Post

Chrome looks set to ship an LLM Prompt API to the web platform. At Mozilla, we oppose this API. We feel it has a large interoperability risk, and Google imposing T&Cs on a web API sets a dangerous precedent. Full details: github.com/mozilla/stan...

There's some legitimate concerns here. Let's try to be concrete about how we will identify the harm in practice so that years from now we can all learn from this difference of perspective. github.com/mozilla/stan...

I.e. @jakearchibald.com I've already admitted you were right and I was wrong in one case (view transitions vs. portals / high-vs-low level web APIs). Will this be one where we even the score or will you double your money? 🙂

One thing I've struggled to get an answer for: Is Google serving the same Prompt API model to e.g. users in China Vs UK? I'm wondering if the user's location is leaked due to restrictions in what the model can and cannot say about things like Tiananmen Square.

I do not understand privacy concerns around prompt API because the non-determined nature of LLMs makes me think of it more like a geolocation API than querySelector, where the nature of using it will result in data leakages. What am I clearly missing?

Geolocation is permission gated?

Google services largely don't operate in China AFAIK, so no model download at all (if you can get even Chrome installed given no plat store). I'm proud that the company would rather be blocked as illegal than compromise on principles like striving for the most accurate information.

So the lack of the model may be a leak? But otherwise it's a consistent one-for-all model?