Summary I use AlephAlpha’s multimodal LM to generate alttext for the eww web browser. This is in order to keep websites textual. When browsing the imaginary web, it is more useful to have textual representations of images. This way, you can do things such as run google searches, just by selecting their description, as opposed to relying on a multi- modal prompt entirely for 𝑖web search. Prompt pf-given-an-image-describe-it/1 http://github.com/semiosis/prompts/blob/master/prompts/given-an-image-...