My summary of the SXSW Interactive event in Austin:  The panels and speakers were only a small step up from the level of do-it-yourself barcamp speakers, the bands were loud, and I learned a lot from conversations with other attendees.  Maybe some of it will be interesting to you.

"Seed combinator" incubators
From the time that I sat down on the plane to Austin next to Allan Tear of BetaSpring, I ran into a lot of people working on microventure incubators. These are patterned after the original Ycombinator model. They select a batch of 10 or so startup teams, and provide very small investments of $10K or $20K. When I first heard about this from YCombinator founder Paul Graham in 2005, I initially wrote it off as being insignificant. Who would want a $10K investment? It's designed only to pay the founder's rent while they work for a few months on a prototype. But, it turns out that other parts of the model that generate a lot of excitement. Putting the startups teams together in batches, with local "mentors", gives them ideas, motivation, and support. At the end, they do a "demo day" where they show their work to potential follow-on investors. The results from that can be astounding. In the last Boulder Techstars group, 8 out of 10 teams received follow-on financing. VC's are supporting these efforts because they provide valuable services for screening and preparing entrepreneurs.

Most of the people that I talked to are working on microventure incubators as a way to do local economic development. Boulder is a small city far from the tech centers. They used the Techstars incubator to make their software and startup scene much more interesting. If it can work for them, it might work for the folks from Boston, Providence, San Antonio, Omaha, etc. Each city effort involves finding local investors and mentors who already have success in technology or business. The theory is that a good mentor team will draw in good candidates and make a successful program. Go team!

At Assembla we focus on being global, not local, but I think it would be great to spread some microventure opportunities to Assembla users around the world.

Working with investors safely
Ycombinator founder Paul Graham had some specific advice that was a little different, including "Move to Silicon Valley." He was running startup batches in both Boston and Silicon Valley, but he notes "I always felt like the Boston startups were getting short shrift... especially because of a lack of follow-on investors." Paul has a baby and a tan since he moved to California, so it must be working out for him. I asked how he prevents his early stage investments from getting crammed down by later stage investors - the classic angel problem. His answer "they wouldn't dare!" They rely on Ycombinator for dealflow, and they also know that there are hundreds of Ycombinator alumni in Silicon Valley ready to defend their comrades. According to Paul, this "provides [equity value] protection for founders and for everyone else further down the chain." This gives us another couple of actionable hints for founders who plan on raising money from professional investors in the "how to use them safely" category. 1) Move to Silicon Valley, and 2) Sell shares to a microventure group, VC partner, or respected angel investor that is needed in to the startup ecosystem. Get that protection.

I guess it's not cool to be an old style VC at SXSW.  All of the investors that I met were nice guys who themselves complained about the practices of big funds who push out founders and find ways to force in a new CEO and a lot of extra money. They all talked about "Capital efficiency". They appreciate the situation where software deals are getting smaller and more numerous and more capital efficient.

Software development ideas
A Google Android developer threw out a good idea on how to evolve a good user interface. It uses the Android idea of "intents", which can call a user-selected applet to fulfill a request. In his recommended process, each developer makes a different UI that can plug in to the "intent". According to this developer, "it only takes a day or two." Android UI's are pretty small, and most of the functional code is in routines that are shared by all of the UI alternatives. Then, users can decide which one to use, and "we naturally work on the one with more users."

I want to try this for Web apps - making multiple page templates, and letting users decide which one they like, and then throwing away the losers. I think it will give a better result than arguing over mockups, and take about the same amount of time.

Which reminds me that a LOT of people are working on mobile apps.

The folks from Ellis Labs (their Codeigniter and ExpressionEngine communities totally rock) talked with me about using Assembla for agile software development. Like many of us, they are fully distributed with team members in Germany, Illinois, Missouri, and Oregon. Like us, they ask each team member to use the Scrum tool to report on their work and obstacles, at the beginning of his or her day, in his or her time zone. If you use Assembla, try this type of daily report. It allows you to provide scrum master services and clear obstacles effectively.

Ellis uses their own Mercurial server that is hooked into an Assembla workspace and ticketing with the "External SVN tool". Derek Jones said "We had a moment of joy when we realized that "External subversion tool" just meant "External repository".

Allen Stern of Centernetworks is managing his own distributed development team, and using some best practices like paid trials for new developers. So, he was able to comment on problems with a couple if different freelance recruiting services that he has covered, and how they impact the workers. For example, 99Designs is getting good results from freelancers, but it asks them to do work on "speculation", without any guarantee of payment. Multiple designers can submit designs, and only the one selected as the winner gets paid. That's not as good for a reliable provider as being selected early on, and getting guaranteed payment.  Allen was more critical of oDesk, saying that requiring users to submit to continual monitoring by screencam and screenshots "amounts to slavery", and is disrespectful of people from foreign countries.

I saw a lot of different pservices for recruiting and qualifying freelancers or buying freelance services.

Questions
Where are the foreigners? Almost everyone here is from the US of A. I didn't meet a single person from India, China, or Germany. It seems like we are exporting things like Facebook, Twitter, and jobs, and not importing anything in return. Aren't Korean interactive services doing amazing things? This one way mirror can't be sustainable.

Are users really"traffic" and "assets". For example, Brad Hunstable of uStream said "American music awards was a live event. Once we got the audience together, we had an asset that we could use..." It just sounds bad. At least the porn vendors gave the advice "respect your audience."

Is the revolution in on-demand cloud hosting bypassing Texas? San Antonio has a cluster of hosting companies that were exhibitors, including Rackspace, ThePlanet, and ServerBeach, and they all came up to Austin with booths. These are very successful companies with loyal customers, but they sell mostly the old-style dedicated servers. I think on-demand servers are resources are important for development speed and deployment reliability and scaling, so why aren't they more important for hosting companies?

Finally, a note from the airplane where I am writing this on the way to San Francisco. I am over the flyover American west, the great basin, and there is no place remote or barren or dry or rugged enough that when I look down, I do not see a road, a straight line, slicing across the landscape of serrations, curves, bumps, and webs. There is no place of sanctuary. All has been cut under the wheel.