A thought-provoking look at how LLMs might lower the burden of writing useful alt text for images, and – given the right prompts – produce better results than humans doing it manually.
It’s far more descriptive than what I would have written without it.
Jared goes on to argue that the benefits really multiply when working at scale. This part made me think of the content team at my work who maintain a customer-facing knowledge base and need to write alt text for hundreds of screenshots.
A huge time-saver is if you’re working with a large set of images. You can write the prompt only once and then say you want the same applied to each image you upload.