31 thoughts on “Can Mobile Phones Think?”

      1. Tasker on Android is perhaps amongst the top few best applications in the entire mobile ecosystem. Fabulous application that justifies calling Android phone with Tasker as real smartphone!

    1. +1 Awesome app and it has been updated quite frequently to my surprise. I use it to actually tell me the time when I want it to, not every 15 minutes or 30 minutes (and it works unlike many clock/alarm apps) You can set volume settings according to date or time or when you click on a widget. I have also made personalized slide shows that run at the click of a widget, or you can make your own soundboard of sorts (sounds at the click of a widget) You can also do this according to location, I haven’t because of battery sucking GPS.

    2. Wow, this looks to me as a very sensible and great app. The first time that I run into something that my iPhone cannot do.
      Maybe next phone will be an Android? 🙂

  1. Definitely a trend with watching, and I wish I still had my N8 review unit to try this out. I’ve noticed others with this approach as well. HTC is adding software to turn your ringer off when you place the phone face down. T-Mobile added My Modes for easy switching of phone profiles in different situations, i.e., work and home. And 3rd party app, Locale for Android, can use a phone’s GPS for certain actions: I’ve used it to automatically turn off my WiFi radio and turn on Bluetooth when it “senses” that I’ve left home. It’s time to make smart phones “smarter”! 🙂

    1. +1 to that. I think the HT C Incredible has the latest Sense UI (amongst all the phones I have) and I like how the background changes with the time of the day. I am sure there are a lot of other apps. Maybe we should talk, you know phones are actually good for talking 😉

    2. Kevin, there is software for Symbian called GSM Navigator which works similar to Locale but can also work based on the cell sites you are accessing so that it doesn’t drain the battery like GPS does. It can switch profiles, launch applications, send messages based on your location and when you enter/exit a locale. There’s a lot you can do if you find the right app but there is not just one app store for Symbian.

  2. So what exactly is exciting here? You can tag times and locations and all kind of stuff. Then you make it interact. Simple filter rule.

    For lady gaga you need pills and counseling. A smartphone can’t save you.

  3. Thinking != following pre-defined rules only
    Context != programmed categories with pre-defined attributes

    What you describe is just location based categorization. No thinking from the app required, nor context. In a simple case context augments data to make them useful, not required here since it’s pre-defined.

  4. Great to see Nokia turning into the innovator again.

    Nokia has this working on Symbian^3. I’d assume they must also be looking at running it on MeeGo as well next year, though the open-source nature of that OS would allow other device makers to grab it.

    On a side note, I hope the flopped launch of Windows Phone 7 doesn’t shake Nokia’s confidence at bringing out a new mobile OS. Nokia just needs to learn from where Microsoft went wrong, and not repeat those mistakes.

    This news shows that Nokia still has the ability to innovate.

  5. So after reading the entire post i came up with this,i just starred at my Nokia for a while and smiled a bit.
    I must say that the times have changed a lot,smart phones are getting more smarter than human beings.
    Nokia with its n8 provided a really cool mobile device,its too feature full but Android is having unlimited apps for smart phones,so people are following the path of Android,nokia must beat it.

  6. The technology is very advanced today but thinking phones, I don’t think so. They just do what they are programmed to do.
    mainstreethost

  7. Tend to agree with Ronald. This appears to be an implementation of ‘situational awareness’.

    What a lot of us want is a trusted personal advisor that we can ask for suggestions as well to ask what it thinks about something. If the system is continually learning then the suggestions (advice) get better. Intelligent learning systems (also called cognitive systems) for mobile are not far off. It may sound Star Trek’ish but that is where we are heading.

    At the other end are intelligent systems that are task-oriented and Siri, which Apple acquired a few months ago, is a good example.

  8. An interesting comment by Yuri Milner, who has invested north of $100m into Facebook, at last weeks Web 2.0 Summit:

    “I think that when you have billions of people connected and the unprecedented pace of exchange of information then you need filtering mechanisms. And now we’re in the very early stages where your friends and networks are doing it for you, like on Twitter and Facebook. Or the Google approach, where there are a whole bunch of machines that are learning fast. And I think there will be a convergence between those two models. And I think Facebook will be one of those platforms from which AI will emerge in the next ten years.”

    Yuri states the problem well and is naturally biased towards FB but I think there will be alternative solutions to the problem.

    The full interview is at http://www.youtube.com/watch?v=YbhXcw5ItmU&p=2737D508F656CCF8 and the above is covered briefly at the end.

    1. Dinesh,
      I have been comparing notes with Jeff Jonas over the past few weeks. He coming from very large data sets, me just doing a lot of theoretical work and some small data.
      Yet we have the same understanding, which makes me believe we are extremely close to some major break troughs. Since theoretical work shows the same trajectory as very practical work.
      BTW: one is “mother nature doesn’t do statistics”. Hence Google for example is not on the same trajectory, based on what I have seen in their research they are heavily invested in statistical modeling.
      Don’t know where you stand, none statistical modeling is a hard pill to swallow for most people.

      1. Yes, I know of Jeff Jonas and a truly wonderful person. I/we are of the Bayesian statistical modeling school. Not sure why this would be a hard pill to swallow? Please elaborate.

        Our premise is based on the notion that people learn concepts from examples and generalize to new items which is one of the cornerstones of intelligence. You can read more at http://thenoisychannel.com/2010/04/04/guest-post-information-retrieval-using-a-bayesian-model-of-learning-and-generalization/

      2. I don’t know what intelligence is. So sorry.
        We learn not from example but from observation, which covers examples. It’s the same with learning, people say they have build a learning system. Which can just recognize or memorize, but learning is self organization. Which covers recognition and memorization(with delays, speed differential is important here, builds the underpinnings of the
        addressing scheme, which is important for smart search …) and the building and organization of abstracts.
        Context is very important in all of this. I could go on for ever.
        Anyway to give an example of the power, a Notebook i3 with 4g Mem, can read over 90% of all(tested, ~100) (re)captcha’s in under 3 minutes/captcha. Context does not only give meaning to “mean”, but also
        to the abstraction of chars. In other words it scales from a number of pixels to very large data sets.
        Here’s another one:
        The author either has no clue what context is and how it’s created and processed or he’s doing an incredible good job of concealing it.
        What he really want’s to do is calculating the information value of any given article for any given person. But for that he needs context which consists of past(memory),present, and future (predictions, goals). My guess, a 12 core personal cluster(2 AMD 6 core, 16G, with 2-4 fast bonded networks and a really big pipe [to Google, somebody got to categorize it first]) will be 100 times better at this than he can ever imagine what’s possible.
        No “intelligence” required.

  9. just a more advanced version of Settings/Preferences

    even in the old days you could set yr power saving preferences to what state yr computer was in …

    at the end of the day the user still has to define the states … is that going to really happen? … all those microwaves flashing at 12:00 all the time tells me not

    what is REALLY required is an intuitive understanding of the where a person is with minimal setup … which means artificial learning, situational awareness without defined inputs, and social cognition

    1. Eric

      Clearly this is a starting point and perhaps we will end up at a point where the phones become more self learning and intuitive. That is why I said — Nokia needs to make it their top priority and push the envelope.

  10. I can second the recommendation of Tasker for Android. I have it set up to turn on the GPS when an app needs it, and to turn it off when the app ends. When I leave home it turns off WiFi, and turns it on when I come home. When I turn the phone facedown it triggers silent mode. It takes effort to set up, but does a lot when it’s done properly.

  11. You are absolutely on target with this post. Context awareness is progressing rapidly and at Nokia it is one of a number of ways we can deliver more consumer value. Just to be clear, when it comes to connecting the real world with new opportunities enabled by the digital world, my team and others at Nokia are indeed focused to follow through on innovations (our own and those of 3rd party developers) by increasingly integrating Ovi services into the core functionality of our devices.

    P.S. Nice to see you are using your Nokia N8 birthday gift!

  12. there are quite a few apps for maemo that do similar things.

    it will be exciting when meego is released and we can have apps like this too because it will make porting them a breeze.

  13. I am also a former Local (for Android) user. The app does everything you mention basically, but it also comes with plug-in and additions you can download once you have the app itself (for example, I download a plug in for Google talk for my phone which starts up my Google talk up everytime I leave my house and turn it off when I get back)

    Other uses I had for it:
    – turning off my ringer and put the phone on vibrate when I’m teaching at my part time job

    – integrate with my to do list and notify me when I’m around a place that’s on it (for example, if I pass the supermarket on the way home, an alert would go off and a message saying “get meat” will show up.)

    – turn off the wifi and mobile network at night automatically, so the phone because a “dumb” alarm clock and doesn’t bother me with loud alerts when someone updates their twitter (sometimes I forget to do that manually)

    – more stuff…

    I uninstalled the app recently because my phone is crawling with too many services already. I will probably install it again when I get a new phone. Try it out!

  14. I purchased a utility on my Nokia E71 that allowed me to set up rules like “Silence phone when any calendar event has the word ‘meeting’ in it; back to ringer when event ends”. (I think my Treo could do this, too, with an add-on.) OK, not intelligence per se – very nice, though, and something I wish my iPhone could do. Apple focused on UI responsiveness and simplicity, both very good things. I hope, as hardware gets better and better, we can eventually have the best of both worlds.
    Maybe a device should not be called a smart phone unless it does things for you. Things like noticing patterns in usage, locations, contacts and calendar events. Warning me when traffic is unusually bad on the commute I do every morning. Automatically downloading podcasts it notices I like to listen to.
    There are some apps now that change their behavior based on time of day, e.g. Instapaper and TomTom GPS automatically use night mode when it’s nighttime. The umbrellatoday.com app isn’t a phone feature exactly, but it also simplifies my life. I guess it’s a start.

  15. I’ve been using Locale for my Nexus One (and now my G2) for this specific purpose – it’s basically the same type of app – it doesn’t really ‘learn’ or ‘think’, but rather reacts to various triggers based on location, situation, etc.

    The problem that these types of apps have, currently, is the false triggers. Users will expect these apps to work flawlessly, and few people want to give software the chance to ‘learn’ from them.

    It took me about a month of actively noting my actions/preferences before I got Locale to work exactly how I want it. There are still a few things I need to tinker with, and I’ve been using it for 6 months or so now.

    Nokia Situations has some serious potential – combined with Nokia Bots (which really did learn from your behaviour) – it can be a real game-changer.

    Unfortunately, as Om pointed out, Nokia’s awesome at innovating and then not capitalizing on it. For examples, see QR Codes, front-facing cameras, full desktop browsing, app stores, etc.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.