Effortless Experiences

Ahead of the Apple Watch event, I thought I’d take a minute to lay out what it is I’m thinking about these days, what types of companies I’m looking for right now and what I think is coming next:

Effortless Experiences

What, exactly, do I mean by that? The easiest current example of an “effortless experience” is Google Now. If you haven’t used it, it is a contextually aware, predictive computing platform that tries to anticipate what information you need next based on a mix of your location, calendar, email and Google searches. Things automagically appear in Google Now, so – “effortless”.

“Smart services” is another way to start thinking about this. We invested into Waze years ago – this was (and is) an extremely smart service, that optimizes your travel path based on current traffic conditions. Very smart if you haven’t used it – and at times you get very unexpected turn by turn directions. There’s some work involved on the user side, but the results are something you or no human would have ever thought of…

Which leads us to yet another way to describe this: Artificial Intelligence. In the case of Waze, we’re talking about very narrow, domain specific AI (although the service has never been described in those terms) that’s delivering a smart service. But, I’m not suggesting you’re going to get Hal 9000 on your wrist (just yet).

Smart Watches in my view are going to usher in an era of effortless experiences. Why? The form factor is simply too small to be a heavy computing platform; but is a fantastic platform for things to automagically appear. I’m not clever enough to guess what types of great ideas are going to come out of this, but directionally, I’m comfortable stating where I think things are going.

We have the computing power in the cloud and on the handset, but the UI/UX  on a handset just isn’t that great for push services. A watch form-factor is an entirely different matter… Google Now takes on a whole different feel when it’s on your wrist vs taking a phone out of your pocket, unlocking it, opening Google Now to see if there’s something useful there. With a watch – it’s just there. Effortless.

So, where does that leave us? I think we’re on the cusp of some great new services, and that smart watches are going to be a key enabler that unlocks that future. That said, and watches aside; effortless experiences are what’s next in my mind.

P.S. If you really want to read more on A.I., I can recommend the following (lengthy and terrifying) blog post: The AI Revolution: The Road to Superintelligence

The Invisible Internet

One area I’m thinking about these days is what I call the “Invisible Internet”, so I thought I’d put a post out there in case there are apps and products I should be looking at. Feel free to reach out if you’re building a company in this space…

Big data is gaining a lot of interest from VCs. e.g. companies that can crunch data and find patterns for large corporations. I’m interested in Big Data for the Little Guy. We invested into Worksmart Labs, who make a great product called Noom. Noom is a pedometer for your Android device. It’s fantastic because once you’ve installed it, you don’t need to do anything else to gain benefits from the app: it sits there and watches you, tracking every step you take. Of course, you probably installed it because you want to know how much exercise, walking etc you’re getting. But that’s the beauty of it – it just sits there silently tracking you. Piling up that data, zero effort from you (except the walking).

I have a withings scale at home, which works on the same principle. I weigh myself every day, but my withings account captures all the data and starts crunching. Fitbit has just announced its wifi scale as well. Strava.com is a cycling site that compares me to others, crunches my data to tell me how hard any given ride was (Strava’s Suffer Score). I love it.

Clearly, wireless healthcare is an area where I’m seeing the Invisible Internet become productized, but there are other areas as well – self writing journals that track where you go, smart home applications that monitor your energy consumption, etc.

Maybe it’s age, but more and more I’d prefer my apps just figure out what it is I’ve done, and what it is I want vs having to work on getting the information in and out of them – I want to be surprised and delighted by insights or information that I would have missed out on otherwise.

I hope to see more of these apps in the future. Or maybe not 😉

Feeds 2.0 and the Netflix Prize

I’ve blogged about my friends at Feeds 2.0 many times over the past few years; I’ve followed the Netflix prize out of interest since it was covered in Wired back in 2007, mainly because I was rooting for the Feeds 2.0 team…

For those of you who may not spend your weekends and evenings reading up on artificial intelligence, here’s an overview of the Netflix competition:

Netflix released a large movie rating dataset and challenged the data
mining, machine learning and computer science communities to develop
systems that could beat the accuracy of their in-house developed
recommendation system (Cinematch) by 10%. In order to render the
challenge more interesting, the company will award a Grand Prize of $1M
to the first team that will attain this goal, and in addition, Progress
Prizes of $50K have been awarded on the anniversaries of the Prize to
teams that have made sufficient accuracy improvements. Apart from the
financial incentive however, the Netflix Prize contest is enormously
useful for recommender system research since the released Netflix
dataset is by far the largest ratings dataset ever becoming available
to the research community. Most work on
recommender systems outside of companies like Amazon or Netflix up to
now has had to make do with the relatively small 1M ratings
MovieLens data or the 3M ratings EachMovie dataset. Netflix provided
100480507 ratings (on a scale from 1 to 5 integral stars) along with
their dates from 480189 randomly-chosen, anonymous subscribers on 17770
movie titles. The data were collected between October, 1998 and
December, 2005 and reflect the distribution of all ratings received by
Netflix during this period. Netflix withheld over 2M most recent
ratings from those same subscribers over the same set of movies as a
competition qualifying set and contestants are required to make
predictions for all 2M withheld ratings in the qualifying set.

I was thrilled when I saw the final Leader Board: The Ensemble (Feeds 2.0 + others) was at the top:

The final winner will not be announced until next month; Netflix still has to decide which of the leading algorithms perform best and how they score on various tests…

So, as the latest Wired article says, it ain’t over til it’s over.

Good luck Nicholas!

You might like… Recommendation Engines

If you're reading this blog, then you might like recommendation engines.

I've been continuously impressed by Amazon's recommendation engine. As I've expanded what I buy through Amazon (heart rate watches, cooking utilities, computer peripherals, etc.), Amazon has done a very good job of processing those likes -across categories- and making very intelligent suggestions that have resulted in purchases.

Apple, on the other hand, can't seem to figure out what MUSIC I like despite having bought numerous albums and even being signed up for their artist alerts… Apple constantly alerts me to new hip hop and r&b content. I bought some snoop dogg two years ago…but I'm no big hip hop fan. What about Dave Matthews? I've probably bought about 5 albums. Alerts about Dave? Zero.

I'm aware of the "gift drift" that you can get with Amazon if you buy a friend's kid a copy of The Gruffalo- for the next several months you get offered children's books non-stop (I've actually never had a problem with Amazon's recommendation engine).

Looking forward to the mobile internet, recommendation engines- or agents- are going to become even more relevant- who's around you, what they are listening to, and why you might like it, that your favorite store (based on transaction history) has your favorite stuff on sale and the bus is 12 minutes away… etc.

Apple is smart (understatement), but the problems they have with recommending good music underscores the challenges facing recommendation engines.

If this is something that you're particularly interested in, O'Reilly has published a great book on the topic:Programming Collective Intelligence.

Highly recommended reading.

The Next Web

Good post over at Genuine VC yesterday covering the History of the Web:

…a transition among three distinct phases of consumers’ primary activity online from receiving, to hunting, and now doing…

Receiving, Hunting, Doing is a good indication of what we’ve seen so far- and I think we’re going full circle to "Receiving" again- only this time from intelligent sources.

Two examples are Kwiry or Tripit. With Tripit, you email them your travel itneratry and they scour the web in the background and send you a nice package of maps, directions, thoughtful suggestions, etc.

I’ve posted many times about "intelligence inside" which lives in the same neighborhood as the Semantic Web:

…I also think there’s a huge opportunity to get to data sooner via the sensor revolution. When phones report location, when phones listen to ambient sound, when credit cards report spending patterns, when cars report their miles traveled, when we’re increasingly turning every device into a sensor for the global brain, there will be more and more sources of data to be mined…

All of which means the process is reversing: doing by machines, hunting by spiders and receiving by users- remix and repeat.

Intelligence Inside

I’m always on the look out for smart new apps that save me time.

Feeds2.0 is one of my stand-bys for RSS readers. It ranks incoming articles based on your previous reading patterns and tends to border on clairvoyance. Pretty amazing stuff.

That said, I’m up for trying new services. A few new ones I’ve come across this week are:
silobreaker (thanks Rob)

All of them are slightly different, but very useful. I’m trying out feedhub sent through Google Reader. We’ll see if their results can top Feeds 2.0. Pipes is just so hackable, it begs to be loved and silobreaker is a news-junkie’s best friend. Enjoy.

Open Source Artificial Intelligence

Numenta made their software available (for educational use) earlier this month. In case you haven’t heard of these guys, their platform for intelligent computing (NuPIC)

…implements a hierarchical temporal memory system (HTM) patterned
after the human neocortex. [They] expect NuPIC to be used on problems that,
generally speaking, involve identifying patterns in complex data. The
ultimate applications likely will include vision systems, robotics,
data mining and analysis, and failure analysis and prediction.

Numenta’s platform builds intelligence from scratch- in the same way a human baby does. What is immensely powerful is that machine-based intelligence is only a download away for other machines, no learning required. I can only imagine what happens when multiple applications built on the NuPIC platform are merged into an integrated system.

I’m looking forward to working with the HTM technology. One of the areas I’m interested in looking into is seeing if Numenta’s algorithm can help me discover the most relevant articles in my multitude of feeds based on my previous reading habits (a perennial favorite of mine).

Anyone else seen any interesting AI applications I should be looking at/testing?

Continue reading “Open Source Artificial Intelligence”