Okay so basicially apple users are dumb as rocks which is why iOS is so profitable in the first place, and they are corraled into installing apps and making in-app purchases.

Users are gonna do what's best for them, and if they install apps then it's also what's best for Apple shareholders ;)

Yes. You owe Apple and Patreon as much ad they want to charge, because you are a mental slave

How does your custom AI-built browser challenge the current browser triopoly?

The LLMs are trained on the code of existing browsers. This is essentially a massive process of turning code you don't understand into code you don't understand.

The code details are pretty much all of the details, other than the protocols and standards.

If you understood the codebase of existing browsers (or at least could be confident in making arbitrary changes to existing browsers, perhaps with AI assistance?) then the triopoly wouldn't be threatening because you could just patch out manifestv3 whenever you want.

There is also the problem of people not testing their websites to be compatible with your custom browser. But I would say this is a problem on the protocol level.

someone might build enough hype around it to challenge the oligopoly.

As for training on existing browsers: they are trained on the whole corpus of the human thought process and can take many insights from other fields into the browser, they can write a browser in a completely new language without just transpiring but building it from first principles or as karparthy calls it spec based programming.

I didn't say that they would be successful, just why it's tempting.

Prior to LLMs, creating a browser from scratch seemed like an insurmountable task for a single person. LLMs lower the barrier to entry, and it's a space that is tempting because it would be cool to be the one to create a new browser that people use.

I am not convinced either of these are good test prompts for generic complexity tasks. Many solutions have already been included in the training data!

You can trivially produce a web browser by copying and compiling the code for firefox, no transformer needed.

Can still be a good capability test. Building a car is a real world equivalent. It’s highly complex and has been done billions of times. Still hard to pull off if you ask me.

Famously, one cannot download a car

But that would produce Firefox.

The goal with these tests is to see if the models can make something new, not just copy an existing solution.

That is the goal, at least.

But how do you define, or indeed assess, novelty?

It's not that difficult to take an existing mature codebase and morph it such that it looks quite different but is functionally unchanged.

This is a very different task than building something that's not been built before.

Obviously Microsoft felt bad when they had to kill the old Edge browser that was based on their own HTML rendering engine. Must feel like a second-rate tech company when you can't write some code to render HTML with sufficient quality.

Now they can get back in the game with a 3-word prompt!

And then every time there's some change to web standards, it's just one more prompt where you say "Hey Copilot, take a look at this page that describes the change, and update our browser code to add this!"

/s

it is difficult to wrench someone when you do not know who they are

Someone knows who they are and they can bash different skulls until one of them gives them what they're looking for.

Who is someone?

I mean they have a lot of tools to figure out who you are if they catch you at a rally or something like that. Cameras and facial identification, cell phone location tracking and more. What they also want is the list of people you're coordinating with that aren't there.

Could the chinese construct a sufficient anti-missle defence?

Obviously the actual number of missiles Taiwan has is not public, but I suspect they have enough that reliably intercepting a full barrage is not something even the us could pull off.

The general calculus is that an interceptor costs as much as the missile it intercepts, so defenses are only effective against an adversary with much less resources. Hence Israel can defend against Hamas/Hezbollah, and the US can defend against North Korea, but Israel struggles against Iran and the US doesn't even try to defend against China/Russia.

China obviously has a lot more resources than Taiwan, but then you have a concentration effect where an attacker can focus their resources on a single target, but a much more resourced defender can't necessarily afford to defend that target. We saw that play out with the UK's nuclear deterrent strategy in the cold war, where they focused on overwhelming Moscow's defenses, and were (probably) able to do it despite the USSR being so much bigger.

TW missiles can't "concentrate" because TW geography = all missile flight paths travels through boost, midcourse and terminal interceptors gauntlet along PRC easter theatre command which probably has the densest IADS in the world. PRC has like 3-4x more interceptors in eastern theater (8-12x more total) than TW has missiles. That's just land based, there's also 1000s of naval picket interceptors. Imagine if all of US patriot batteries in Florida, multiplied by 3, then asking what Cuba can do to saturate. Then add in USN DDGs and the answer is realistically nothing, because the industrial math is brutally lopsided. That's assuming TW gets to coordinate salvo their entire inventory, realistically most TELs would be glassed first, every part of TW is withing 5-7 min of PRC missiles and mlrs, less if fired from strait or loitering munitions, i.e. basically faster than abbreviated TEL setup cycle TW has for their tunnel to launch strike complex.

and then the population exploded such that it could only be sustained through modern agricultural methods. We are married to the technology more than before

In what manner do they extract less data

Unless you're running Graphene or a similar security minded distro the sandboxing isn't very good. Okay let's be honest it's fairly abysmal at preventing fingerprinting. It could almost be accused of not even bothering to try.

But one example: https://news.ycombinator.com/item?id=43518866

Even with graphene I don't believe it mitigates much as far as apps collecting data. The idea for more privacy is you run open source apps instead that just don't collect data.

AFAIK Graphene is oriented towards strong device security with privacy as more of a side effect.

One thing with the sandboxed Play Services being that Google has fewer permissions on the device, so presumably they can collect less data.

Which I believe is GrapheneOS' argument when people praise microG: microG being open source does not fundamentally add privacy: apps using microG will phone to Google's servers (that's the whole point of microG). What microG solves is that it removes the Play Services that are root on your device, and it turns out that sandboxed Play Services do that as well.

> The idea for more privacy is you run open source apps instead that just don't collect data.

Yep exactly, I just wanted to add about the sandboxed Play Services, because it was not obvious to me at first :)

> Unless you're running Graphene or a similar security minded distro the sandboxing isn't very good. Okay let's be honest it's fairly abysmal at preventing fingerprinting.

Hmm... the sandboxing is a security feature, it's not there to prevent tracking (not sure what "fingerprinting" includes here). The sandboxing of Android is actually pretty good (a lot better than, say, desktop OSes).

There is pretty much nothing you can do against an app requesting e.g. your location data and sending it to their servers. Fundamentally, the whole goal of apps is that they can technically do that. Then you have to choose apps you trust, and it's easier to trust open source apps.

What GrapheneOS brings in terms of sandboxing is that the Play Services run sandboxed like normal apps. Whereas on Android, the Play Services run with system permissions.

The mobile operating system developed by the enormous ad tech company doesn't try to prevent fingerprinting?! :O

>Unless you're running Graphene or a similar security minded distro the sandboxing isn't very good

Grapheneos doesn't prevent the installed apps fingerprinting you linked either.

Color me surprised. But if you run the app using the sandboxing feature that it provides surely it will only be able to see other apps installed within that same sandbox?

What is "the sandboxing feature" you're talking about? The standard app sandbox built into android allows apps to discover each other for various purposes, and grapheneos doesn't do anything to attempt to plug this.

Apologies. I was thinking of Android user profiles which are available from mainline and (AFAIK) prevent the linked workaround from revealing any apps not installed in the same profile. So it's an example of an unfixed leak in Android but not (as I had previously implied) something that Graphene corrects.

Honestly the state of anti-fingerprinting (app, browser, and otherwise) is fairly abysmal but that's hardly limited to android or even mobile as a whole.

>Apologies. I was thinking of Android user profiles which are available from mainline and (AFAIK) prevent the linked workaround from revealing any apps not installed in the same profile.

But there's no evidence that stock android leaks apps installed across profiles? The link you provided doesn't discuss profiles at all, and stock android also has private space and work profile just like grapheneos.

... yes? That's what I said? Feature available in mainline, motivating leak unfixed, graphene doesn't correct.

no ublock

How can whatsapp be better? Android at least has features like scoped storage.

Where is the ios equivalent of newpipe? Where is the iOS equivalent of pojavlauncher? where is the iOS equivalent of libretorrent or syncthing?

Open source is essentially banned on iOS.

What is the advantage of iOS? "Feels smoother"? Totally subjective.

Safari just got uBlock back!

iOS isn't particularly open source friendly, but mostly people don't do it because of personal incentives, not because it can't be done.

It's subjective, and I get that, but what you miss is that features are subjective too. Missing parity apps are only relevant when you care about that feature; at no point in my life have I ever thought my life would be better or more convenient if I could only torrent on my phone.

But having an app that is responsive and works well has made my life better. Standing outside a bar in the rain trying to get a stupid Covid app to work, not work well, just work, on Android has made my life worse.

(Ironically, I've kind of noticed this is part of the Unix ethos writ small: do one thing and do it well. It's not exact, and iOS for sure has tons of crud everything apps. And they sure don't work together! I just think it's amusing.)

You say that they are ideologically driven when they say browsers are better on Android, and then go on to defend that having LESS features is not necessarily bad. Honestly, you are the one sounding ideologically driven. Having more options is good, specially if there are better options out there (which is the case). Firefox on Android is a better browser than Chrome or whatever, and having the option to use it IS better than not having. You have the right to say that Safari is great, but you cannot say that Gecko on iOS would be worse because, well, you don't have that option.

I'm just gonna put it out there, more choice always being good is the ideology, but when you measure user experience, they consistently rate smooth, fast experiences over feature count unless it's a feature that's important to them.

I don't think iOS is less feature rich except in some specific areas, like web browsers, but you can see in the extreme example that if you could use any web browser for 20 minutes before running out of battery vs safari for hours, one is clearly better. Then you're just haggling over scale. Having the choice to use bad options is not really a choice, unless you have to eg for certain functionality.

And like, in other contexts this isn't even a debate. You talk about the useless feature bloat of Microsoft Word and the associated UI crud, and people are like 'yeah'. But in this context people will straight up make an argument that n+1 features better than n features.

Synctrain is an open source (MPL2.0) iOS Syncthing client (which I made) with full native mobile-first UI and tight iOS integration (shortcuts, background processing, etc).

Are you really expecting me to read this paragraph all by myself? What am I supposed to do, load some text off of a Hacker News comment section? I only read paragraphs written by teams of highly paid experts.

Sent From my iPhone