Hacker News Reader

Comment by trophycase | Hacker News Reader

Exactly. There is a distinction that needs to be made between something that is trying to solve my problem for me, and something that will let me best solve the problem for myself, in my own way.

Nothing frustrates me more than when software thinks it knows better than me.

View on HN

> Nothing frustrates me more than when software thinks it knows better than me.

Would it be fair to rephrase that as "... and fails"?

Because that's what we're really talking about, right? Those times when the smart system failed to correctly predict user intent, but lacked a fallback method for the user to clarify?

I'm sure there's an uncanny-valley-esque term out there for the difficulty "99% correct" smart systems have when users hit that 1% of incorrectness.

I'm not sure of a good term for it, but by design when these "smart" system work you'll hardly even notice. They "just work." However that 1% of the time they fail disastrously, you'll definitely notice.

  > Would it be fair to rephrase that as "... and fails"?

No. Unless you consider failure to be the vast majority of interactions with smart software

> There is a distinction that needs to be made between something that is trying to solve my problem for me, and something that will let me best solve the problem for myself, in my own way.

Max I borrow that phrase for all future references from UX colleagues regarding smart UI? I really hate smart UI as it changes while I keep expecting other things and constantly have to relearn or search for items/stuff that was there just a moment ago, but cannot be fount after a reload.

This sounds like the classic "tool vs agent" design concern.

It's always fun backspacing one character at a time to surface the result I wanted, that showed some number of characters ago. Windows 10's search is especially bad about showing things that don't start with the characters I type until I land on (or chisel out) the right characters.

Reply on news.ycombinator.com