Some thoughts on Mozilla's trajectory and Waterfox's stance on AI in the browser 👇

[soaproot@sfba.social]

@miff @Waterfox Agree, I'm looking into flatpak sandboxing now, it would appear to be the reason (I can access some file URLs, just not this particular directory).

[kohoso@mastodon.social]

@Waterfox

[grahamperrin@mastodon.bsd.cafe]

@Waterfox honestly, I fail to see how it undermines trust, transparency, and user agency.

How, exactly?

Cc @skinnylatte @jbz

[wagesj45@mastodon.jordanwages.com]

@Longplay_Games @grahamperrin @Waterfox @skinnylatte @jbz Yeah I think the issue Waterfox is concerned about is the mechanistic interpretability of the model, not the intention of the developer/browser/prompt. They're fairly unpredictable almost by design. Not something you want in a browser, design-wise.

At least not at baseline. They do have information on how to add that kind of functionality in their documentation.

https://www.waterfox.com/support/ai-chatbot/

[cf@todon.eu]

@Waterfox Hi, We're currently looking around for a new webbrowser on account of all the "AI" stuff, this statement looks promising, though, if We may, We'd like to hear some more direct statements about where Waterfox as a project stands. In decreasing order of relevance:

What are the politics of Waterfox? In general, but in particular regarding marginalized folks such as black and/or queer individuals? What critiques does the project usually get about its politics and what's the stance on those?

The blogpost focuses on the usage of "AI"/LLM in the context of webbrowser usage, but what is the stance on the technology of "generative" "AI" as a whole? In particular relating to the financial bubble around it, the slave labor necessary to create it, the poisoning of the web, climate destabilization and energy grid degradation that it causes? What about locally ran models?

The blogpost mentions that "LLMs have utility, measurably so" - Which is that?

The Waterfox repository has a folder for the Zed editor, of which a key feature is the integration of LLMs. Is "AI"/LLM produced code used in the Waterfox project? Are "AI"/LLM produced code contributions accepted?

The Waterfox.com 404 page includes the following image: https://www.waterfox.com/_astro/astronaut-nobg.CW-JfQlf_Z125kPS.png - Is this made by an artist, and, if so, whom?

[nickkarpowicz@mastodon.social]

@Waterfox I just downloaded Waterfox and am trying it out because of this statement. Looks good so far :)

[sodiboo@gaysex.cloud]

@Waterfox@mastodon.social thank you.

[gncardoso@ursal.zone]

@Waterfox Best Firefox alternative browser.

[ppw@social.redflag.ps]

@msokiovt @Waterfox there is nothing wrong with having AI features in your browser, but it should be up to the user to decide whether and how this happens, rather than Firefox simply imposing it on you and then not even giving you the option to disable it.

[waterfox@mastodon.social]

@ulfi where are you seeing that?

[potassium@mindly.social]

@Waterfox well put!

[ef@mastodon.bsd.cafe]

@Waterfox can you do a musl build as well as the glibc one?

[scrottie@bsd.network]

@Waterfox "Their lunch is being eaten by AI browsers." -- Chrome was dominating before Google started to add "AI" features to it. I'm sure there's an example out there, but I don't think I've once heard someone say they prefer Chrome over Firefox because of "AI" built in to Chrome. But I often hear people say they use Chrome because it works (noting that Google abused their monopoly position when repeatedly sabotaging their sites for Firefox). It seems like there was an uptick in Interest in Firefox after Google banned ad blockers, but "AI" in Firefox and tone deaf governance from Mozilla has more than counteracted that effect. So I'm not sure I agree with your premise there.
I often see smaller companies trying to do the same thing as larger companies, by making an assumption that the large company knows what people want, but this is often a kind of "cargo culting" effecting. Large companies very often don't do things because people want them (was just reading an article about Paraquat in the US, then there's rBST and everything Monsanto has ever done). Large companies often succeed exactly because they can do profoundly bad things and survive while their smaller competition destroys themselves trying to play follow-the-leader. Microsoft can add Clippy (Clipit) to Microsoft Word that pops up and refuses to go away without a complex, non-intuitive sequence of clicks, that is nearly universally despised, because we ascribe a "no other choice, have to deal with it" superpower to Microsoft (as with Monsanto), but it would be suicide for eg LibreOffice to attempt the same abuse of power (or for independent dairies selling at farmer's market to proudly advertise rBST in their milk). They do not have the power they can abuse. Avoiding getting killed in games of "follow the leader" led by tech giants requires understanding the market and not assuming tech giants are providing what people want. It requires understanding why giant corporations are doing things. That's a whole other discussion, but I see this as more like billionaires buying up the news and social media to intentionally wreck them than in any way providing people what they want.

[cf@todon.eu]

@Waterfox It's been almost a month since We asked, do you wish to put in any words about these topics or are We to assume that besides the very specific application of LLMs piloting a webbrowser the project is uncritical of the usage of "AI" and has LLM "generated" code?

#AI #NoAI #LLM

[tiotasram@kolektiva.social]

@Waterfox this is good. I moved to Waterfox because of Firefox' position on AI, and your stance here us the correct one which I believe will stand the test of time.