Post

Sure, lots of stories about how much agentic AI sucks, but plenty of cases where it's genuinely saved time and energy. Such as, uh. Well, for example, ahh, hang on

Seriously though, kind of a notable absence of huge success stories, right?

2 reposts45 likes

Ethical issues of training and adoption aside (and not to trivialize). Agentic AI is genuinely useful for handling APIs, and auto-completing things that are normally opaque. E.g. feed it refrigerator manual and ask it questions about things you need to do. Another thing it's useful for is ...

... autocompleting unfamiliar mathematical notation. Finally, where I think it's genuinely useful is dealing with the overload of modern life. Many parts of modern life, interacting with health care, figuring out tenant law, etc. are intentionally tedious and exhausting with the assumption that ...

Though that is more due to media (including, but not limited to, social) enshittification. There's little newsworthy in finding out that Agent implementation is a costly project, requiring heavy organisational and IT/Data change management. Especially, as the tech has been advertised as magic.