Where Waterfox stands on AI and LLMs: version 1 compared withcurrent version
Unified text comparison: − removed,+ added. Formatting and non-text content may not appear.
Removed: Back in December, I wrote No AI* Here. I talked about how I don't believe LLMs have their place in a browser such as Waterfox, due to the inherent privacy risk of sending your browsing data out or having it processed by third parties. Even if you were to use a local LLM, there can still be a privacy risk, unless you develop it yourself from scratch. My position will likely never change for LLM usage in the browser.
Removed: But since then, some users have taken it upon themselves to extrapolate what Waterfox is into more than a privacy focused, customisable browser, and instead treat it as anti-AI in every single aspect. I don't even know where to start with unpacking this, considering the "AI" field in computer science stretches back many decades and contains a vast array of technologies, many of which are inescapable.
Removed: Cutting off your nose to spite your face
Removed: A few days ago, I raised a warning flag on the privacy policy of Mozilla's collaboration with Mistral. I finalised the series of posts with this:
Removed: This isn't anti-AI just for the sake of it either - LLMs are useful tools. However, even useful tools need boundaries.
Removed: This seemed to cause quite a stir, because a lot of people don't seem to understand that AI isn't just LLMs. In fact, some users now demand an impossible puritanical standard: that their software must contain NO AI in its entirety. Any technologist who understands computer technology beyond just the surface knows how unrealistic this is. That includes demands to drop so called AI commits (for which I have no idea what these are, I can only assume they mean upstream commits from Mozilla) completely.
Removed: Maybe there was an issue with the wording "LLMs have their uses" - of course, pitchforks at the ready, it seems these users want a web browser that has had all LLM usage scrubbed from it, something that is impossible. Every single major browser engine in existence will have had LLM involvement in some form. In fact, Chrome has had multiple zero days this year, many discovered and assisted by LLMs.
Removed: ffmpeg had 21 zero days discovered by LLMs, Firefox itself had over 423 security bugs discovered by LLMs and patched by over 100 volunteers.
Removed: It is not just browsers, or media tools. Debian voted to allow responsible use of generative AI - Debian neither endorses nor prohibits such tools, but recognises they can improve productivity when the contributor remains responsible for quality and legal compliance. The Linux kernel itself documents how to use AI coding assistants and curl is patched with AI help on an ongoing basis.
Removed: The sheer fact that people want to actively have these fixes dropped because of LLM involvement is the very definition of cutting off your nose to spite your face. And what's worse is the attitude and vitriol that accompanies it, as if logic no longer has a place and only emotions matter because understanding that the world is more than black and white is too complicated.
Removed: Pandora's jar has been opened, and love them or hate them, LLMs do provide value in security research when used properly. It is just another tool.
Removed: This does not mean I support the wanton environmental destruction, the ethical concerns surrounding the development of LLMs, nor the fact that generative AI music and art often amounts to theft.
Removed: So to the users who so proudly label themselves "anti-AI" in all its forms, I have to ask - do you use any technology at all? Because unless you've sworn off modern computing entirely, you already rely on decades of AI research every day, none of which is an LLM:
Removed: Spam, phishing and fraud filtering - email spam filters, SMS spam detection, your bank flagging a suspicious card payment
Removed: What you type - autocorrect, spellcheck, swipe typing, predictive keyboards
Removed: Search - ranking, autocomplete, "did you mean?", voice search
Removed: Photos and cameras - Face ID unlock, face grouping in your photo library, Night Mode, HDR, Portrait Mode, background blur
Removed: Voice and calls - Dictation, live captions, noise suppression and echo cancellation in Zoom, Meet and AirPods
Removed: Understanding documents - OCR for check deposit and document scanning, machine translation
Removed: Getting around - Maps ETA and traffic prediction, ride share matching and pricing, lane keep, adaptive cruise and emergency braking
Removed: Safety on device - antivirus heuristics, crash detection, fall detection and AFib detection on watches
Removed: Behind the scenes - weather nowcasting, delivery routing and demand forecasting
Removed: Almost all of that is narrow, single purpose machine learning, computer vision, speech recognition and optimisation - the AI field as it has existed for 30+ years. Rejecting all of it is not a privacy position, you're just decided to opt out of modern computing. But don't pretend that using any modern technology validates your stance, because it does the opposite.
Removed: I will also say, this field has been a boon for accessibility - whether it's the translation feature that has been a part of Waterfox for over 7 years, pre-dating the public release of LLMs, or alt-text descriptions to assist users with low vision. Taking a blanket anti-AI stance is by all means an ableist position.
Removed: Where is the shame
Removed: As I posted on Reddit a few weeks ago, human decorum seems to go out of the window when it comes to Waterfox. People seem to forget there is a human being behind Waterfox - only one at that - and that polite discourse is the way to approach, well, anything.
Removed: I do hope the users who are "anti-AI" in all its forms take a step back and realise maybe they could be kinder, but on top of that, realise that the world is complex, often times shitty, and I am doing my best to make it just that little bit better, because I do believe my morals and ethics are in the right place here. Not only that, but unless you drop technology completely - and I mean even going back 20 years - you have already interacted with the field of "AI" today.
Removed: FAQ
Removed: Does this mean you support GenAI?! What about art and music?!
Removed: No, it does not. Art, music et al. are all intensely human experiences. I will never come to appreciate or want its usage for that. Not to mention all the ethical concerns surrounding it.
Removed: Does this mean Waterfox is slop?
Removed: Apart from my detestation of the overuse of this word, no, it is not. Every line of code is my responsibility, the art and design is either done by myself, oftentimes poorly, or I pay someone.
Removed: How can I trust I'm not reading LLM output right now?!
Removed: I think after 15 years of dedicating myself to a community, often thanklessly, the whole point of comments, blog posts and the likes are me communicating to you. Apart from the prose output being grating, it is quite frankly a pile of crap.
Removed: Does this mean you hate the environment?
Removed: No. Quite the opposite in fact - and there are many open weight, locally run LLMs that get close enough to state of the art to be useful. Believe it or not, it's not a zero sum game.
Removed: Does this mean LLMs will infiltrate Waterfox and/or its services?
Removed: No. I have been so clear on this. I just don't see how privacy will ever align with the usage of LLMs in the browser or even the services I want to offer in relation to Waterfox.
Removed: Well I'm going to go use X browser!
Removed: Good luck to you. Unless it's NOT a Chromium/Gecko/WebKit based browser it will have had LLM involvement, because it is an impossible standard to meet.
Removed: With all that, it is the weekend. I have spent countless hours this week working on Waterfox, making very little from it. I am going to turn off all my notifications and hope that logic prevails.
Removed: That said, I would like to thank everyone who engages in good faith, sincerely. To those who disagree respectfully, ask honest questions or simply say thanks every now and again.
Added: I first published this article when I felt a lot of pressure, tensions were high and it led me to post a more emotional retort than intended. Having stepped back and reflected over the past week, I have spent time rewriting this post so it explains my position in a clearer manner. For transparency, the original version remains available here.
Added: Waterfox is a privacy focused, customisable browser. It's the guiding principle in how I do anything related to the project and I don't want an LLM processing your browsing activity as part of the experience Waterfox provides. Running a model locally doesn't, by itself, convince me that integrating it is appropriate for this project.
Added: In December's No AI* Here, I acknowledged that LLMs have utility while explaining why I don't want them integrated into Waterfox. Keeping LLMs out of Waterfox wasn't a promise to oppose every single use of them and for some reason that seems to have been what was extrapolated from that post.
Added: I don't want to be a poster boy for blanket opposition to AI or LLMs. I have my own views about these technologies and agreeing with one of my decisions doesn't mean we will agree about every other application, or that I have adopted an entire set of beliefs associated with it.
Added: Keeping LLMs out of the browser doesn't require me to deny their usefulness elsewhere, reject every security improvement that involved one, or promise never to use one myself. You can hold me to the commitments I have actually made. Broader expectations inferred from those commitments aren't promises I have broken by having views of my own.
Added: Recognising useful applications for LLMs doesn't mean I approve of everything involved in their creation or everything their developers do. Using software doesn't amount to endorsing every decision made during its development and choosing to sparingly use an LLM can also be consistent with believing that the industry's overall effects are harmful. An LLM helping to discover a vulnerability, an LLM generating a patch, and an LLM running inside the browser are different things.
Added: For example, in May 2026, Mozilla described fixing 271 security bugs in Firefox 150 identified with Claude Mythos Preview. What the the account describes the human work actually needed to go from a report to an actual fix.
Added: The curl project has documented useful findings from AI-assisted analysis as well with a more honest example, showing the necessity of human verification.
Added: I won't reject an otherwise appropriate, verified security fix solely because an LLM helped identify the problem. Accepting such a fix corrects a defect; it doesn't add an LLM to your browsing experience.
Added: At the end of the day, I control my decisions about Waterfox, not every upstream contributor's tools or every researcher's methods. I can't certify that no LLM has ever contributed to any work in the history of the software Waterfox depends on and I don't believe that's a promise I need to make. I suppose refusing work to discourage practices someone considers harmful is an argument that can be made, but when an appropriate security fix has been verified and is available, my judgement is that protecting users takes priority over excluding it solely because of LLM assisted discovery.
Added: A person can accept those kinds of findings and still believe the technology's wider consequences are negative. To me, finding a useful application challenges the claim that LLMs have no use at all, but that doesn't mean I make a judgment on whether their overall benefits outweigh their harms.
Added: To that point, my concerns about environmental costs, training practices and the treatment of creative work haven't disappeared. For example, when it comes to the arts, I want human work and strongly oppose generative AI replacing what I regard as a fundamentally human experience. Waterfox's art and design are things I make myself or pay someone else to create.
Added: I think it's also important to clarify, for removal of all doubt, that someone using "anti-AI" as shorthand for an objection to generative AI isn't necessarily rejecting every application of machine learning. My disagreement is with a literal rejection of all AI, in a more absolutist fashion.
Added: For example, Firefox's translation system works on the user's device. It deserves to be assessed separately from a general purpose LLM assistant.
Added: Accessibility tools, including those that use generative AI, deserve to be considered on their own merits. Nature has reported on the benefits of these tools in terms of accessibility, as well as the potential consequences of restricting their use in academia. Be My Eyes provides blind and low vision users with AI generated visual descriptions and access to human assistance.
Added: People using such tools should be involved in decisions that affect their access to them. This does not excuse poor privacy practices, unreliable output or the forced implementation of unwanted features.
Added: Waterfox is a passion project. I put substantial amounts of time into it, make it available for free, and earn next to nothing from the work. I do it because I care about the browser and the people who use it but it also means that I don't have unlimited time or energy.
Added: Constructive criticism is part of maintaining software; a specific bug report can help me fix something, a reasoned objection can point out something I have missed and a disagreement doesn't have to be vitriolic to be fruitful. But acknowledging useful applications for LLMs, or drawing different boundaries around their use, isn't grounds for personal abuse or harassment.
Added: To wit, making Waterfox freely available isn't an offer of unlimited access to my time. It doesn't oblige me to accommodate every demand, defend the same decision indefinitely, or adopt somebody else's entire position on AI. I am responsible for the commitments I make about this project, not for fulfilling an identity assigned to me by people who want it to represent something else. That means I will set boundaries around abusive behaviour and step away from conversations that become harassment. Truthfully, I would rather spend my time working on Waterfox.
Added: To everyone who engages in good faith, including those who disagree with me: thank you. The useful reports, honest questions and occasional words of encouragement matter.