This isn’t going to work whatsoever with people who don’t know how to express what they want to do.
Tons of people have just been taught a fixed workflow involving a sequence of buttons with known labels and icons and locations. Lots of people already can’t find programs in the start menu even if they know the name (because they don’t know how search works and often even will think it’s not the same program / will think it won’t have the same data because the icon was found in a different place).
How are they suddenly going to talk to an AI about things that the AI don’t even have information about? The AI won’t know all the nicknames people have, it won’t knew how people describe the icons, can’t handle all misspellings (they don’t even understand phonetics), it won’t under people’s description of the UX parts, and when programs have 20x start options where people usually follow a guide to pick the right one then the AI won’t be able to reliably recognize which one the user intends to open.
Every single company would literally need a team of AI training experts and capture EVERYTHING the employees does with the computers and says about them for a few months to capture all the context it needs.
corsicanguppy@lemmy.ca 1 year ago
I like how the copilot button will allow Microsoft to run what it thinks you want to run.
That would be anticompetitive, but Microsoft learned from the last time. (And what it learned was “nothing’s going to happen to you so carry on”)