Even if they didn't realize it, I don't believe they were arguing that firefox and chrome didn't/wouldn't use machine learning already, rather that they just thought the use cases you provided don't really sell the cost of having a full LLM integrated into every browser install.

"AI" as it's used nowadays is unfortunately usually a shorthand for LLM. When firefox talks about "AI features", I think most people interpret that as "LLM integration", not the page-translation feature that's been around for ages.

That’s different from an agentic browser in a few key ways.

Most importantly it’s far more difficult for a bad actor to abuse language translation features than agentic browser features.

Nowadays they call AI everything. Browsers translate websites from decades, when AI was only a word you would see in science fiction movies.

Okay, what's the problem? The UX of Google Translate is fine

- it will pop up when it senses a webpage in a language you don't speak.

- it will ask if you want to translate it. You have options to always translate this language or to never do it.

- it will respect your choice and no pop up every-time insisting "no please try it this time". Or worse, decide by default to translate anywyay behind my back.

- There are settings to also enable/disable this that will not arbitrarily reset whenever the app updates.

There are certainly environmental issues to address, but I've accepted that this US administration is not going to address this in any meaningful way. Attacking individuals will not solve this issue so I'm not doing this. So for now, my main mantra is "don't bother me". the UX of much AI can't even clear that.

Alternatively: they’re already taking advantage of the AI features they like without at all needing “AI in the browser” and do realise it.