Bill Boebel

What am I going to do with all these startup ideas?

Loyal typepad followers... I've launched a new personal website: http://billboebel.com
Key feature: a new blog where I'm going to start posting the random startup ideas that come to me all the time.

I'm running Google Chrome OS

I installed Google's Chrome operating system on my laptop (Dell Latitude D420).  In fact, I am writing this blog post from Chrome OS.  Pretty cool, huh?

To install, I downloaded the source code and followed the build instructions on the Chromium website.

Here are a few problems I hit, and the resolution:

• I couldn't login at first because it wants to validate my Google account credentials against the Google server, however my wireless wasn't working initially.  So to fix I followed the optional steps listed in the build instructions to "Enable a local user account" and "Set the shared user password".  Now I can login with the local user account I created, and I can run sudo commands.

• Wireless wasn't running.  In a terminal (ctrl-alt-t) I saw this error in the dmesg output:
       iwl3945: iwlwifi-3945-1.ucode firmware file req failed: Reason -2
       iwl3945: Could not read microcode: -2
  
  So I copied these driver files from my laptop's Ubuntu (Karmic) install:
      /lib/firmware/iwlwifi-3945-1.ucode
      /lib/firmware/iwlwifi-3945-2.ucode

Unresolved Problems:

• Mouse is annoyingly slow by default and changing it didn't work until I closed the laptop lid and it went to sleep and came back.  But it loses it when I restart.
• Closing lid makes the laptop hibernate correctly; and it comes back super fast, but then it errors and hibernates again.  Then if I hit the power button it returns and starts working normally (not a reboot).
• Can download file attachments, but can't directly open.  Need to manually upload to google docs
• Have to boot from a USB drive since my laptop doesn't have a solid state had drive :(
• Booting isn't that much faster than Ubuntu yet
• Periodic "A server error occurred. Please contact the administrator" errors with the start menu (its a web page)

In general, this OS is very simple (just a browser) but exactly what most people need... they just don't know it yet.  Its not reliable enough for normal use yet, but this will become my primary OS sooner than my "Desktop 2015" prediction 4 years ago.

Desktop 2010

I told you so... Google OS is coming.

Btw, gOS, I'm still waiting for my beta Cloud OS account.  Better hurry.

Black Lips + GZA #sxsw

Was at the Black Lips show tonight at Emo's and something incredible happened...

http://search.twitter.com/search?q=bboebel+black+lips

GZA rapped for like 30 min with Black Lips backing him with wu-tang beats, then the Black Lips did their show.  They ended by announcing a combined album.

100% Time

I just walked out of a fascinating panel at SXSW Interactive by the Meetup guys, about how their company's R&D culture evolved from fun/fast/startup-mode to, boring/slow/process-oriented, and then back to fun/fast/"self-organized".  They call what they have now "100% Time".

Here are my unedited, loosely organized notes...

Great entrepreneur/startup culture first few years. One day announced "we're no longer a startup and we need processes". Within a year, Product and R&D were at each other's throats and engagement had plummeted. The processes had made their business like every other business and crushed their company culture.  Leadership had 2 week offsite to figure out what to do. Could not decide. Decided to hold a 6 week Hackathon to more buy time to figure out what to do.

Held company meeting to tell everyone to put current projects on hold. Told them to pick a 6 week Hackathon project.  3 rules:
- Project must add value
- Must convince 3 people to work on project w you (1 product, 1 frontend dev, 1 backend dev, 1 support)
- Must impress your peers

First day, people pitched their ideas to the company, including leadership.  Leadership tossed out their feedback and then walked out and went to see a movie.

Teams then formed. Teams commandeered conference rooms and other spaces. Teams got to work.  They got more built in those 6 weeks than prev year. Staff's engagement level was all time high. It was a clear choice to keep this "self-organizing" team structure.

Every so often they have a deadline and slip structure back in, and it always turns out to be a mistake. They learn this lesson frequently as structure creeps in.

On usability:
The theory that Developers/Engineers can't understand humans and need product and design processes to drive usability is false. Typically developers/engineers just don't have access to the users. At Meetup Developers/Engineers routinely watch normal users use their product, and it us very humbling for them and the usability of the product improves because of this.

On hiring:
- To get things done, their developers have to be able to work with others. People who just want to sit alone and hack will not get anything done, since a rule is you must work on your project w 3 others. Makes hiring really difficult. But finding the right people that fits this mold pays off.
- interview process: written test, then meet w 4-5 people from all sorts of groups, then meets w leadership. Takes Joel Spolsky approach where one maybe = No

On management:
- Tried peer-to-peer reviews. Created too much work. Now peers are asked to voluntarily give feedback to an employee's manager 
- everyone has a manager. Managers are essentially the top technical guys. Managers are also on dev teams themselves. People on the same team do not have to have the same manager.
- everyone works at their office. Not remote.

They encourage lots of communication channels. Small company. 60 people.  Their office is mostly one big open space. Lots of overhearing.

Interesting comment... "Leadership can't create a company culture. It rises out of who you hire. Hire entrepreneur startupy people and you'll have and entrepreneur startupy culture. Your job as leaders is to not squash the culture."

Q&A:

How do you resolve product vision conflicts?
- we're extremely transparent, so everyone understands the vision, sales/finance, product adoption, etc

Do you have a product roadmap?
- no. This made it really hard to raise money. When people ask, they make shit up. Normally it doesn't matter.

Do you have a company wide development methodology?
- no. Teams pick their own methodology.

Do you have any dev metrics or goals?
- nothing company wide. Each team tells us how we should measure their success. And I spend a lot of time asking the teams questions and offering suggestions as they design, develop and set milestones.

Release Cycles?
- week releases every Tuesday
- QA person signs off

Do you think this model would work if you guys had Customers, not Users?
- I don't know.

Do you think this model would work for a much larger organization?
- we don't know yet.

Have you written about this?
- no. But here are 2 books we read after we started doing this:
Maverick
Seven Day Weekend

Your email got to me fast

Bill: your email got to me fast because you're sitting next to me
Beth: email doesn't work that way
Bill: where did you learn that?
Beth: Oprah

Too many distributed key/value store projects?

I seem to hear about a new distributed non-relational data storage project every few weeks.  While I'll admit that all of them are pretty damn cool, most of the coolness is in their roadmaps.  Richard Jones posted a great overview of many of these open-source projects last week.  Given the fact that there are so many projects each with so much more to develop, and the fact that most of these projects are still early-stage, many will die off before they produce anything you can actually use.  I'd love to see some of these developers get together and start merging their efforts.  What do you think?

Auto bailout

If not rail, then maybe cloud infrastructure.

Revive dead business or Create better business?

Why don't we take the $25 billion, $50 billion, $100 billion, or whatever it is that automakers are begging congress for, and instead of helping them keep their dying business models alive, invest the money in building a kick-ass US rail system like Europe and China have?  This would be better for US workers because it would create new sustainable jobs to replace their inefficient union jobs that still may vanish even after a bailout.

Too big to fail

If a company is "too big to let fail", such as Citi, AIG, GM, shouldn't we not let them get that big in the first place? I'm not sure they actually are too big to let them fail... But if the government thinks they are, then isn't the government negligent for putting us at risk by allowing these companies to get this big? Do we need anti-monopoly style laws to break up these giants early so that some can fail without government bailouts? If not, then these companies aren't too big to let fail.