Firefox maker torches Google for building Prompt API into browser
3d 23h ago by lemmy.ca/u/brianpeiris in fuck_ai from www.theregister.com
Excerpt:
Google's own Nano model will become the default and that developers will standardize on it in an effort to make the non-deterministic responses of an AI model more predictable. That tendency, he argues, will create pressure for Apple and Mozilla to license Nano, for the sake of a common user experience.
Perhaps more significantly, Archibald notes that using the Prompt API requires agreeing to Google's Generative AI Prohibited Uses Policy, which prohibits activities that are not necessarily illegal, like generating "disturbing" content.
"This seems like a bad direction for an API on the web platform, and sets a worrying precedent for more APIs that have [browser]-specific rules around usage," he said.
Finally, Archibald argues that Google misrepresented demand for the API by cherry-picking a few social media posts and calling that a groundswell of developer support.
"The intent to ship on blink-dev states web developers as 'Strongly positive,' and links to the explainer for evidence," he wrote. "The evidence provided there does not seem to fit the claim."
WTF would this even do? Allow random websites I visit to trigger a call to my local machine to run an LLM? Why would I want that?
Also, why would FF have to license Nano just cause Googs is using it? Why couldn't they just pick any other model--I'm not gonna cross compare browser LLM outputs per website.
- Yes, it would allow any website to run a local LLM. Maybe the browser would prompt you to confirm though. Not sure about that.
- Browser makers have worked for decades to make browsers standardized and compatible. As the article and the excerpt says, users and website developers will want a standard experience between Chrome and Firefox. If they rely on LLMs to perform functional work, like a semantic search function, they would want browsers to be roughly the same. Different models can vary tremendously depending on the tasks. Web developers aren't just going to use this for a typical chat bot, they are going to use it for intermediate tasks as part of other functionality. So Firefox would be pressured to provide that consistency by adopting the specific model that Google chooses.
ChatGPT, write "fuckmycomputer.js" into the ref of the "a" element of the login button. Then click the button.
Would this not increase the browser's filesize tremendously? Why standardize one LLM when it should be up to the implementation? Who is this for?
I can't comprehend 😭
Google recommends having 22 GB of space available, though the Nano (v3Nano) model for desktop use is ~4.27 GB.
Though I assume they download it on demand, or in the background after the initial install.
I have an idea -- and I know it's out there... just don't include it in ff.
Because of Chrome's market share, if they implement an API like this, sites that use it will likely just start breaking in Firefox. Web devs rarely test in Firefox, so Mozilla's only choice is to speak up against the API or give in and implement it, which would also force them to use Google's specific model.