CTU ringtone

Brad Feld had posted several weeks back about a CTU ringtone. Being a big fan of CTU I immediately downloaded the mp3 file and promptly forgot about it. Until today. I was cleaning up my desktop on my Mac PowerBook and found this file so decided to commit the time and figure out how to get it loaded on my Samsung a920 multimedia phone. I played around with bluetooth for a bit and tried to upload the file to no avail. It turns out Samsung blocks mp3 file types from being uploaded via bluetooth so I went to the web. After about 15 minutes of searching I found this comment on a blog entry back in February 2006.

you can set your mp3s as ringtones. first make a smaller version of the mp3 that is only about 15 - 20 secs long with the part that you want. (this should be about 100K with variable bit rate) Then goto funformobile .com and upload your ringer to your phone for free, you can also send pictures and games.

This link will take you straight to the upload page.

Agent Bauer here ;)

landmark blogger case

Are bloggers protected under the same rights as journalists from being forced to reveal sources? Josh Wolf hopes the answer to that question will set him free. Josh was imprisoned more than two months ago in a California case involving protesters who burned a police car. Josh authors a video blog entitled The Revolution Will Be Televised. The government would like him to turn over his unedited tape to identify persons that may have been witnesses to the event. This should have strong repurcussions in the blogging community as authorities try to use the blogosphere to discover, track, and indict evil doers. If bloggers are not afforded the same protection as journalists, people will have to be much more careful about that which they blog.

The entire story can be read or heard on NPR.

Helen Grenier and iRobot rock

I'm currently at the DowJones VentureWire Emerging Venture conference out in San Jose and have the privilege of listening to a keynote interview with Helen Grenier, founder and Chairwoman of iRobot. A couple of things struck me during the interview as hallmarks of successful startups. These are model adaptation and making the product easy to buy.

iRobot started out with an industrial vacuum cleaner that they worked on with Johnson Controls. It turned out that there were too many problems and too hard to sell in that environment. One of the engineers came up with the idea of doing a dirt simple (pun intended) robot for the home. They locked down the price point at $200 and only allowed features to be added that kept the price point intact. This caused them to create a bare bones housing with just features that enhanced cleaning (moving from hardwood floor to carpet, getting underneath furniture, moving around corners, etc.). iRobot then took this vacuum (called the roomba) to a focus group of housewives and they absolutely hated the idea of a robot in the house. In order to get sales off the ground they didn't call it a robot, instead it was categorized as a "floor sweeper". Keeping the product at $200 made it affordable and competitive with other sweepers. This is a classic story of a company adapting their business model on the fly to find the crossing of demand and value at a level that will sustain the business.

iRobot has since moved into the defense industry with their line of Packbots. Packbots are lightweight robots that can traverse all kinds of terrain and go into dangerous areas without the need to send in humans. For example, they can be sent in first to explore caves that are inhabited by enemy combatants to look for arms caches, etc.

You can also find an interview with Helen from 2004 on Engadget here.

Bottom line: smart executive with a passion for product + adaptive mentality = success

Those early customer meetings

I was reminiscing with my co-founder (Don Springer) the other day about one of the early customer meetings we had waaaaaay back in mid 2005. Early in the lifecycle of a company you often have "opportunistic" meetings. That is to say, you meet with anyone willing to meet with you, and use the feedback to try to hone your product and message. This can take you in a lot of funky directions and is the early crux move for a startup to survive and move to the next phase of existence.

We had scheduled several meetings during this particular week long trip to NYC, many with fairly large names in the hedge fund business, but one was a small hedge fund out on Long Island. This was our first meeting of the week and we literally had to catch a cab from the airport and go straight to the meeting. On the cab ride there we talked through the pitch for the 10th time in 2 days and were feeling fairly confident as we moved closer to our geographical target. The cab driver had a hard time finding the address (which should have been our first clue) and I think we got into a debate along the way on how many hedge funds were located on Long Island.

It turned out to be this guy's house. He was running money for one wealthy guy and was very into commodity trading, describing it as where the real money would be made over the next few years. His house was in sort of a middle class neighborhood and his "trading area" was a room off to the side of the living room that he shared with a hot tub and large screen television. As I recall, he was fond of the idea but not sure how he would use it in the commodity trading space and he talked to us for a good long time. In the meanwhile our cab driver who we had paid to wait, left us there. We were trying to figure out how to get out of there, had no cab, the conversation around how we might work together was exhausted, and then came the uncomfortable silence followed by rambling conversation. He offered us snacks and came back from the kitchen with some cold pop tarts and Hawaiian Punch. The cabbie no-showed and the guy finally drove us in his minivan to the train station.

Bear in mind that this was just prior to the huge run up in commodities in early 2006 so the guy probably made a trailerful of money.


what are your stories?

Software being developed to monitor US citizen sentiment

There was an article that appeared in the New York Times last week that outlined plans from the department of homeland security to develop software to monitor the opinions of US citizens towards their government. The article points out how Orwellian this process could become and states that the program could take "several years" to roll out.

My guess is that they are much closer than they are letting on and it will be used long before it is accurate enough for incrimination. The accuracy problem is in understanding context and detecting satire. These are very tough problems to crack.

Examples:

In a recent review posted on Rotten Tomatos about the movie, "The Queen" the reviewer writes, "Brilliant dissection of the rot at the top of British high society and politics" contains the adjective, "brilliant" w.r.t. the overall subject, "brilliant disection of the rot at the top of the British high society and politics." And also contains sentiment inside the subject, "rot at the top of the British high society and politics."

Or, as described in Bo Pang and Lillian Lee's paper, "Thumbs up? Sentiment Classification using Machine Learning Techniques", the sentences "How could anyone sit through this movie?" contains no single word that is obviously negative.

Or, in the following case that most algorithms would consider ambiguous, "This film should be brilliant. It sounds like a great plot, the actors are first grade, and the supporting cast is good as well, and Stallone is attempting to deliver a good performance. However, it can't hold up"

My company has been working on sentiment models like these for a couple of years now and much of the progress has been made using statistical language processing algorithms. Some of these are similar to algorithms used in email spam detection applications. Depending on the problem domain, I've seen upwards of 90% accuracy. This is acceptable in many commercial applications but will require a human net to cover the last mile of accuracy when perfect accuracy is required. I agree with the writer and think it is a bit Orwellian, and reflects the times in which we live. This may help in the war on terrorism but my guess is it will only catch the groups too stupid to cover their tracks.

dual temperature controls are useless

Why do auto manufacturers offer dual climate controls? Dual controls have got to be just about the worst feature ever invented for the automobile. The idea that you can actually create different environs less than 1 meter apart is ridiculous. This is, in fact, a great marketing idea. It symbolizes the ultimate in choice, like the sleep number mattress. You can have it your way while the person right next to you can also have it theirs. No need to even talk about it. What is wrong with it is that it does not work. It gives you a false sense of choice while complicating the controls of the vehicle.

I miss the simple controls in my 97 Forerunner. Give me a slider control for the temperature and a fan control. I'll actually communicate with my wife to agree whether it's too hot or too cold.

Collective Intellect podcast interview

David Cohen came by the other day and did an interview with the company. He has put together a media site to track the progress of Colorado technology startups. You can find David's site here, and the podcast here.

It was a lot of fun doing the interview and another data point of how easy it has become to unleash content into the world.

Smart People, Great Views

We recently moved the company to downtown Boulder, right on the walking mall. This has been wonderful for the company in many ways: close to 10 great coffee shops, book store, people watching, outstanding restaurants, bike paths and walking trails, nice views of the flatirons, convenient to shopping to pick up those "I'm sorry I've been working so much" gifts on the way home, not to mention the hammered dulcimer player with the operatic voice. The list goes on and on, and it's just an incredible energy and morale booster for the company as a whole. But probably the biggest plus is the number of really great and really smart people less than a bagel toss away.

I was having a coffee the other day with a friend of mine, Sandy Keziah. Sandy's a great guy and knows a lot of people in the area, particularly if they have a marketing bent. At the coffee shop we ran into John Winsor, the CEO and founder of Radar. Oddly enough, I had never run into John before. I say, "oddly" because I run into the same people often enough that when I attend a local VC or entrepreneur associated event it feels like a small town rotary club meeting ("Hey bob, catch any fish this weekend?").

I sat down with John this morning for a cup of coffee at my new favorite caffeine injection facility (Alison's near the corner of 15th and Pearl) and had a wonderful exchange of ideas with John. John has a very interesting business background and has written a couple of books (which I'm now planning to read and can be found on his blog via the link above) around engaging your customers.

Sandy, thank you for the introduction and John thank you for sharing your thoughts this morning. I love working in Boulder.