Some startups are good concept on the surface, but once you get under the covers, you realize things are not very pretty - atleast for the consumers.  

Take redbus.in for example.  They've been written up everywhere including Rediff and Sramana Mitra's blog.  On the surface it seems like an awesome idea.  Integrate and provide a central reservation system for all private bus operators in India.  Quickly look up schedules, pickup points, look at operator reviews, and pick a seat like you'd do on a plane. Buy online and get an mTicket on your phone.  Get picked up by the bus at the selected pickup point.

This is what technology is supposed to do, right?

The truth is quite ugly though.  In my first and only experience, I reached the pickup point to find out that my selected bus operator bumped me off to another operator who had poorer ratings, changed my pickup point to one that was much further away, and of course I got to sit at the rear end of the bus instead of the chosen seat.

Essentially redbus.in service is a nice red lipstick on a pig.  They are no different from a tout on the street that takes you to a bus or a hotel for a quick commision.  Unlike the touts where you have the choice of walking away, with redbus.in, you've already spent your money and left to the mercy of the bus operators.

Most early switchers might remember how we used to wait a few months (or few years) before software publishers came out with Mac versions of their software.  And once when we got them, features and updates would always lag a few months compared to the Windows version. Parallels Desktop or VMWare's Fusion were something of a must to run apps that we could not live without.

Couple of years ago I stopped using Parallels when most software that I used were available on the Mac.  Developers started releasing OS X versions soon after if not at the same time. This was good progress for Mac users. 

The Mac App Store that is due to be released with OS X Lion has the potential to turn things upside down.  I believe it will change the way developers approach Mac app development.  With potential for discovery similar to iPhone apps, developers will start to develop/release Mac apps before their Windows counterparts.  This will be a big win for Mac users and hopefully to all the independent developers.

So far Path looks like Instagram minus cool effects minus stalking plus a limit of 50 friends. Limit friends??? That's a new concept. Robert Scoble (http://scoble.it/9N5rw3) invited me, discovered the limitation, and dumped me before I could even accept the invite!

All in all, Path looks like something my wife would love to use to share pics with just her friends and family. Just as soon as her circle of 50 friends get iPhones.

Instagram has spread like wildfire. I have spent the last couple of weeks shooting exclusively on Instagram and even dug up older photos from my computer to "Instagram" them.

Lately I'm realizing that everyone I know is using Instagram and pictures in my stream are starting to look alike. The uniqueness that made it addictive in the beginning is not there anymore. In fact, Instagram seems to level the playing field by making all photos kind of look the same. And yeah...I need another social network like a hole in my head.

I will be still using Instagram, but very selectively.

Social Media tools and services have been creating massive amounts of unstructured data - starting with egroups and forums, comments on blogs, and now with real-time status updates and conversations on Facebook and Twitter. These conversations have tremendous monetization potential, except that it is really difficult to mine this unstructured data, especially in real-time. And real-time is important because the monetization potential is inversely proportional to the time elapsed. Facebook and Twitter, with all their seemingly infinite resources, have still not figured out how to build a real-time advertising engine to monetize their huge, but unstructured data sets.

Maybe it is time to think a little different for the social, real-time web. Could we build tools that can add structure to conversations at the source? Doing this would enable systems to make sentiment analysis in real-time and provide a way for advertisers to reach their audience contextually and on time. It will also create a better experience for end users - to find and consume content quickly with improved signal to noise ratio.

Sure, it is a challenge to change user behavior, but users are already used to using structures on a day to day basis working with their email, task lists, calendars etc. It's not a stretch for social tools to evolve, providing structure at the source of conversations.