Andre Natal and I did the Firefox translation feature. It started as a research project with Mozilla and some EU schools and was called Bergamot. We adapted tha

Andre Natal and I did the Firefox translation feature. It started as a research project with Mozilla and some EU schools and was called Bergamot. We adapted that for Firefox starting a couple years ago and shipped last year. The technology was a perfect fit for the user feature. I think the same holds for summarization (a Reader View feature maybe?) and as someone what was on Firefox's A11Y team for almost 5 years, I think there's some work Mozilla could do with image alt text and perhaps fixing up some common quirks that trip up screen readers so they don't have to do that fixing themselves. There are several places I believe this relatively new tech fits perfectly with web browser use cases. Those mostly aren't monetize-able though, so they're "for users" not "for money" and users don't really seem to be up for grabs these days, too locked into their routines, and so those features may not even be competitive advantages. Still, doing good for your existing user base by incorporating new tech that fits their use cases well, even if it doesn't make you more money or win you more uses is a good thing and the fact that Mozilla's not chasing growth like FANG and others means they can do more of that unprofitable "good stuff." I worked on Mozilla for 25 years. It was an amazing ride and knowing what I know about Mozilla I feel confident they'll behave not just responsibly with AI, but as paragons of virtue and a north star for the industry (and perhaps even for governments as the Mozilla Foundation has been working on responsible AI for half a decade and is in regular contact with the US and EU governments on just these kinds of issues.)