Firefox 148 Now Available With The New AI Controls, AI Kill Switches 71
Firefox 148 introduces granular AI controls and a global "AI kill switch" that allows users to disable or selectively manage the browser's AI features. Phoronix reports: Among the AI features that can be toggled individually are around translations, image alt text in the Firefox PDF viewer, tab group suggestions, key points in link previews, and AI chatbot providers in the sidebar. Firefox 148 also brings Firefox for Android, support for the Trusted Types API, CSS shape() function support, Sanitizer API support, WebGPU enhancements, and a variety of other changes. Developer chances can be found at developer.mozilla.org. Binaries are available from ftp.mozilla.org.
Well, there is a positive way to consider this. (Score:5, Insightful)
Firefox appears to have gotten some dev money from the "AI" war chest, which, if the kill switch indeed kills this anti-feature is, I guess, acceptable.
We replaced the google search box successfully, we'll survive the "AI" box as well.
Re:Well, there is a positive way to consider this. (Score:4, Insightful)
Re: Well, there is a positive way to consider this (Score:2)
I voted against it as well.
Sadly, a community that donates small amounts isn't as attractive as big cash dumps, and most of the ff community isn't donating all that often.
Re: (Score:3)
Re: Well, there is a positive way to consider this (Score:5, Insightful)
It's not about your politics, it's about basic privacy and functionality issues. A commonality among those that don't want to be the product, is finding ways that prohibit being involuntarily monetized.
The AI rubric serves only the tech bro fortunes, and not those of the individual. It appears on the surface to be of value, but AI inconsistency, and the addictive quality of short cuts then strangles users and their needs.
This isn't about anybody's lockstep. It's about values and liberty, and human worth as opposed to shipping one's value to somebody for their kleptocracy purposes.
Re: Well, there is a positive way to consider thi (Score:2)
There doesn't really seem to be a shortage of vocal right wing opinion on /.
It's part of what keeps the site decent. Generally intelligent debate on issues from multiple perspectives.
I'd say the comments lean left, but not extremely so. And there's usually intelligent opposition in any thread.
Re: Well, there is a positive way to consider thi (Score:2)
Even if every user donated something we would be hard pressed to have a voice compared to Google. And that's even assuming we all wanted the same thing.
Cover the costs ONLY for WANTED changes (Score:1)
Can't tell if you're joking, but there is a way to do it. At least I have a theory of a solution.
What if you could donate specifically for a change you wanted? If enough people agreed with you, then they'd collect the money and make the change. If too few people agreed with the change, then it would never happen.
As it applies to this story, Firefox wouldn't be in the position of trying to limit and undo the impact of a change that would never have gotten enough support to be implemented in the first place.
Re: (Score:2)
I'm in favor of your plan, but it is fundamentally incompatible with the Mozilla foundation's Google-related income.
Re: (Score:1)
Concurrence, but the basic situation is that every organization and company tends to wrap itself around it's primary income sources. That's why I think a "real" charity model focused on just recovering the costs might help.
For a while, anyway. The gamesters always find new ways to break the rules of every game. Progressive taxation was another interesting example of a change that had some positive results for a while... (Reacting to a economic philosophy book.) I remember seeing a t-shirt that said "Time he
Re: (Score:2)
Mozilla seems to be obsessed with getting new users in with gimmicks like AI. Stuff that is far from ready or reliable. What the current users want doesn't seem to be of much concern to them.
Or maybe they have just given up on the basics. Compatibility when the browser landscape is basically a Chromium monoculture is hard. Speed is just hard full stop, and while to be fair they have made some good progress, they still have some issues.
Re: (Score:2)
They just suffer from fomo, or actually, just joining the latest fad.
Re: (Score:2)
The closest I can find is one conducted in 2019/2020 surveying how people feel about "AI" in general. Suffice to say, that survey's results are nothing close to what you are indicating.
I then searched for any articles referring to Mozilla AI assertions, looking for any evidence of a "backlash", or some primary source indicating there even was one.
One cited an article containing the following: "As a member of the open-source/Linux community who has been speaking
Re: (Score:2)
Re: (Score:1, Troll)
Firefox does built in local translations of pages. That's an "AI feature" and it's avoiding you having to use google translate or other systems that end up sending the data "to the cloud" (in other words, sending you private data to other people's servers where they can do whatever they want with it).
I don't see how local, relatively low energy, private, reasonably effective translation of a page is an "anti-feature".
Re: (Score:2)
Example: it's an anti-feature if I don't want bloat and want proper high end translations by a much larger model instead. I.e. google translate, or chatGPT.
Re: (Score:2)
Re:Well, there is a positive way to consider this. (Score:4, Insightful)
For those that want in-browser translation, it's no problem to have this as an available option. However, that doesn't mean it must be integrated into the browser, distributed to all, and enabled by default. What is the argument against having it be an add-on?
Or maybe the "translate this page" feature could be as it had been before integrating a local AI translation feature - fork that out to an external service - but provide configuration for the "translate this page" akin to the search engine config, along with an option to pull down an add-on for a local AI? I note this option because I see the rationality of adding the relatively light weight and simple "translate this page" when it's just sending the data out to some other service for the heavy lifting... why not facilitate that for those times when a user wants to actively choose to do that by clicking on that option? But why jump to including a local translation AI with the browser?!?!?
Also, maybe I already run a local LLM (I do). And maybe it's already more capable than the one they're shipping (it is). Then maybe it makes even less sense for the browser to bundle one in with it?
Long story short, most people don't want this feature creep, and those that want a translation feature AND want to use a local LLM are very very very likely to be fully capable of adding an add-on or configuring it to use their existing local LLM.
My guess... it's probably a lot easier (less friction, and less documentation/setup/support needed) to build it in when you're in that position than it is to do the same in an add-on.
Re: (Score:2)
I'm kind of with you on that, but building features as optional is a bunch of extra work. For me, although I would like what you ask for, it's good enough that I can turn it on or off. If you disagree and see this a major priority then I'm sure there are many like myself that would applaud your contribution of the code for it.
I guess it will start built in, a new framework will be built and in the long run we'll all be able to choose whatever AI components we prefer.
Re: (Score:2)
I get why it happens this way, but it's backwards IMO.
I guess it will start built in, a new framework will be built and in the long run we'll all be able to choose whatever AI components we prefer.
It's like test-first development. Most people will get behind the idea in theory, but they wind up developing before making the tests most of the time. I think this hits a similar mark.
IMO, all such features should start as add-ons, period. Got some new idea? Great. Can it be done as an add-on today (IE: does the add-on API have sufficient features to implement it)? If yes, then directly go to creating the add-on. If no, then the work to do on the codeba
Re: (Score:2)
rather than have people running to third party forks.
Lots of stuff I kind of agree with. Software development is generally awful and people just don't care about doing it right even when it costs them hugely.
However, there's one thing. We desperately need third party forks. Zen browser, for example, is a really important concept because its an open fork which scratches someone's itch. I hope eventually that Firefox really comes to understand that importance. It's what made webkit so important.
Unfortunately, firefox is under a bad license which doesn't fully s
Re: (Score:2)
However, there's one thing. We desperately need third party forks.
100% agree. It's a terrific way (as a society?) to explore different implementations and features. In this context, Mozilla could have done a full on AI/LLM browser fork and littered it with integrations. If more people started using the fork, then they'd know where to dump more resources.
Unfortunately, firefox is under a bad license which doesn't fully support collboration. The reason that there are so many webkit derived browsers which come together under the one HTML renderer is because the KDE project chose the LGPL instead of a weaker license like the Mozilla license.
IANAL, but I am a bit of a licensing nerd. While MPL vs LGPL may have played some role here, I think you're overstating the impact. KHTML / Webkit initially took off because of Safari. AFAICT, the MPL probably would have s
Re: (Score:2)
Firefox does built in local translations of pages. That's an "AI feature" and it's avoiding you having to use google translate or other systems that end up sending the data "to the cloud" (in other words, sending you private data to other people's servers where they can do whatever they want with it).
I don't see how local, relatively low energy, private, reasonably effective translation of a page is an "anti-feature".
But do you trust that Firefox is not sending the data to a third party to train the AI?
Re: (Score:2)
But do you trust that Firefox is not sending the data to a third party to train the AI?
I trust that the firefox code is out there. The build process is open and that, if the were doing that, we could and would discover it. This is not some obscure NPM package used by 20 people.
Re: (Score:2)
Re: (Score:3, Interesting)
Re:Well, there is a positive way to consider this. (Score:5, Informative)
Telemetry doesn't always mean spying, especially when done by a non-profit in a responsible anonymised way.
Telemetry is extremely useful in many cases, and it's the reason most websites implement analytics of some sort.
In Firefox's case, telemetry can reveal that many people don't notice a feature or rarely use it, so it doesn't need to be as prominent. It can also show that many users visit `about:config` to change settings, which suggests the org should expose those options in a more user-friendly way instead of burying them.
Telemetry can also reveal widespread use of an add-on, which may justify building that functionality into the browser. Examples include tracking protection, dev tools, tab controls, and vertical tabs, etc.
Regardless, Firefox asks on first install and in new profiles whether to enable technical data collection.
Re: (Score:2)
Telemetry can reveal a whole lot of stuff, but none of that matters if the developer doesn't care.
There's nothing wrong with simply listening to community feedback, but why do that when you can forcibly collect data and use it to construct pie charts for the pointy-haired bosses?
Besides, a long time ago I used to work in a warehouse where our productivity was monitored by a telemetry system. We all knew the numbers were "adjusted" by the tech team after every shift -- they even told us so right to our face
Re: (Score:2)
Telemetry doesn't always mean spying, especially when done by a non-profit in a responsible anonymised (sic) way. ...
Telemetry can also reveal widespread use of an add-on, which may justify building that functionality into the browser. Examples include tracking protection, dev tools, tab controls, and vertical tabs, etc.
I would love to believe this; however, I have never seen any improvements come from telemetry; only negatives. "Only a million people change their task bar position, so we will remove the feature". Meh. I don't want to participate.
Re: (Score:3)
I have never seen any improvements come from telemetry
That could be because many technologists disable telemetry out of fear of spying, and so what we need doesn't really get exposed to mozilla / other non-profit projects. Which would also explain why the interface keeps getting dumbed down and oversimplified.
Regardless, it's a simple toggle under: Settings -> Privacy -> Data collection.
Unlike Google's aggressive and extensive data collection via countless avenues which we mostly can't escape (recaptcha, analytics, doubleclick, googlesyndication, tagmana
Re: (Score:2)
As you demonstrated in your post, telemetry's main purpose is to advance what you already want to advance. Your add-on example is a good example. There's no reason to take a widespread add-on and re-develop it as new built-in functionality. It's already working for tons of people as an add-on. Making a new version means you've now split that market, you probably introduced bugs, your implementation will be slightly different, and the features may conflict with each other when the browser automatically u
Re: (Score:2)
And as a bonus, in the EU this is legally required to be off by default as it potentially processes personal data may sent it to the cloud.
ok (Score:1)
Developer chances can be found at developer.mozilla.org.
There's a chance Slashdot "editors" are incompetent and don't deserve employment in this job market, and that chance is 100%.
Re: (Score:2)
Imagine having the ability to post anonymously in 2026? It can only lead to real people getting pissed off and trolled.
That's 100% why they allow anonymous posting. They turned it off and found out that ad impressions decreased because of the lack of pissed off posting, so they turned it back on.
Re: (Score:2)
They turned off anonymous posting because it was non stop nazi ascii artwork.
Re: (Score:2)
They turned off anonymous posting because it was non stop nazi ascii artwork.
I did not speak to why they turned it off, I spoke to why they turned it on again.
Maybe you missed it when they actually disabled all anonymous posts by any users, logged in or not? It was glorious. Most anon posts are trolling by far.
Re: (Score:2)
It's part of
Except for silvergun wasting pixels on novella sized posts that put you to sleep by the 3rd paragraph and Nazee Asci Art, anon is a quirky and valuable part of
Re: (Score:2)
Except for silvergun wasting pixels on novella sized posts
Most of those are AI-generated, ostensibly not by him, but by whoever keeps advertising their "fridae" site.
Re: (Score:2)
Operating system level (Score:1)
Every operating system should come with a global AI kill switch that requires apps to not show AI features, alerts, dialogs, nor run any AI at service or application level.
Re: (Score:2)
Re: Operating system level (Score:2)
Re: (Score:3)
I think it does - even Windows 11 seems to do it.
But Windows Defender freaks out because modifying the hosts file is a super common way for malware to do hijackings as well as being persistent - they could hook Google.com so it reinstalls when you visit Google. Just like another way is to redirect your DNS settings.
Of course, it's a lot more complex nowadays since man
Re: (Score:2)
I think I've seen where it is possible, but difficult, to get win 11 to respect hosts.
Re: (Score:2)
The exception being Windows Update, which for a long time has ignored the Hosts file because it would make it too easy to block updates. That includes malware definition updates for Windows Defender.
Re: (Score:2)
How does the hosts file prevent a bundled LLM from executing within a program? (it doesn't)
Re: (Score:2)
I'm working on one right now. It will require AI to determine how to kill every AI component.
The browser wars are heating up again. (Score:2)
Re: (Score:2)
The whole battle for "human-based" browsing vs "ai-assisted" browsing is also causing (additional diversity) in the internet's culture.
FTFY
Re: (Score:3)
Firefox is the only browser that supports the full uBlock Origin plugin now. Any Chrome-based browser, including Chromium can no longer run it thanks to plugin manifest v3 that Google crammed down everyone's throats. No, supporting uBlock Lite is not good enough. Even Vivaldi, which has its own, much-less-effective ad blocker, does not support even uBlock Origin Lite. So yeah really Firefox, along with its forks and derivatives, is the only one left that lets you decide what you want to see. So for all
Feature request for new button (Score:2)
Will the AI be off by default? (Score:1)
Keybindings (Score:2)
You can now finally change the keybindings. Better late then never I guess.
Wait.. no it was just the wind (Score:2)
So who here asked for AI features? Crickets or just the wind.
Re: "Developer chances" (Score:1)
Yes, it is in normal Release Notes. First entry.
https://www.firefox.com/en-US/... [firefox.com]