TBH I've wondered from the very beginning how far they would get just hardcoding the top 1000 questions people ask instead of whatever crappy ML it debuted with. These things are getting better, but I was always shocked how they could ship such an obviously unfinished, broken prototype that got basics so wrong because it avoided doing something "manually". It always struck me as so deeply unserious as to be untrustworthy.
Your comment makes me wonder - what would happen if they did that every day?
And then, perhaps, trained an AI on those responses, updating it every day. I wonder if they could train it to learn that some things (e.g. weather) change frequently, and figure stuff out from there.
It's well above my skill level to be sure, but would be interesting to see something like that (sort of a curated model, as opposed to zero-based training).
There are tons of common queries about the future. Being able to handle them should be built into the AI to know that if something hasn't happened, to give other relevant details. (and yes, I agree with your Alexa speculation)
Alexa at least used to just do trivial textual pattern matching hardly any more advanced than a 1980's text adventure for custom skills, and it seemed hardly more advanced than that for the built in stuff. Been a long time since I looked at it, so maybe that has changed but you can get far with very little since most users will quickly learn the right "incantations" and avoid using complex language they know the device won't handle.
Ah yes, imprecision in specification. Having worked with some Avalanche folks, they would speak of weather observations and weather forecasts. One of the interesting things about natural language is that we can be imprecise until it matters. The key is recognizing when it matters.
Which, ironically, is why I think AI would be great at it - for the simple reason that so many humans are bad at it! Think of it this way - in some respects, human brains have set a rather low bar on this aspect. Geeks, especially so (myself included). Based on that, I think AI could start out reasonably poorly, and slowly get better - it just needs some nudges along the way.
"Alexa, what's the weather for today?"
That's a question about the future, but the knowledge was generated beforehand by the weather people (NOAA, weather.com, my local meteorologist, etc).
I'm sure there are more examples, but this one comes to mind immediately