Anil DashOne of the top stories on Hacker News today was a post arguing that Mozilla shouldn't accommodate any usage of AI in Firefox because (understandably) people were mad at Big AI companies for all the horrible things they've done to users and the internet and society. But I think people are ignoring the reality that *hundreds of millions of users* are using LLMs today, and they need to have tools from platforms that will look out for their interests. https://www.anildash.com/2025/11/14/wanting-not-to-want-ai/
Igor (BotchedRPR)By all means, use Firefox and don't use AI. I don't give a damn about that, that's great. But think about this question deeply: What browser do you want your friends and family to use when they're using AI tools? We told people "social media from big companies is bad, get on Mastodon and Bluesky". And that works — provide useful alternatives. But if we simply say "stop using this thing you find useful, and I won't give you any alternatives", they'll just ignore us. And Big AI will win.
Rpsu (326 ppm)@anildash I really wish we had a LARGE button to turn off all AI in Firefox, or at least a clear choice how to use it - explicitly.
I think not allowing users to choose is a huge underestimation of users. Designers know how to help people in this kind of decisions if it’d be a priority. AI has its uses, but currently putting everything under one single label of “AI” makes choosing and discussing about it difficult.
Rowan Tommins@rpsu @anildash This seems contradictory to me: if the problem is putting everything under one label, then surely the *worst* thing we can do is give everyone a single toggle labelled "AI is scary" <-> "AI is awesome".
What's needed is a better vocabulary for talking about what an "AI" tool is actually doing - Is it running locally or sending data to a giant corporation? Is it using a prompt as a natural-language interface for an existing specialist tool, or attempting to simulate what the tool would have done? And so on.
In other words, I want effort spent on building an *AI control panel* that gives meaningful choices, and maybe some opinionated defaults.
yoasif@anildash All indications are that Mozilla plans to roll out more integrations with existing AI platforms.
How does that help the situation exactly? You say that "social media from big companies is bad" -- what is the corollary here with AI? Is there a "good" AI that Mozilla is secretly working with?
You say Big AI will win.
What is little ai?
Karl Heinz Häsliprinz@KarlHeinzHasliP @anildash If that is what Firefox's AI Window is, I think people will be very surprised. I guess we'll see.
Linux and praxis: part of the resistance 🇵🇸 ☮️@linuxgnome @anildash I don't actually want to leave Firefox. 🤷
Dustin Westphal@anildash We want an alternative without any AI bullshit at all, which is what Firefox used to be and what it should be.
Mercutio@Bachus @anildash
https://www.waterfox.net/
A privacy-friendly Firefox fork without DRM, AI and other intrusive techniques.
Dan Wineman@anildash If we don’t have a smoking section in our restaurant, people will eat somewhere else and Big Tobacco will win.
Chris Ferdinandi@anildash “what brand of cigarettes do you want your kids smoking when they light up?”
Jeff Atwood@anildash yes. do not say "don't use this" without providing an alternative.
@codinghorror I mean, I have sympathy for "Don't do this, and let's figure out what to do instead!" Because sometimes it's hard to solve problems and we have to band together to find alternatives. But "don't do this bad thing, dumbass, be smart like me" sure is guaranteed to fail.
@anildash ok but smart.. like how? what do the "smart" people do, exactly? Can you describe it? Just move to a remote island in the pacific and sit motionless doing nothing? I mean.. what?
Jennifer Kayla | Theogrin 🦊Horse, water, drink.
I can proselytize about the dangers of 'AI' until the cows come home. If they won't listen to me on the matter of a world-killing mind virus which will never produce anything of meaningful value, they're certainly not going to listen to me when it comes to changing their desktop wallpaper -- or anything in between.
Ben Curthoys@anildash I find it ironic that one of the things I find LLMs most useful for is helping me with the documentation gap on getting FOSS things to work.
Before, if I was trying to install something, following the instructions, and I got an error because my distro was ever so slightly different than the one the instructions had been written for, my choices were:
1. Spend a month reading the code to understand how the thing actually works
2. Spend a week reading obscure forums
3. Give up
@anildash And in that context an option "4. Typing the error into Co-pilot and getting an explanation of how it has gone wrong and how to fix it" is incredibly useful. Sure, sometimes the answer is off. So is 90% of the advice given in obscure forum posts!
@anildash One argument I've seen against using an LLM for this is that "this isn't what they're being advertised as for or funded for". So the fuck what? Just because my use case will never justify the bubble valuation doesn't mean it's not a use case.
⁂ Fish Id Wardrobe@anildash i don't want my friends and family to use "ai" tools — and if they find them useful, that says more about them than it does the tools.
smyru
Christophe B. 
@anildash I think it's rather a matter of explaining them why something they find useful (AI) is actually not that useful because it's all smoke and mirrors and bad at the tasks they assign it. There are usually AI-free alternatives for their use cases that do the job better.
lia
Grey the earthling@anildash
> What browser do you want your friends and family to use when they're using AI tools?
I think AI tools are harmful in the first place, so I don't think they should use them at all. But, suppose they're going to anyway: you're saying Mozilla AI can be a harm reduction strategy (like monitored drug consumption rooms). I hadn't thought of that before.
I think a lot of the emotion is because there are precious few reliably AI-free options, and Mozilla (or people's idea of what Mozilla used to be circa 2005) seemed like the last best hope.
Perhaps a harm-reduction approach for those who are already AI users is sensible, but those of us who are clean just want to stay away from it completely. And we don't want to encourage non-smokers to start vaping. So I think it would have been smarter to have a very clear bright line between AI harm reduction and AI-free — two separate products.
Mozilla could have used this as an opportunity to introduce a new brand for their AI browser, and dodge any years-old associations with “Firefox” being slow or incapable.
> What browser do you want your friends and family to use when they're using AI tools?
I think AI tools are harmful in the first place, so I don't think they should use them at all. But, suppose they're going to anyway: you're saying Mozilla AI can be a harm reduction strategy (like monitored drug consumption rooms). I hadn't thought of that before.
I think a lot of the emotion is because there are precious few reliably AI-free options, and Mozilla (or people's idea of what Mozilla used to be circa 2005) seemed like the last best hope.
Perhaps a harm-reduction approach for those who are already AI users is sensible, but those of us who are clean just want to stay away from it completely. And we don't want to encourage non-smokers to start vaping. So I think it would have been smarter to have a very clear bright line between AI harm reduction and AI-free — two separate products.
Mozilla could have used this as an opportunity to introduce a new brand for their AI browser, and dodge any years-old associations with “Firefox” being slow or incapable.
Michael K Johnson@anildash Thank you for articulating this.
I am so, so tired of self-righteous folks here proclaiming that literally no one, not a single soul, wants these features, and that the same literally no one will ever again recommend Firefox as a result. It's arrogant posturing on their part, and detracts from a reasoned conversation.
I've struggled with how to respond to those posts, and you succinctly and clearly articulated every half-formed thought I've had about the topic, and plenty I hadn't. 🙂
🚲 murph
dxzdb@anildash @codinghorror I left Firefox when they couldn't say they weren't selling user data.
Brave seems like a great alternative. There are quite a few others like Vivaldi & Orion for example.
Yeah - saying “Stop Using Amazon” and do random searches to find things has definitely NOT worked with relatives.
Pierre Gab@anildash That’s what I'm doing right now. On every device I have installed Firefox, my main browser, I have a list of things required to have a nice experience:
Remove sponsored shortcuts, remove Google Search, disable AI.
The list is growing. There is a finite amount of work going into Firefox's development, I hope they choose features users want. But I know that’s not going to happen.
There has to be some enshittification. If there was none, everybody would use it, so it would need to run ads to cash in on that, that’s how the modern world works. Without Google Search default, Firefox would already be dead and buried. We can be glad that the pile of shit exists in Firefox, for the pile of shit is feeding developers and the browser would cease to exist without it.
But we can be honest and recognize that this is a pile of shit that no sane Firefox user wants (please forgive the no true Scotsman fallacy, but I'm really curious about users who would want AI in Firefox. Has anyone conducted interviews with them?).
Ben Rosengart@anildash why does it need browser integration tho … ?
Ed
Fabiano Celentano@anildash Big AI will win either way, whether we destroy Firefox for it or not. That's not the point. The point, as with all these questions, is whether there is ONE alternative left for people who don't want all their data to be hand delivered to American fascists.
Calicosine@anildash "AI" is inherently demon tech, it must be destroyed if art is to survive.