I'm not running a study, I'm making an anecdotal statement. That said, my actual position is that inconveniencing every single person who posts an image on a m

I'm not running a study, I'm making an anecdotal statement. That said, my actual position is that inconveniencing every single person who posts an image on a microblogging service by forcing them to write a description of it on the off-chance that a tiny percentage of the population who are visually impaired will want to read it is a net evil. It will, in the short term, antagonize people who just want to post something quickly and get on with life, and in the long term drive lower engagement on the platform as it raises the amount of effort required. It's a waste of potentially a massive amount of tiny slices of a nonrenewable resource: human life time. What should happen, instead, is that the tech wizards developing the microblogging software use computer vision and language models to provide on-demand transcription of any arbitrary image. If anyone should understand that, it's the audience here.