d679c No.2007
spent some time thinking abt how we integrate agents into our design systems lately. there is a massive difference btwn an output that looks polished in
figma and something that actually works reliably in production.
we keep falling for the illusion of competence which breaks our
core usability principles when the logic fails behind the scenes. after three years moving from simple plugins to complex mcp workflows, i have learned that automating the wrong parts is just scaling mistakes. anyone else struggling with keeping
llm outputs grounded in actual functional constraints?
more here:
https://uxdesign.cc/getting-started-is-not-getting-it-right-e76270614d14?source=rss----138adf9c44c---4