All sounds pretty sensible. I do think it might feel annoying waiting minutes to download a model for the sake of generating a line of alt text the first time, though. It would probably be quicker to write the alt text.
I think it’s about generating alt text for people with disabilities when they are missing from pictures.
Yeah if someone that is blind comes across a page I’m sure waiting for the description to generate is more valuable/worthwhile than a faster experience with no information at all.
This is the kinda stuff AI should be used for, this could be big for accessibility whenever alt text isn’t actually provided
It’s a neat option but the example proof of concept alt text “The Firefox logo”, as I understand it, it isn’t ideal for describing the image.
Maybe something like this?
The Firefox logo which consists of a Fox wrapped around a sphere
That would depend on the context. How the logo looks like does not matter most of the time, only when the logo itself is the topic.
Makes sense. And if it’s not relative to the content you just put decorative only right?
So…
Welcome to Firefox ([Logo for Firefox] marked as decorative)
vs
Our sponsors are [Logo for Microsoft] [Logo for Firefox] [Logo for Google]
