Post

this is the really ineffective head-in-the-sand reaction a lot of folks on Bluesky tend to have if you talk about any harm reduction-based approach to the reality that millions of people are using AI tools today. The current approach to critiquing AI is obviously not working, but they don't care.

Been debating writing my "AI is not going away" article about why people reacting with the "shut it all down" response are making it more difficult for people who want to build better user-empowering AI tools

There's no such thing as a user-empowering AI tool.

Glasses that identify objects for people with vision problems? That seems user-empowering to me

Respectfully, that’s not what anyone is talking about. This is about LLMs.

I am not an expert, but I am nearly certain that LLMs are usually involved in object recognition eyedaptic.com/seeing-beyond/

so there is a smartphone app called Seeing AI, designed by a blind coder at Microsoft and a group of ML researchers at Microsoft using the largest 2015 models, which are what grew into LLMs; originally it used the various Cognitive Services that Microsoft built to eg identify objects