How would life be if you could tell your computer to cut down a 3-hour movie to one hour?

Sounds impossible?

From what I understand of this paper called “Feature fusion and redundancy pruning for rush video summarization” by the people at the Vision Research Laboratory at UCSB, it is very much possible!

The basic idea is to find ‘distinctive’ parts of the video, for example, someone talking at a high pitch or lots of moving scenes which, intuitively, would be more important than a slow scene or repeated shots.

They consider multiple facets of the video such as speech, camera motion, significant differences in color, suppression of repeated scenes and of course, identification of visually distinct segments.

The caveat is that their test data set are drama “rushes” video which are raw footage including the clapboards, the color tones, repeated takes, etc. This is very conducive to such an algorithm, which could probably explain why they had such good results (details are in the paper).

But if this is the state of things today, I can imagine that around five years down the lane they would really be applying it to commercial movies and television shows. It is amazing on what can be done with a combination of mathematics, statistics and computers.

Interestingly, the final summaries were around 4% of the total video length. If this was applied to the 8-year long Kyunki Saas Bhi Kabhi Bahu Thi show, I wonder how much it would be reduced to…

Update : Now Microsoft Research has done it for audio as well!

Yesterday, I attended The Smart Techie Startup City event. It was intended as a showcase of startups as well as for learning/sharing/mentoring.

I had taken some notes during the day. As I was expanding it into a blog post, I realized I was just adding filler words which was a waste of bits, so here it is as-is:

• Ashish Gupta, Helion VC on “Concept to Success : Milestones for startups”
  • India is a startup (positive way of looking at things)
    • High energy
    • Lots of growth
    • Small absolute number (relatively)
    • Little infrastructure or process
    • Lack of talent
    • Lots of optimism
    • Need to innovate to survive
    • ⇒ Once in a lifetime opportunity
  • Significant change in dynamics (negative way of looking at it)
    • Whatever can be made efficient will be done so.
      • We can in turn get bangalored and some other country will benefit.
    • Creative folks will thrive.
    • ⇒ We have no choice.
  • Hardest evolutionary steps
    • Those that requires behavior change
      • For example, starting to think “Become cash flow positive”
      • Next level CEO, process, tech, business model, etc.
  • Put in place metrics to measure everything – will help identify whether one has already hit an inflection point.
  • Rules of thumb
    1. Focus on customer/issue
    2. Focus on continuous improvement
    3. Intellectual honesty
    4. Results matter – only for MEASURING (measure progress on a larger scale)
    • ⇒ Same rules for person, family, company

(more…)

People often say there’s nothing much to see in Bangalore. But if you get to know some of the old parts of Bangalore, you’ll slowly start to see something special. Avenue Road is one such place:

This mini documentary was created by my friend Vineetha who was a business analyst at an IT software services company some time back and today is a journalist at CNBC!

I did the Sunfeast 10K Open Run today. I finished in 1 hr 10 min 26 sec. Finally, a timing I am happy about.

Best of all, it was a good run. I didn’t have any of my usual performance anxiety symptoms, mostly because I made sure that I didn’t plan or think about the run. I’d just get up late, hurry and reach the place, just wait to run and hope everything goes well.

I’m thankful that it worked out exactly to a T and it was a good run. A really feel-good steady-pace run.

I needed an extra boost in the end to bear the searing sun and that was provided by “Get Up! Go Insane!”. Thanks Fatboy Slim!

There were people running for their charities, people running for fun, people running in costumes, and even people running out of curiosity. I was running to fight against my off-late tendency to give up easily.

On a different note, I must appreciate how well-organized the entire event was. After the run, they gave everyone bun and biscuits. It may not mean much in a normal situation, but after a run, it’s really important and I was thankful they had thought to this level of detail.

After I came out of the Kanteerava stadium (the start and end point of the run), I bumped into an old school mate after a really long time. He looked at me and asked:

Him: Oh so you came for the run?
Me: Yeah
Him: So you completed?
Me: Yeah…
Him: (has the ‘not bad’ look) So how long did you take?
Me: 70 min.
Him: Oh. (face expression changes) I took 90 min. Okay. cya later.

So next time, don’t underestimate that fat people can’t run.

Running is the greatest metaphor for life, because you get out of it what you put into it.

– Oprah Winfrey

One of the hard lessons that I have learned this year is “Always remember Carpe Diem“. The corollary is that “If you don’t execute on your idea quick, someone else definitely will.”

For example, long back Vikram had this idea that there should be a company which takes care of odd chores such as electrical maintenance or plumbing, basically handyman work. Yesterday, I saw www.handiman.in on the back of an auto rickshaw. I came home and checked it out and it does exactly that. It’s a very useful service and seems reasonably affordable, at least for IT people. I’m sure lot of people in Bangalore will go for it.

Today, Mrinal pointed to www.indimeme.com, a TechMeme for the Indian blogosphere.

I started kicking myself.

I’ve had this idea for months but I couldn’t really move on it because I don’t have the knowledge yet, for example, about clustering algorithms. However, I did brainstorm it with a couple of friends and thought we’ll work it out. But a single person beat us to it.

There is a range of reasons why such a website is a good idea, probably the same reasons why TechMeme is indispensable too:

• Allows people to see what are the latest topics that Indian bloggers are talking about.
• Allows people to see the discussions across blogs, not just one blog and its comments.
  • Encourages the above type of discussion.
• The portal can become the gateway of the Indian blogosphere.
• For the website creator’s point of view, it can bring in a lot of visitors. And subsequently, advertisers.
• An indispensable website means the creator of the website is indispensable too. Just like Gabe Rivera is everything behind the scenes of TechMeme. (Let’s face it, we’re all replaceable in our workplaces.)

And so on.

Anyway, the only downside I’ve noticed about IndiMeme.com is that the clustering results aren’t good yet, but the thing is it is already out there. It has been executed. It needs refinement. And I’m sure it’ll get there.

I don’t know whether I should add this idea to my already-long personal ‘deadpool’. Sigh.

When I started thinking about this idea, I came across one paper called Mining blog stories using community-based and temporal clustering which explained how this is a special type of clustering that takes time into account. They call it:

“[the] Content-Community-Time model that can leverage the content of entries, their timestamps, and the community structure of the blogs, to automatically discover stories. Doing so also allows us to discover hot stories.”

I was thinking whether the same idea can be applied to an RSS aggregator and then I found that was done too as well.

I guess there are simply no low-hanging fruit left in this accelerated world.

I have been impressed with the urgency of doing. Knowing is not enough; we must apply. Being willing is not enough; we must do.

– Leonardo da Vinci

Philipp Lenssen recently had a good post on tips on information overload by various people. It got me thinking about the various tips and tricks I’ve imbibed in the recent past and which work reasonably well for me. So I tried to collate them into one place:

Email

• Always bring the inbox down to zero regularly. ‘Regularly’ is defined by you.
• Never allow anything to be in your inbox > 2-3 days
  • If you’re not going to reply in that time frame, you never will. So simply archive it or reply with a one-liner saying you can’t look into it now.
• If you don’t have anything to add, don’t reply.
• Make sure you are clear on what is the action you are expecting from the recipient.
• Reply in bullet points. Because everybody skims.
• Once you’re done with the email (replying, taking action or reading), archive it.
• If it is not actionable, archive it. Don’t let it remain in your inbox.
• Use keyboard shortcuts.
• Mailing lists go into folders. I simulate them in Gmail using “Apply label, Skip Inbox” in the filters. The reason is that mails not directly addressed to me are not urgent, so I can process them whenever I have the inclination. Whatever is in my inbox is what deserves immediate attention.
• Minimize the number of times you need to check email. The minimum that is required for you to stop worrying about it. The beauty of email is that you can reply at your pace. Make use of that feature. If you end up constantly checking email, you’re better off resorting to phone calls or instant messenger.
• [new tip] Before you send the next email, go through the checklist.

Feeds

• Use your feed reader once in a few days. The world won’t stop without you.
• Use a desktop feed reader because it is faster to use.
• Have a ‘Try Before You Buy’ folder where you add feeds. If it doesn’t turn out to be useful, delete it.
• Have a number in mind, say 100 feeds. If you add a new feed, delete an old feed that is no longer interesting.
• If you end up doing a ‘Mark all as read’ on a feed 2-3 times in a row, delete it.
• Separate them into categories and/or priorities.
• Most importantly, read interesting things. Do not aim for reading 500+ blog posts a day. Optimize, don’t maximize.
• Remember that the goal is to derive some value out of this reading and that value is usually knowledge. If it is not helping you towards that goal, delete it. Don’t think twice, just delete it.
• While working, if you feel the need to distract yourself once in a while or read something interesting, don’t use your feed reader but use good filters like TechMeme or programming.reddit or a good link-blogger on your subjects of interest. Have a separate dedicated time for reading feeds.
• Take notes. Over time, you’ll judge if a feed is useful or not depending on whether you’re taking (any) notes or not.
  • You can also use a mental version of the “McDowell Grid”.

Inlets

• Cut down on the types of inlets – Email, Feeds, Twitter, IRC, Messenger, Phone, etc. (this one is particularly hard for me)
• Spend at least 50% of your time at the computer with all these inlets shut down.

Focus

• Personally I find productivity inversely proportional to information overload. The days when I’m productive and “in the zone” turns out to be the days when I’m less affected by information overload. The vice-versa is true as well. So if you focus on the right things, the information overload problem will get solved by itself.
• Maintain focus by having a todo list. Have a big todo list and then pick random tasks from that list depending on your energy levels and get things done.
• Never indulge in tasks outside of your todo list. If you’re not in the mood for any of them, don’t indulge in wilfing. Go out instead – whether for a walk, or call up a friend or even read a paper book. If you’re not being productive, just get out of the chair.
• Don’t use fancy software for writing lists. Use a good plain text editor (like Vim).
• Use GTD.
• Use an auto-pilot schedule (I’m still learning this).

P.S. Many of these ideas have been borrowed from elsewhere. It’s been a long time since I imbibed all these, so I don’t remember all the sources from which I gleaned them.

There have been many times where I’ve been asked for “career advice”, especially after a talk. I usually suggest them to ‘build a repertoire of things you have done, things you are capable of, things you like to do’. But I’ve never really been sure of this advice nor do I feel I have the credibility to answer such questions.

The good thing is that I have now found something to point them to – Garr Reynolds‘ presentation on a career advice book. This presentation explain things very well and is so beautifully done that it can capture the attention of a young mind:

| View | Download

The other resource I have found useful is Aaron Swartz’s “How to Get a Job Like Mine”.

Cultivating a good career is like creating the Mona Lisa. The right tools and strategies will only get you partway there – the soul of the artist is necessary to create something worthwhile.

– Steve Pavlina

It all started on Monday last week when Lakshman tweeted if anybody was game for a weekend trip. Ashwin tweeted back saying yes.

Later they asked me over email. My reply was “Why Ooty!?”. They replied saying “It doesn’t matter. We’re going for the drive.” Two days later, we three were driving to Ooty on bikes at night.

My descriptions below are in twitter style as an ode to how the trip happened. ( But of course, my usual writing style will resume after this post )

We started off at 8 at night. First stop was some lip-smacking food at Kamat Lokaruchi:

9.10 pm : We’re on the way! 3 twitterers on bikes.

9.25 pm : Getting out of Bangalore is the toughest thing.

9.30 pm : @scorpion032 says 2020 will also be the year of the linux desktop.

10.32 pm : @cruisemaniac and @scorpion032 are tweeting away…

1.40 am : Taking a break.

2.33 am : Admiring the mysore palace…

3.16 am : Sleep getting to me… But we’re taking breaks and having fun. In Nanjangud.

3.20 am : Another break.

3.34 am : Listening to My Sacrifice at 330 am at 70 kmph on bike with the wind in your hair is something to be experienced.

4.10 am : We find a freakin’ coffee day in the middle of nowhere. Waiting for capuccino.

6.15 am : Mudumulai forest.

6.20 am : Animals! Elephant, deer, peacock, mongoose, eagle, woodpecker, … All right there next to us… Thank heavens the elephant didn’t think we were pesky…

6.45 am : Exiting Mudumulai forest.

My favorite photo from this trip (notice the clouds and the bike):

(more…)