Firefox uses on-device downloaded-on-demand ML models for privacy-preserving translation.

@firefoxwebdevs How are the models not LLMs, if they are trained on large datasets and generate text?

@Fnordinger https://www.neuralconcept.com/post/ml-vs-llm-key-differences-applications-engineering-impact seems like a good overview

@zzt That would be funny.

But look at the Firefox forks... some had to bring back translation after (mistakenly) disabling it. I don't think any of the local ML API should be suppressed. The discussion should be about shoving LLMs into places where they don't belong.

@firefoxwebdevs

I think the challenge with everything going on here is one of clarity.

@sil, you are asking them about disclosure of models and sourcing. But that is far from the only AI that is in the system.

The tool that does grammar checking and language identification does not leverage an LLM, and while there may be some type of model underneath, the context is very different. Tools that detect spam pages or faulty JavaScript that locks the pages, that's another type of AI hard at work.

Is the browser allowed to support speech to text?

@jmax You're calling out that Firefox may not be able to do this, but I think that mischaracterizes the scope of what's happening here.

The browser has several types of non-deterministic, probabilistic tools in it that provide useful services. Now there's a backlash against one very specific version of those non-deterministic, probabilistic tools. But the backlash is vociferous, often unsolvable, and incredibly broad.

It's hard to engage with non-specific anger.

@liquor_american @firefoxwebdevs shit, I use it all of the time

@firefoxwebdevs @zzt This doesn't feel honest. Maybe from you personally, sure. But not from Mozilla or the Firefox team.

That is like, I decide the car you get. The brand, the model, the color. But hey, don't worry, your voice is important too, so you are allowed to decide what bumper-sticker I will put on your car.

Seriously, this fake inclusion is kinda insulting.

Again, nothing personal against you. But where else should I share my opinion, consider Mozilla even ignores its own feedback platform

#firefox #mozilla

@tasket if you want a serious discussion about the role translations should or shouldn’t have in a browser, let me refer you to steve: https://hci.social/@fasterandworse/115849566354469222

I don’t really feel anything about the translations feature other than disappointment, a bit of concern over how the data was sourced, and a strong feeling that it shouldn’t be a core browser feature

@wes @firefoxwebdevs Sure. But can we agree that it does not represent a core functionality of a web browser?

Like "this meeting could've been an email," but "this feature could've been an add-on."

A web browser should load web pages, allow you to interact with them, and offer add-on support for functionality that doesn't match the definition of "web browser." It's all pretty straight-forward if you're not a marketer, whose brains are all broken.

@firefoxwebdevs Poll is missing a radio button for "fuck you and the horse you rode in on"

@jaffathecake @Fnordinger that really reads like chatbot text. are you *sure* it is not?

@jaffathecake @Fnordinger john? this article appears to be marketing copy written by a company that sells LLMs, quite probably also written by an LLM

what are we doing here man

@firefoxwebdevs

Just give me an easy to find switch that removes _all_ LLM and "AI"-features in Firefox, thank you.

@firefoxwebdevs IMO the killswitch should kill all "Ai" features, but make it easy to re-enable ML-specific features. The main thing I don't want in my browser are LLMs/GenAI, but I know some are against ML too. Maybe there ought to be a checkbox next to the Ai killswitch button to include machine learning features in the killswitch, checked by default - but it can be unchecked before hitting the killswitch if the user desires.

@tedstechtips
100% this ️
@firefoxwebdevs

@davidgerard @Fnordinger hah, I'm not sure. Do you know a better source? I just found one pretty quickly

@zzt @Fnordinger hey Mr Object. I'm sure there's a better source for this. If you know one, let us know.

@firefoxwebdevs What do you mean "open data"? https://firefox-source-docs.mozilla.org/toolkit/components/translations/resources/01_overview.html points to https://browser.mt/ points to https://paracrawl.eu/index.php which says "We do not own any of the text from which these data has been extracted."

@firefoxwebdevs Ask me again later

@firefoxwebdevs hey team, I'm happy that you're doing this, but I think you're caught in a really bad loop.

There is a significant Community backlash against a very specific tool, the LLM. Often just "the chatbot LLM". But that tool has become so ubiquitous, that it has become known as "the AI". For some people, anything that looks like AI might as well be Skynet.

People are spiraling, they're questioning not just LLMs, but anything that looks like it could be non-deterministic probabilistic code.

And look, I know that you run Common Voice, I'm a contributor. But we're not at the point where people are going to question even products that fall out of that project.

Little mini surveys like this are not going to quell any of that backlash.

I think we're at the spot where talking more broadly about where Mozilla uses these non-deterministic tools and how they were sourced is probably a good broad initiative.

@cassidy @firefoxwebdevs The term "AI" has existed since 1956 so of course it's going to have a very broad definition.

Things don't just stop being "AI" when AI researchers invent newer "more AI" stuff.