If anyone with 4S has some time, I'd like to know how Siri responds to:<p><pre><code> * Map from here to the nearest gas station
* Map from here to the nearest gas station, avoid highways
* Turn screen darker (or lower brightness)
* Redial (last called number)
* Turn off 3G when near home
* Turn off Wifi when near work
* Turn vibrate mode on when near work
* Turn vibrate mode off after 6pm
* Repeat song (when iPod is playing)
* Open HackerNews (saved as bookmark)
* Delete Safari cookies
* Update AppStore apps
* Which actor plays Ryan in The Office?
* Alert me when battery has less than half charge</code></pre>
Does the choice of being unable to call 911 seem strange to anyone else? I think the risk of occasional accidental calls [1] is nothing compared to emergency situations where the user is too panicked to operate the touchscreen, or blind, or [insert your own edge case here]. Or is this solely a legal restriction on emergency services?<p>[1] I once somehow butt-dialed 911 on my old dumbphone, which was embarrassing and terrifying. I awkwardly apologized and hung up, and there were no repercussions.
This reminds me of playing around with Dr. Sbaitso when I was a kid. <a href="http://en.wikipedia.org/wiki/Dr._Sbaitso" rel="nofollow">http://en.wikipedia.org/wiki/Dr._Sbaitso</a>
Siri is going to be a great marketing machine for Apple. People will love to show it to their friends. And it all happens on the servers, so there's no end to Siri's incremental 'intelligence'.
Did the original Siri also do stuff like this?<p>I think it’s a great move on the creators’ part. Building silly stuff like that into Siri creates massive amounts of goodwill. That will make it much harder to really hate Siri when it – inevitably – screws up sometimes.
As random as some of these seem, clearly Siri has amassed a ton of input data.<p>With the volume of people playing around with Siri I'll bet darn near every clever question has already been asked a dozen times.<p>Heck with that amount of data pouring in Apple could hire a team of comedy writers to just hit the daily top 1000 new questions.<p>Wonder how they are ranking the quality of the responses / satisfaction.
Anyone knows if Siri can keep a todo list alongside with the calendar? That would be great, actually. "Siri, what's the three most important things I need to do today?"<p>Also, something I really need assistance with is to be aware of the time it will take me to get to the appointments I have. I always leave my office just a few minutes before I need to be somewhere, no matter if it's within walking distance or fifteen minutes by car. I'm like an infant in this regard. "Siri, let me know when I need to leave for my two o'clock appointment" would be great if it worked.
Damn it, now I really want Siri.<p>Too bad I only have an iPhone 4. Pulling the original from the app store was kind of a dick move, Apple. (even though I understand the motivation)
Today, a friend and I have been having fun using the Android version of Siri <a href="http://speaktoit.com" rel="nofollow">http://speaktoit.com</a> (best one we tested).<p>It understands 90 to 95% of the conversation we spoke to it. Also, it understands 8th grade humor type words and reprimands you by saying things like, "Would you talk to your mother like that?"<p>Fun stuff for the 8th grader in all of us!
This stuff is pretty immature, and indicative of it being something Apple didn't build internally. It makes Siri seem gimmicky and highlights the programmed nature of the responses (instead of being more AI-like, as I had hoped).<p>These answers serve no purpose, except for one's buddies to go "heh heh watch what happens when I tell it I'm drunk"
This has everything I've thought to ask Siri and more... I'm very impressed. I hope Apple plans on supporting and iterating Siri rather than forgetting about it like they occasionally seem to do...
Just an observation I've made, but isn't this almost identical to what WolframAlpha's goals are? With the exception of handling your own personal data, they both seem to work in the same ways: Understanding natural language and interpreting massive amounts of data in different ways. I'd be interested to see how they are different underneath.