It's in wide use now, so you think of it as pattern matching.

How is that not a straw man? Or at least telling me what and why I think?

I said I thought arbitrary speakers was crappily implemented, but it's the specific speaker voice recognition that I said was pattern recognition.

You said:

When an AI application hits the mainstream, you don't think of it as AI any more.

I thought you were posting to me, since you were quoting my post. I didn't realise you'd switched to a second person that didn't include me.

Yes, but mathematics doesn't stop being mathematics when it's used by physicists. Logic doesn't stop being philosophy when it's used by mathematicians.

If you do think that you can't make a useful discovery in another discipline while researching another, then we have an unbridgable semantic/philosophical divide. I see it as quite simple, and I can't imagine how to explain it. I can discover a physics principle doing chemistry, I can discover something during aeronautics that turns into medicine -- why not?

I think we should look him up in the Codex first.

What am I supposed to do with a talking points memo on why a project it currently over budget to the tune of, excuse me has grown by a billion dollars? Laugh?

I'm thinking you're supposed to have it on your desk in case a reporter gets to your line accidentally. No?

I can't come up with a useful definition of intelligence that doesn't boil down to "really really really good pattern matching."

I think the problem with this:

When I went to college in the 1980s, voice recognition for an arbitrary speaker was considered incredibly difficult, and was indeed an AI topic.

Is that as we've gotten better at creating this sort of thing, the threshold for what's considered AI by the public has gotten higher. What was defined under the AI umbrella in 1980 doesn't ping us as "intelligence" any more, and possibly shouldn't. It seems too strong a word.

I can't come up with a useful definition of intelligence that doesn't boil down to "really really really good pattern matching."

Heuristics. Learning over time (not just training), and the ability to make new rules. No doubt a personal view on AI (like I said -- most chess? not intelligent enough -- impressive number crunching, though -- do they have programs that can learn how to beat specific people over time? That'd rock), but that's what elevates it for me.

My ease in describing OCR as AI is probably slanted by knowing a decent amount about it, so it's not just familiarity breeding contempt.

Hey, isn't Nutrax for Nerves, per Peter Wimsey??

Heuristics. Learning over time (not just training), and the ability to make new rules.

Oh, yes! Absolutely.

(I don't know why my brain lumped this in with pattern-matching, because I really did have it in mind. Today would probably not be the best day to give me a Turing Test.)

My ease in describing OCR as AI is probably slanted by knowing a decent amount about it, so it's not just familiarity breeding contempt.

any other theatre/music geeks thinking "original cast recording" or is it just me? or that I'm listening to the Wicked soundtrack, which I will be seeing in THREE AND HALF HOURS.

Why does Nutrax want you to take pictures of your food?

How will a camera phone help me eat better?

Great question - one that we are bursting with excitement to answer. They let you record and learn from everything you eat. This is the most important thing you can do to eat better for a lifetime. And, they let you do this easily.

You may have kept a food diary before. Chances are it was a great healthy decision you made. Chances also are, that it took a lot of time to maintain. Paper food diaries are a great end but not a great means to eating better. With your camera phone, you won't need to write down what you eat!

Best of all, your dietitian will look at photos of what you actually eat - not what you think you eat to give you custom food coaching. In effect your dietitian will "see" every bite.