! Less than a month to find alternatives? That's hardly any time! Microsoft would only negotiate with the big players, leaving small developers in its wake scrambling.
Kagi successfully avoided this business-destroying crisis by having alternatives. Yes, that does also mean using Yandex as a search indexer. But, in the alternate search engine market, that isn't something easily avoided — at least, if you want to avoid the Microsoft/Google duopoly monopoly.
There actually is a real, independent browser in development right now - Ladybird. No idea how long it will be before we see an alpha release, but they have real backing and appear to be both...
There actually is a real, independent browser in development right now - Ladybird. No idea how long it will be before we see an alpha release, but they have real backing and appear to be both passionate and dedicated to the project.
With all these browsers building on Chromium, I fear of the incoming proprietary-fication of Chrome, or worse, "this website works in Google Chrome" (which if you're using Firefox you might...
With all these browsers building on Chromium, I fear of the incoming proprietary-fication of Chrome, or worse, "this website works in Google Chrome" (which if you're using Firefox you might already have seen Google Workspace products limit the use of some features, and on Firefox Android it use legacy Google Search mobile UI).
We saw that in VSCode, where Cursor have used their code to build a competitor. Microsoft's proprietary extensions that used to work with open source builds now strictly enforce their TOS and only work in Microsoft's official builds. Sure it is Microsoft's right to do that, but if you were using VSCodium for privacy you're getting worse experience because Cursor exists.
And then there was Red Hat Enterprise Linux, that they hate what Oracle was doing with their RHEL clone and stopped the "open" source releases and killed CentOS.
And also many new open source software that already have changed their license (or start with newer license) just because otherwise the big clouds will just provide them as a service.
I am convinced that Google's reCAPTCHA is now more aggressive for Firefox users. I can't remember a single occurrence of seeing it and not having to complete a stupid challenge for the last few...
I fear of the incoming proprietary-fication of Chrome, or worse, "this website works in Google Chrome"
I am convinced that Google's reCAPTCHA is now more aggressive for Firefox users. I can't remember a single occurrence of seeing it and not having to complete a stupid challenge for the last few months, or even longer.
Meanwhile, every so often when I see Cloudflare's implementation, I usually don't even have to click the checkbox. Some others I've seen trigger occasionally, but not all of the time. reCAPTCHA is by far the most popular though, so it's hard to fairly compare.
I also see better YouTube performance in Chrome, but that is difficult to measure as well.
I worked at a company who unironically shipped a web app to paying customers that only supported Chrome. They only developed in chrome, tested in chrome and shipped to chrome. Customer using...
I worked at a company who unironically shipped a web app to paying customers that only supported Chrome.
They only developed in chrome, tested in chrome and shipped to chrome.
Customer using Firefox or Edge? Install chrome, you need it to use our app you paid for!
This is definitely coming. I work for a enterprise SaaS company and they recently tried banning literally any browser that wasn't Chrome, including Chromium-based browsers, from being used...
I fear of the incoming proprietary-fication of Chrome, or worse, "this website works in Google Chrome"
This is definitely coming.
I work for a enterprise SaaS company and they recently tried banning literally any browser that wasn't Chrome, including Chromium-based browsers, from being used internally. Immediately engineers complained that this would hurt our customers who use Safari and Firefox but corporate said that our automated testing tools would be good enough to catch any issues in those browsers. Absolutely laughable. The majority of our bug reports for non-Chrome browsers literally come from employees who don't use Chrome so the experience of our products on those browsers would just get worse over time. Thankfully the ban has been indefinitely paused for the moment since our CEO uses Arc but I think the moment Arc stops working and he needs to switch, corporate will be forcing Chrome on everyone.
I can say the other side of that. I worked on enacting that policy two months ago, but it didn't end up going live as the primary reason to do that was deprioritized. In enterprise, Chrome is the...
I can say the other side of that. I worked on enacting that policy two months ago, but it didn't end up going live as the primary reason to do that was deprioritized.
In enterprise, Chrome is the only browser that has "write once, run everywhere" policy enforcement. (Maybe Edge does too, but we aren't Microsoft shop so I didn't find out). You set policy in Google Workspace (doesn't cost money unless it is on ChromeOS) which is then enforced on all desktop and mobile platforms with Chrome. You don't even need a MDM to control some settings - users simply just log into their corp google account to be enforced (you'll need MDM to enforce it on other profiles and incognito).
Compare that to Firefox where you'd need to ship a group policy (we don't have AD), and translate that policy to whatever format Mac, Android and iOS uses by hand without a GUI. Now you can see why enterprises love Chrome.
I use Firefox myself, I was heartbroken but it had to be done in the name of enterprise security.
It sounds like this is a paid Google Workspace/Google Business Profile feature, not a feature of Chrome itself. But it sounds like it really stems from Google's tight integration of Chrome.
It sounds like this is a paid Google Workspace/Google Business Profile feature, not a feature of Chrome itself. But it sounds like it really stems from Google's tight integration of Chrome.
I already have issues copy/pasting into Sheets -- every copy/paste brings up a 'helpful' dialog instructing me how to use command+c and command+v, and the only way to dismiss the dialog is...
I already have issues copy/pasting into Sheets -- every copy/paste brings up a 'helpful' dialog instructing me how to use command+c and command+v, and the only way to dismiss the dialog is clicking X in the top right of the dialog. Seems like the dialog only shows up once in Chrome, but it happens all the time in Firefox. Can't even get Google to care about this when my employer is a paying customer.
I use Kagi to avoid that issue entirely, but there's an extension to fix that on Firefox Android! https://addons.mozilla.org/en-GB/android/addon/google-search-fixer/
and on Firefox Android it use legacy Google Search mobile UI
I use Kagi to avoid that issue entirely, but there's an extension to fix that on Firefox Android!
OpenAI is close to releasing an AI-powered web browser that will challenge Alphabet's (GOOGL.O) market-dominating Google Chrome, three people familiar with the matter told Reuters.
OpenAI's browser is designed to keep some user interactions within a ChatGPT-like native chat interface instead of clicking through to websites, two of the sources said.
A web browser would allow OpenAI to directly integrate its AI agent products such as Operator into the browsing experience, enabling the browser to carry out tasks on behalf of the user, the people said.
Perplexity, which has a popular AI search engine, launched an AI browser, Comet, on Wednesday, capable of performing actions on a user's behalf. Two other AI startups, The Browser Company and Brave, have released AI-powered browsers capable of browsing and summarizing the internet.
Chrome's role in providing user information to help Alphabet target ads more effectively and profitably has proven so successful that the Department of Justice has demanded its divestiture after a U.S. judge last year ruled that the Google parent holds an unlawful monopoly in online search.
This greatly upsets and frankly, terrifies me. "Just trust me bro" but as a search engine, that will actively discourage people from visiting websites. So, so many sites depend on clicks and...
OpenAI's browser is designed to keep some user interactions within a ChatGPT-like native chat interface instead of clicking through to websites, two of the sources said.
This greatly upsets and frankly, terrifies me. "Just trust me bro" but as a search engine, that will actively discourage people from visiting websites. So, so many sites depend on clicks and visits to generate revenue - and this just ... circumvents that with bullshit.
People will trust what the "search engine" tells them, and lacking curiosity will never visit the site of interest. That site will be affected, and may shut down - or worse, may engineer itself to take advantage of ChatGPT and other "AI" engines. So instead of content being engineered for people to read, watch, etc. ... it'll just be made for bots and AI agents to consume.
I have mixed feelings about this seemingly inevitable near-future. For starters, I’ve been a part of the “information wants to be free” crowd with Cory Doctorow, Rick Falkvinge, and Jamie King for...
I have mixed feelings about this seemingly inevitable near-future.
For starters, I’ve been a part of the “information wants to be free” crowd with Cory Doctorow, Rick Falkvinge, and Jamie King for many years. One of the great promises of the early web was its potential to bring the world’s knowledge to our fingertips, without gatekeepers or censors. And from almost the beginning there has been pushback from financialized interests to create things like DRM, the DMCA, data brokers, and paywalls of varying rigidity.
The free and open web was always a revolutionary idea. It upsets economic systems. It disincentivizes consumer capitalism and encourages sharing and collaboration. It’s idealistic. It provides no guaranteed way for contributors to be rewarded for their work. It costs money to run, and when the money is gone, so is the network. It’s a flawed ideal, at least in our current economic order.
But I still believe in it. Information does want to be free. In the LLM age I think that’s more evident than ever. But now we’re going to have to reckon with its darker implications…
Misinformation and disinformation also want to be free. The original vision naively assumed good faith from all participants. But there will always be a contingent of liars and saboteurs pursuing their own diametrical goals. AI amplifies this.
Worker compensation is drying up. Journalism has been on the rocks for ages as it is. Remember AMP? Even here on Tildes most paywalled article links are accompanied by an “archive” link to a mirror. Content’s been widely shared without compensating its creators since the beginning; social media and YouTube reaction videos threw fuel on the fire, not to mention the world of BitTorrent. What I’m saying is, this isn’t a new problem, but content compensation has been fighting a losing battle for ages, and AI is (merely) an accelerant.
Today’s AI is too opaque. We assume that the models are as “aligned” (neutral, benign, honest) as possible, but there’s no certainty in that. We’re at the mercy of OpenAI, Anthropic, Meta, Google, and Perplexity. We have to trust that their closed-source black boxes have our best interests in mind. Meanwhile these companies are making money hand over fist for access to the tech.
So yeah, the situation’s not great. But there’s a hopeful voice in me that says maybe there’s still a path through the quagmire to a world where information can still be free. What would that look like?
Local, high quality, FOSS AI models for everyone. This will require significant hardware advances but should become at least possible as a matter of course.
A way to make content creation rewarding, whether or not people pay creators directly or need to visit their sites to access content. This is a harder nut to crack. We have crowdfunding, which may continue to be viable. I’ve seen some convoluted blockchain-type solutions peddled but nothing I found compelling. IMHO nothing short of UBI is going to facilitate the economic adaptations we need for the AI future. In today’s political climate that’s a tall order. But tomorrow’s? Who knows.
I have some optimism that AI advancement will result in efficiencies that aren’t available today, when it comes to automating content moderation and fact-checking. Just as bad actors are empowered to shitpost slop everywhere, good actors should also be able to catch it and block it near-instantly. It will be a dizzying cat-and-mouse game but I’m hopeful that the days of rampant reality bifurcation might be stymied—or at least slowed—by sensible AI moderation.
I know I'm probably being naive about this. But I also can’t believe the whole internet is going to collapse into a meaningless post-truth void. The way forward will probably be somewhere between the two extremes. We’re probably going to have a bad time, for a while, but then we’ll adapt. We’ll find a way to make this tech serve us, but maybe only after we hit rock bottom first.
I grew up influenced by the “information wants to be free” crowd. But all my life, a glaring contradiction stood out to me: high-quality information is inherently expensive and difficult to...
I grew up influenced by the “information wants to be free” crowd. But all my life, a glaring contradiction stood out to me: high-quality information is inherently expensive and difficult to produce. So all the high-quality information online is subsidized by either organizations or individual volunteers.
The "crowd" involved in the discussion that quote originated from were actually discussing that exact "glaring contradiction". Wikipedia's article on the quote is well-sourced, and its citations...
One idea I've been pondering is the concept of paying some amount of money through your browser on a monthly basis, which is then distributed across the different sites you've visited within that...
• A way to make content creation rewarding, whether or not people pay creators directly or need to visit their sites to access content.
One idea I've been pondering is the concept of paying some amount of money through your browser on a monthly basis, which is then distributed across the different sites you've visited within that month. This is very challenging in practice, as you would have to make some differentiations on sites that have user generated content (paying video uploaders instead of just YouTube, or a blog author instead of Medium, or perhaps a split), as well as address privacy issues (this could be a footprint of sites visited by a single person).
This is already established to some degree within specific sites, though many people are unhappy with the low payouts - Spotify and YouTube premium are examples. I also think Nebula does this, but I am less familiar with that platform.
I’ve been mulling this possibility over more and more these days. I think a way to break out of the clickbait-y and ad driven web is to shift to a “pay-per-play” model similar to your browser...
I’ve been mulling this possibility over more and more these days. I think a way to break out of the clickbait-y and ad driven web is to shift to a “pay-per-play” model similar to your browser wallet concept.
What makes this shift seem more possible today is the move of the EU towards a Digital ID Wallet. Since websites are now requiring age verification (and who knows, potentially identity verification), this wallet idea holds the possibility of being your SSO for the web. Since it can be tied to your personal finances it could in theory be tied to a monthly charge for content.
While I prefer the idea of fixed microtransactions moreso than an arbitrary/discretionary monthly budget, I get nauseous thinking about the current state of crypto getting tied into my Internet access. We already pay to gain Internet access, but we don’t pay for all the externalities that come with that like paying journalists a living wage or the water costs of AI.
If the browser is truly to be a user’s agent, especially including AI agents as some new forks purport as possible (e.g. OpenAI and Dia Browsers), then maybe we need to get serious treating it as a separate entity. People used to pay a travel agent to find and book their trips, should we expect the AI agent of the near future to be at zero cost? Even if locally run, I’m still going to pay for the energy usage at home.
With a monetary incentive (/cost) to producing high quality information, an economic valuation of quality could be the incentive to counteract the proliferation of AI slop.
My fear is that the ignorant will happily pay for misinformation due to the emotional brain-hacking it’s already established, and that the masses will still sell out society collectively through sheer volume alone. Or maybe in the way that luxury goods and fashion are mimicked by cheaper brands accessible to the masses, we might see Truth™ having a certain desirable quality that shuns/shames the conjurors of cheap tricks!
Speaking of Dia, As annoyed as I am with The Browser Co for putting Arc on the back burner, seeing the direction OpenAI and perplexity have gone make me think as a business move it was a very...
Speaking of Dia, As annoyed as I am with The Browser Co for putting Arc on the back burner, seeing the direction OpenAI and perplexity have gone make me think as a business move it was a very smart to shift to an AI centric one. I’ve been meaning to try to beta version out but haven’t had a chance, but based on a comparison of the Perplexity browser demo against Dia it seems like the Browser company is running ahead on design, UX, and features.
This was the idea of a crypto-based standard from the W3C called Interledger. The concept had 3 parts: The website owner would create an address for payments to go to and put it in the meta tags...
This was the idea of a crypto-based standard from the W3C called Interledger. The concept had 3 parts:
The website owner would create an address for payments to go to and put it in the meta tags of the website.
The website's wallet provider would support receiving payments at this address from all types of cryptocurrencies (hence why the project was called "Interledger")
The user would have a browser extension loaded up with some amount of funny money that would occasionally send small payments to the website being viewed.
There are perhaps two interesting things that happened with this system.
Imgur, the image hosting site, used this system for their "Emerald" paid tier.
The only real browser extension to support this was called Coil. You didn't load it up with crypto, but with US dollars. Coil would hide the crypto side of this whole thing away from you, which probably the only reason it saw any usage whatsoever.
But that's about all. Coil is dead. Puma, a chromium fork that also implemented this, has pivoted to AI. If cryptocurrencies were stable enough to actually facilitate transactions and weren't treated as vehicles for wealth accumulation/vacuuming, then maybe there would have been a chance. That didn't happen. Interledger is dead.
I think there's space for the large payment processors (talking Visa/MasterCard/AMEX) to facilitate this kind of payment. They already have infrastructure for a global payment system, but I imagine there'd be a lot of work to scale it for the kind of volume you'd get from this. I certainly don't want to support those companies any more than is strictly necessary, but I think they're really the only ones set up to be able to do a global system. Maybe PayPal, but they certainly wouldn't go for a standards based approach and it would die a slow death.
Thanks for sharing! It sounds like we would both agree that while cryptocurrencies do have some benefits in this space (theoretically no transaction fees that invalidate a $0.05 payment, for...
Thanks for sharing! It sounds like we would both agree that while cryptocurrencies do have some benefits in this space (theoretically no transaction fees that invalidate a $0.05 payment, for example), they are not a good/stable form of payment in general.
Thinking about PayPal & other popular payment platforms, it feels like Venmo and CashApp and Zelle are all leaning in this direction by empowering personal transactions of small amounts. The main difference between these apps and a potential system we're discussing is that you have to know the person you are sending funds to in some way. Even if it's simply scanning a QR code to send a $5 tip, it's executed on an individual basis.
Automating this across however many thousand page loads the average person performs in a month is a lot more complex. But it's interesting to see that this has happened before on the technology side, and maybe just needs more backing from the payments industry to be viable on a larger scale.
I'm not commenting on right or wrong, but maybe the next iteration of this is (paid, high quality) content authors writing for AI agents. Maybe the days of browsing the web will become a blip in...
I'm not commenting on right or wrong, but maybe the next iteration of this is (paid, high quality) content authors writing for AI agents. Maybe the days of browsing the web will become a blip in time.
Setting aside the millions of ways this can be abused, from a technical and naive perspective, society could be served by an AI middle man that is capable of sourcing all of mankind's questions and distributing them to human agents for research and answers. Something highly curated, peer reviewed, and then folded back into AI awareness. Over time, building a trustworthy and high quality source of truth that's traceable back to source material, instead of today's source of "truth" that is based on popular conversations scraped from all kinds of sources.
Using a tool to parse information doesn't seem bad to me. Again, my comment above was knowingly naive and assuming only benevolent actors building such a system. We already have middle men for...
Using a tool to parse information doesn't seem bad to me. Again, my comment above was knowingly naive and assuming only benevolent actors building such a system.
We already have middle men for accessing information. Google, news websites, book authors, printed journals, Twitter. AI could (if done properly, for which I have no hope) remove bias that is inherent to these channels.
So if it's humans providing quality data in, and humans asking human questions and getting high quality answers, I don't see a problem.
I believe you're talking about the algorithms that promote and surface information. LLMs are usually very different. I think we can both agree that an LLM summarizing a book is extremely different...
We already have middle men for accessing information. Google, news websites, book authors, printed journals, Twitter. AI could (if done properly, for which I have no hope) remove bias that is inherent to these channels.
I believe you're talking about the algorithms that promote and surface information.
LLMs are usually very different. I think we can both agree that an LLM summarizing a book is extremely different from a researcher linking to a PDF of the book. If I want to deeply understand the book, the LLM summary is much less helpful than actually reading it for myself.
Well I just finished looking at the blurb page of Comet (Perplexity's whateversauce equivalent). Had no idea why they were showing stylized planets on their site. But ... it made me go down a...
Well I just finished looking at the blurb page of Comet (Perplexity's whateversauce equivalent). Had no idea why they were showing stylized planets on their site. But ... it made me go down a Neptune rabbit hole on Wikipedia. I learned two things:
It takes Neptune 164 years to orbit the sun and as a result, it's been warming/brightening in our telescopes for the last 30-50 years, making better observations possible.
Friggin' no one is going back with an orbiter or anything else. Potentially the earliest thing might be a Chinese orbiter in the 2050s.
I'd like lots more spacecraft to be launched to faraway places ASAP so I can learn more space stuff in my lifetime, damn it. Speed the heck up.
Oh, this has nothing to do with AI browsers other than the rabbit hole I just went down? Imagine how much worse the rabbit holes will be when the goal is to keep us in the Chatbot-browser all day.
There it is. I'd be lying if I said I had my hopes up for a minute, but it irks me to no end that this is being framed as an assault against Google when it's really just Chrome wearing Sam Altman's jacket. Nobody truly makes their own browser anymore. If anyone might have a chance of rapidly starting up a new rendering engine project that could compete, it might be OpenAI... I mean, it's not one of their core competencies or whatever, but they do have a ton of incredible engineering talent and (crucially) the raw AI resources to possibly spin up a vibe coded something that might be interesting.
I'm curious to see what "agentic" stuff they bake into it, but I'll be sticking with Firefox.
Just once I want one of these companies to build on Firefox. I use it every day and it works great. No clue why they'd be so adverse to building on top of it.
I share your view, and maybe this was a rhetorical question. But one practical reason is that Firefox's Gecko rendering engine is licensed under more restrictive terms than Chromium/Blink. The Mozilla license requires that any modifications to source code be made public and available for royalty-free use.
Meanwhile if you are OpenAI and build something on Chromium, you're free to fork any component of the codebase and keep all of your modifications closed source. This is more convenient from a process perspective, and doesn't bear the risk of competitors reusing your changes.
That's definitely not the only reason though; I'm not an expert but I think at this point Chromium also has a much larger developer community and presents more of a paved path for customization.
I believe there are also technical reasons. Gecko, on desktop anyway, is relatively tightly integrated with the rest of Firefox where chromium and blink is much more modular. As far as my understanding goes, the Chromium setup makes it relatively easy to have a browser with a completely different GUI but still be able to apply security patches to the core. Where with Firefox that isn't nearly as easy to do.
Mozilla did put in the work to separate gecko in the android ecosystem a few years ago when they rebuilt Firefox for Android. But that work hasn't been done on desktop as far as I know.
Disclaimer, probably overly broad from my side and I haven't kept up recently.
Gecko also lacks some features like HDR support, but I'll still keep enjoying Firefox since it has uBlock Origin.
I know you didn’t ask or suggest otherwise, but if you have macOS you can also use Orion browser by the Kagi people, which is Safari based, which allows for extensions (including ublock origin). Works on iOS too.
For those interested, Kagi directly funds the Russian conglomerate Yandex.
It's a reason for me to steer clear.
Here are some links people can use to form their own opinion:
The second comment echoes my thoughts as well.
Only 3 large search indexers exist: Google, Bing, and Yandex. All alternate search engines must use some combination of the above for a usable search engine with good results.
This puts alternate search engines in a precarious position. Placed in an undesirable negotiating position, any change from these major indexers could massively impact their service. Or worse, potentially destroy it!
About two years ago, Kagi experienced a crisis after Microsoft raised the price of their indexer API by 225% (from 1.25 cents to 2.8 cents per search!). As a result, their once unlimited premium plan was downgraded to a mere 700 searches per month. This stagnated their previously upward trajectory, leading users to unsubscribe, slowing down new sign-ups, and flatlining growth.
During this crisis, Kagi was forced to reduce their dependence on Bing and insulate themselves from unilateral API changes. Like what the other commenter said, if you don't like the entrenched Google/Microsoft duopoly, your only other option is Yandex.
This setback was not easily resolved. It took half a year to find alternatives and reduce their reliance on Microsoft. Only after they felt comfortable reinstating their unlimited search plan, did their trajectory finally improve.
Two months ago, something devastating happened in the search API sphere:
Here's the fundamental problem: had Kagi still relied on Microsoft and not insulated themselves from these unilateral API changes, they would have been completely screwed! Less than a month to find alternatives? That's hardly any time! Microsoft would only negotiate with the big players, leaving small developers in its wake scrambling.
Kagi successfully avoided this business-destroying crisis by having alternatives. Yes, that does also mean using Yandex as a search indexer. But, in the alternate search engine market, that isn't something easily avoided — at least, if you want to avoid the
Microsoft/Googleduopolymonopoly.Notably, they're also still working on their own search indexes, but it's not easy to beat the big players' decades of data!
https://kagifeedback.org/d/7107-microsoft-bing-retiring-search-api-what-does-that-mean-for-kagi/4
There actually is a real, independent browser in development right now - Ladybird. No idea how long it will be before we see an alpha release, but they have real backing and appear to be both passionate and dedicated to the project.
Ladybird as well as Servo, although that one will likely take even longer.
I didn't know about Servo! Thank you, I like being aware of these development efforts.
The Ladybird team is targeting summer 2026 for their alpha release. Delays are highly likely with a project like this, of course.
Given how security sensitive browsers are, they are the last thing I would want vibe coded.
With all these browsers building on Chromium, I fear of the incoming proprietary-fication of Chrome, or worse, "this website works in Google Chrome" (which if you're using Firefox you might already have seen Google Workspace products limit the use of some features, and on Firefox Android it use legacy Google Search mobile UI).
We saw that in Android, when Amazon started to build the Fire phones Androids are getting more and more close sourced. The default apps you see on the "pure" Android phones today are not the same as the AOSP version. Now Google is also moving to Cathedral model where they'd only do Android source drops instead of developing in the open.
We saw that in VSCode, where Cursor have used their code to build a competitor. Microsoft's proprietary extensions that used to work with open source builds now strictly enforce their TOS and only work in Microsoft's official builds. Sure it is Microsoft's right to do that, but if you were using VSCodium for privacy you're getting worse experience because Cursor exists.
And then there was Red Hat Enterprise Linux, that they hate what Oracle was doing with their RHEL clone and stopped the "open" source releases and killed CentOS.
And also many new open source software that already have changed their license (or start with newer license) just because otherwise the big clouds will just provide them as a service.
I am convinced that Google's reCAPTCHA is now more aggressive for Firefox users. I can't remember a single occurrence of seeing it and not having to complete a stupid challenge for the last few months, or even longer.
Meanwhile, every so often when I see Cloudflare's implementation, I usually don't even have to click the checkbox. Some others I've seen trigger occasionally, but not all of the time. reCAPTCHA is by far the most popular though, so it's hard to fairly compare.
I also see better YouTube performance in Chrome, but that is difficult to measure as well.
I worked at a company who unironically shipped a web app to paying customers that only supported Chrome.
They only developed in chrome, tested in chrome and shipped to chrome.
Customer using Firefox or Edge? Install chrome, you need it to use our app you paid for!
This is definitely coming.
I work for a enterprise SaaS company and they recently tried banning literally any browser that wasn't Chrome, including Chromium-based browsers, from being used internally. Immediately engineers complained that this would hurt our customers who use Safari and Firefox but corporate said that our automated testing tools would be good enough to catch any issues in those browsers. Absolutely laughable. The majority of our bug reports for non-Chrome browsers literally come from employees who don't use Chrome so the experience of our products on those browsers would just get worse over time. Thankfully the ban has been indefinitely paused for the moment since our CEO uses Arc but I think the moment Arc stops working and he needs to switch, corporate will be forcing Chrome on everyone.
I can say the other side of that. I worked on enacting that policy two months ago, but it didn't end up going live as the primary reason to do that was deprioritized.
In enterprise, Chrome is the only browser that has "write once, run everywhere" policy enforcement. (Maybe Edge does too, but we aren't Microsoft shop so I didn't find out). You set policy in Google Workspace (doesn't cost money unless it is on ChromeOS) which is then enforced on all desktop and mobile platforms with Chrome. You don't even need a MDM to control some settings - users simply just log into their corp google account to be enforced (you'll need MDM to enforce it on other profiles and incognito).
Compare that to Firefox where you'd need to ship a group policy (we don't have AD), and translate that policy to whatever format Mac, Android and iOS uses by hand without a GUI. Now you can see why enterprises love Chrome.
I use Firefox myself, I was heartbroken but it had to be done in the name of enterprise security.
It sounds like this is a paid Google Workspace/Google Business Profile feature, not a feature of Chrome itself. But it sounds like it really stems from Google's tight integration of Chrome.
I already have issues copy/pasting into Sheets -- every copy/paste brings up a 'helpful' dialog instructing me how to use command+c and command+v, and the only way to dismiss the dialog is clicking X in the top right of the dialog. Seems like the dialog only shows up once in Chrome, but it happens all the time in Firefox. Can't even get Google to care about this when my employer is a paying customer.
I use Kagi to avoid that issue entirely, but there's an extension to fix that on Firefox Android!
https://addons.mozilla.org/en-GB/android/addon/google-search-fixer/
Relevant excerpts:
This greatly upsets and frankly, terrifies me. "Just trust me bro" but as a search engine, that will actively discourage people from visiting websites. So, so many sites depend on clicks and visits to generate revenue - and this just ... circumvents that with bullshit.
People will trust what the "search engine" tells them, and lacking curiosity will never visit the site of interest. That site will be affected, and may shut down - or worse, may engineer itself to take advantage of ChatGPT and other "AI" engines. So instead of content being engineered for people to read, watch, etc. ... it'll just be made for bots and AI agents to consume.
I have mixed feelings about this seemingly inevitable near-future.
For starters, I’ve been a part of the “information wants to be free” crowd with Cory Doctorow, Rick Falkvinge, and Jamie King for many years. One of the great promises of the early web was its potential to bring the world’s knowledge to our fingertips, without gatekeepers or censors. And from almost the beginning there has been pushback from financialized interests to create things like DRM, the DMCA, data brokers, and paywalls of varying rigidity.
The free and open web was always a revolutionary idea. It upsets economic systems. It disincentivizes consumer capitalism and encourages sharing and collaboration. It’s idealistic. It provides no guaranteed way for contributors to be rewarded for their work. It costs money to run, and when the money is gone, so is the network. It’s a flawed ideal, at least in our current economic order.
But I still believe in it. Information does want to be free. In the LLM age I think that’s more evident than ever. But now we’re going to have to reckon with its darker implications…
Misinformation and disinformation also want to be free. The original vision naively assumed good faith from all participants. But there will always be a contingent of liars and saboteurs pursuing their own diametrical goals. AI amplifies this.
Worker compensation is drying up. Journalism has been on the rocks for ages as it is. Remember AMP? Even here on Tildes most paywalled article links are accompanied by an “archive” link to a mirror. Content’s been widely shared without compensating its creators since the beginning; social media and YouTube reaction videos threw fuel on the fire, not to mention the world of BitTorrent. What I’m saying is, this isn’t a new problem, but content compensation has been fighting a losing battle for ages, and AI is (merely) an accelerant.
Today’s AI is too opaque. We assume that the models are as “aligned” (neutral, benign, honest) as possible, but there’s no certainty in that. We’re at the mercy of OpenAI, Anthropic, Meta, Google, and Perplexity. We have to trust that their closed-source black boxes have our best interests in mind. Meanwhile these companies are making money hand over fist for access to the tech.
So yeah, the situation’s not great. But there’s a hopeful voice in me that says maybe there’s still a path through the quagmire to a world where information can still be free. What would that look like?
I know I'm probably being naive about this. But I also can’t believe the whole internet is going to collapse into a meaningless post-truth void. The way forward will probably be somewhere between the two extremes. We’re probably going to have a bad time, for a while, but then we’ll adapt. We’ll find a way to make this tech serve us, but maybe only after we hit rock bottom first.
I grew up influenced by the “information wants to be free” crowd. But all my life, a glaring contradiction stood out to me: high-quality information is inherently expensive and difficult to produce. So all the high-quality information online is subsidized by either organizations or individual volunteers.
The "crowd" involved in the discussion that quote originated from were actually discussing that exact "glaring contradiction". Wikipedia's article on the quote is well-sourced, and its citations include videos of the original discussion.
One idea I've been pondering is the concept of paying some amount of money through your browser on a monthly basis, which is then distributed across the different sites you've visited within that month. This is very challenging in practice, as you would have to make some differentiations on sites that have user generated content (paying video uploaders instead of just YouTube, or a blog author instead of Medium, or perhaps a split), as well as address privacy issues (this could be a footprint of sites visited by a single person).
This is already established to some degree within specific sites, though many people are unhappy with the low payouts - Spotify and YouTube premium are examples. I also think Nebula does this, but I am less familiar with that platform.
I’ve been mulling this possibility over more and more these days. I think a way to break out of the clickbait-y and ad driven web is to shift to a “pay-per-play” model similar to your browser wallet concept.
What makes this shift seem more possible today is the move of the EU towards a Digital ID Wallet. Since websites are now requiring age verification (and who knows, potentially identity verification), this wallet idea holds the possibility of being your SSO for the web. Since it can be tied to your personal finances it could in theory be tied to a monthly charge for content.
While I prefer the idea of fixed microtransactions moreso than an arbitrary/discretionary monthly budget, I get nauseous thinking about the current state of crypto getting tied into my Internet access. We already pay to gain Internet access, but we don’t pay for all the externalities that come with that like paying journalists a living wage or the water costs of AI.
If the browser is truly to be a user’s agent, especially including AI agents as some new forks purport as possible (e.g. OpenAI and Dia Browsers), then maybe we need to get serious treating it as a separate entity. People used to pay a travel agent to find and book their trips, should we expect the AI agent of the near future to be at zero cost? Even if locally run, I’m still going to pay for the energy usage at home.
With a monetary incentive (/cost) to producing high quality information, an economic valuation of quality could be the incentive to counteract the proliferation of AI slop.
My fear is that the ignorant will happily pay for misinformation due to the emotional brain-hacking it’s already established, and that the masses will still sell out society collectively through sheer volume alone. Or maybe in the way that luxury goods and fashion are mimicked by cheaper brands accessible to the masses, we might see Truth™ having a certain desirable quality that shuns/shames the conjurors of cheap tricks!
Speaking of Dia, As annoyed as I am with The Browser Co for putting Arc on the back burner, seeing the direction OpenAI and perplexity have gone make me think as a business move it was a very smart to shift to an AI centric one. I’ve been meaning to try to beta version out but haven’t had a chance, but based on a comparison of the Perplexity browser demo against Dia it seems like the Browser company is running ahead on design, UX, and features.
This was the idea of a crypto-based standard from the W3C called Interledger. The concept had 3 parts:
There are perhaps two interesting things that happened with this system.
But that's about all. Coil is dead. Puma, a chromium fork that also implemented this, has pivoted to AI. If cryptocurrencies were stable enough to actually facilitate transactions and weren't treated as vehicles for wealth accumulation/vacuuming, then maybe there would have been a chance. That didn't happen. Interledger is dead.
I think there's space for the large payment processors (talking Visa/MasterCard/AMEX) to facilitate this kind of payment. They already have infrastructure for a global payment system, but I imagine there'd be a lot of work to scale it for the kind of volume you'd get from this. I certainly don't want to support those companies any more than is strictly necessary, but I think they're really the only ones set up to be able to do a global system. Maybe PayPal, but they certainly wouldn't go for a standards based approach and it would die a slow death.
Thanks for sharing! It sounds like we would both agree that while cryptocurrencies do have some benefits in this space (theoretically no transaction fees that invalidate a $0.05 payment, for example), they are not a good/stable form of payment in general.
Thinking about PayPal & other popular payment platforms, it feels like Venmo and CashApp and Zelle are all leaning in this direction by empowering personal transactions of small amounts. The main difference between these apps and a potential system we're discussing is that you have to know the person you are sending funds to in some way. Even if it's simply scanning a QR code to send a $5 tip, it's executed on an individual basis.
Automating this across however many thousand page loads the average person performs in a month is a lot more complex. But it's interesting to see that this has happened before on the technology side, and maybe just needs more backing from the payments industry to be viable on a larger scale.
I'm not commenting on right or wrong, but maybe the next iteration of this is (paid, high quality) content authors writing for AI agents. Maybe the days of browsing the web will become a blip in time.
Setting aside the millions of ways this can be abused, from a technical and naive perspective, society could be served by an AI middle man that is capable of sourcing all of mankind's questions and distributing them to human agents for research and answers. Something highly curated, peer reviewed, and then folded back into AI awareness. Over time, building a trustworthy and high quality source of truth that's traceable back to source material, instead of today's source of "truth" that is based on popular conversations scraped from all kinds of sources.
That sounds awful. I don't think humanity needs a middle man for access to information.
Using a tool to parse information doesn't seem bad to me. Again, my comment above was knowingly naive and assuming only benevolent actors building such a system.
We already have middle men for accessing information. Google, news websites, book authors, printed journals, Twitter. AI could (if done properly, for which I have no hope) remove bias that is inherent to these channels.
So if it's humans providing quality data in, and humans asking human questions and getting high quality answers, I don't see a problem.
This sounds to me like you've inadvertently invented an AI that only uses Wikipedia as a source. Which, might not be terrible?
I believe you're talking about the algorithms that promote and surface information.
LLMs are usually very different. I think we can both agree that an LLM summarizing a book is extremely different from a researcher linking to a PDF of the book. If I want to deeply understand the book, the LLM summary is much less helpful than actually reading it for myself.
Bad day for The Browser Company. Shoulda kept working on Arc
Well I just finished looking at the blurb page of Comet (Perplexity's whateversauce equivalent). Had no idea why they were showing stylized planets on their site. But ... it made me go down a Neptune rabbit hole on Wikipedia. I learned two things:
I'd like lots more spacecraft to be launched to faraway places ASAP so I can learn more space stuff in my lifetime, damn it. Speed the heck up.
Oh, this has nothing to do with AI browsers other than the rabbit hole I just went down? Imagine how much worse the rabbit holes will be when the goal is to keep us in the Chatbot-browser all day.
[noise]