Hacker News new | ask | show | jobs
by skybrian 1141 days ago
Learning how to use freeform text input can be pretty annoying when only a few things work. Some examples are playing a text adventure ("guess the verb") and using an unfamiliar command line interface. Good error messages can help.

Web search changed that. Most queries work, at least somewhat.

There's a point where freeform text input becomes better than structured input. A simple search box is what people mostly use instead of an advanced search form, let alone a web directory (like Yahoo! back in the day).

For web search, there are very few error messages. If you enter a query that doesn't work very well, you get back results that aren't very good or what you wanted, so you try something else.

With AI chatbots, expectations are sky-high, but there are times when they should refuse with a good error message, because they really can't do what you're hoping to do. An example is when you ask it to explain its reasoning. An LLM never knows why it wrote what it did, but it will try to invent a plausible explanation anyway. [1]

Better error messages that help users understand what chatbots can actually do would help avoid misconceptions, but this won't happen unless the error messages are trained in.

[1] https://skybrian.substack.com/p/ai-chatbots-dont-know-why-th...

1 comments

Yeah, exactly. A freeform input promises it can do 'anything at all' but if it can't in reality, it's always a frustrating interface.

Classic example, Siri. It's so easy to quickly find stuff you feel like it should be able to do, but it just can't. "When was my last message from Steve" etc.