Reply to post:

"These are things that standard probabilistic analysis, plus a little location information, ought to be able to help with. No AI magic required."

Yeah, you know why? Because that's all the "AI" we use actually is. We don't have AI, we have trained statistical engines, that's it.

Google and Facebook's data streams don't mean a lot even with every morsel of information - it's almost impossible to sort the chaff from the noise without understanding of context and no "AI" has that.

The reason Samsung can't "compete"... it's because that's all there is. There's really not much else to it. Sure, Google seeing an email that looks like a flight ticket is handy. But that's as far as it goes - recognise flight tickets - add to calendar... erm... what else? Maybe start suggesting holiday gear? That's about all you can do.

We do NOT have AI of any kind. We have statistics and various ways to "train" computer to apply those statistics, which basically boil down to "keep trying to jam these statistics into a correlation against something useful by brute force", which usually ends in a "good enough" approach that'll never get better, and nothing we couldn't do with explicit coding.

Until we start getting machines that actually understand, and can infer context, not just rely blindily on probabilities based on previous data, we'll never have AI.

Literally... since the "AI revolution" of the last, what? Decade or so? What's changed? Siri really does nothing we couldn't do with some statistical analysis, even 20, 30 years ago. The only thing was that we didn't have the capability in the portable devices to do so. But it doesn't actually break any ground. Nor does any of the other stuff.

The irony is, we'll all walking around with chess grandmasters in our pocket, but they still can't tell whether an email is spam or not, or - in my case - do anything even approaching non-trivial voice recognition.

It's the wrong kind of machine, the wrong kind of approach, the wrong kind of the problem for the approach we use, and the wrong kind of hype.

Honestly, in a way, I'd be glad to not having Samsung throw tons of money at AI and just produce a nicer phone. Even if every line of code was written by a human.