514
Firefox's beta feature "Smart Window" shared browsing and search history to AI models without prompting
(www.omgubuntu.co.uk)
Ai in browsers is stupid
this post was submitted on 13 Mar 2026
514 points (98.5% liked)
Technology
82581 readers
4472 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
Do yourself a favour and use LibreWolf or WaterFox instead.
Or further favourable: Konform Browser.
Tor Browser and Mullvad Browser also worthy mentions.
The problem with Firefox doing AI is theyre one foot out always. The features they add are always undercooked compared to the rest of the market. This looks really shit and useless in its current state like a worse version of perplexity browser.
All AI is undercooked: Errors are baked into LLMs and there is no viable solution to prevent the mistakes and outright bullshit they produce other than to assume it fucked up and pay an actual expert manually check literally everything it does.
Errors are baked in but I don’t agree with the “no viable solution” part. One research team actually was able to identify the “neurons” responsible for hallucinations and adjust the contribution to negligible amounts.
https://www.youtube.com/watch?v=1ONwQzauqkc (Linking a youtuber instead of the actual study because he summarizes it pretty well and the research itself is not geared for laypersons.)
If this was implemented industry wide would it completely solve the problem? I don’t know, but I do know it would be a massive improvement.
So sad for Firefox. I try to keep using since it’s the only solution free of Chromium, but I guess chromium will control everything only Safari will not be chromium.
Librefox, waterwolf, mullvad, fennec?
What is this, a cutlery drawer?
Hopefully Ladybird will be production ready at some point.
Ladybird is pretty much dead to me, firstly because the dev has some really bad right wing vibes (he said gender inclusive language is too political for his docs and retweeted a Nazi on Twitter). And secondly because the started using AI to move the code to Rust to make it more secure, which is insane if you know anything about AI or security.
My current hope is in servo, because they have much more capable maintainers and the project seems to make some good progress. Also they have daily builds on their website for every operating system, so you can already try it out easily (but don't expect everything to work right now, they still need some time).
I won't use < insert FOSS Project > because of political vibes.
I see this shit almost everyday. Absolute npc behaviour.
Yea I just don't get it for open source software, like you aren't supporting their agenda by using the free fucking software
Ladybird is gonna be awesome, you are overblowing things, it's an open source project chill out. Also sevro is not currently really being designed as a desktop web browser. Its a web engine for light weight embedded applications. Ladybird is currently the only in development webbrowser that is primarily designed with desktop main browser use as its intended use.
And I know you aren't a fan but in the next 5 years pretty much every piece of software you touch will have code from an LLM in it, so get over it or stop using software.
I tried out the servo alpha on android and it is surprisingly usable for something so early in development
I did not know about the AI stuff. However, I do think that the "inclusion" controversy is way overblown. Why in the world would you need to have "gender inclusive language" in the docs for a browser engine?
That ladybird gender thing is such a load of crap. I find it hard to even believe that people are genuinely passionate about it. Every time ladybird is mentioned, someone brings up the 'extreme views' based on this. But it is the biggest load of nothing you are ever likely to see.
For anyone who doesn't know, here's what happened:
- The build documentation for this unreleased pre-beta software used the pronoun 'he'.
- Someone suggested that they change it to be more inclusive.
- The author didn't think it was important enough to change, so left it.
- More requests and pressure came to change the pronoun.
- The pronoun was then changed, and the author apologised.
To me, that's a minor error of judgement, with no lasting harm caused to anyone at all. But yet somehow this is constantly used as a reason to avoid ladybird.
How can I take this seriously? Is this some kind of organised anti-competition propaganda campaign? We're talking about a free and open source project of a highly technical nature, and somehow people are upset that the word 'he' existed temporarily in a work-in-progress document with a target audience of essentially zero people. The people making this project are not political leaders or public figures with media training. They are focused on the technical side of things. Yeah, the pronoun was a mistake, but it pretty much the smallest mistake you could possibly make in this context. It not like they are donating to right-wing orgs, or publicly denouncing anyone, or promoting hate. I see far worse than what they did on a daily basis from all sorts of people - including right here in lemmy. And in terms of ladybird, I have not heard of any kind of misstep ever since this instant - which was a very long time ago now. It is honestly bizarre that people have clung onto this incident. I'm honestly not sure I believe that the backlash is entirely organic. It's just too disproportionate.
[edit] Let me just follow this by saying that I do think there are other good reasons to be upset with this same ladybird dev. I just don't think the 'he' in the docs thing is anything at all.
Could it be posibble the authors first language is gendered, because this sounds like something I would say in Lithuanian like when I talk about computers I use he because the word has a male gender.
He's swedish
Or we could ask the question in the opposite direction - why would you use language which excluded anybody who doesn’t identify as male from the documentation for an open-source project, to the point where when someone offers to update the language for you your response is to rant about “personal politics” and write a contribution policy which forbids the use of gender-neutral language?
Can you provide an example of this non-inclusive language? I honestly can't even come up with an example for web browser documentation that would refer to any gender at all.
"In case of browser crash, the male, white, cis user should submit a bug report"
If the user is female, a report should be submitted by her closest male relative.
load more comments
(3 replies)
🤢 🤮
Much like Adobe‘s Acrobat which I also have to use for work. At least from what I can tell when it suddenly summarizes a PDF. There‘s no way in hell that happens locally. But the fact that it seemingly automatically processes potentially sensitive data from customers didn‘t even do as little as raising eyebrows when I brought it up.
No way in hell? My understanding is that an NPU could perform that type of processing locally. I welcome info & correction!
(I know other types of local ai processors could too, but there’s little chance Acrobat would be geared to look for them - even GPUs - unlike NPUs.)
Now if we switch to talking about policy instead of capability, I don’t think Adobe would miss a chance to be evil. So yeah they’re probably stealing all the data they possibly can.
I use Firefox as a PDF reader at work. it's better than Adobe Reader or Chrome, and it's good enough, so I haven't bothered finding something else.
I deal with secure information sometimes, in PDF form. I haven't even considered that this information might not remain local.
Sumatra is the notepad++ of pdf readers (minus the state sponsored hackers https://notepad-plus-plus.org/news/hijacked-incident-info-update/)
Looks interesting! I can't see it in their docs, can it display the font type and size of a selected text? I've been using PDF-XChange Viewer mostly for that feature. (Lol I just noticed it has been discontinued since 2018 😬)
I've been using it for ages whenever I need to open PDFs on Windows. Though nowadays browsers handle them too. And I avoid Windows.
It is genuinely excellent software.
load more comments
(1 replies)
The speed at which firefox has been enshittified is as impressive as it is sad.
...and we'll all suffer as climate change increases. None of this shit is worth frying for.
Dont worry the US and Israel are starting WW3 as we speak we will all be dead before then any way
Donald Trump is the most environmentalist president ever. With oil prices as they are, can you imagine how much less oil we're burning?
Setting fire to the oil fields, cutting out the middle men. That's American efficiency.
load more comments
(1 replies)
load more comments
(3 replies)
thier revenue entirely dependent on google, you can see why, they got lazy and dint spend time developing thier browser.
Relevant section:
Smart Window uses ‘memories’, things Mozilla says “…it learns from your activity” to inform its responses.
You can delete memories individually, and you can set any given chat session to not use/store them.
Fine so far.
The problem? My memory list isn’t populated with things Smart Window learned since I enabled it. Oh no.
It has activity going back months. We’re talking searches and website interactions from long before I enabled this. features.
Firefox just handed that history to the AI models to plough from, without telling me upfront.
I found this the creepiest aspect of Smart Window.
Mozilla says this was a flub; it will refine the onboarding around Smart Window to limit memory formation to post-opt-in activity only. That’s obviously the right fix.
Because sharing a user’s prior browsing history with third-party AI models, silently, on feature activation, without any headset? Yeah, a bit icky – but that’s the price of testing features that are finished, I guess.
I'm willing to give them a pass since this was a development build and while someone probably should've thought of it, it's the kind of bug that can happen. If this was the public release it would be a lot more outrageous.
Finding out about this gives me some extra questions, though.
- Was this data summarized on enabling this window, or before?
- Did it use an existing model, or re-use one that someone may have already downloaded for a different feature?
- Is this activity going anywhere else, like Mozilla's recent "privacy-preserving" advertising?
- When this does release, what will the default be?
load more comments
(4 replies)
F*** I really dont want to change away from Firefox. Pleas be good Firefox. Please! Don't F this up.
Librewolf is exactly the same browser with all the security features dialed to 11 and all the AI removed.
And it breaks sooo many sites.
Waterfox, is Firefox that just works.
load more comments
(15 replies)
unfortunately forks depend on FF, and chrome to survive.
load more comments
(2 replies)
view more: next ›