Tim Bauer’s Running Thoughts

This discussion between Scoble and the head of Live Search at MSFT (Brad Goldberg) was rather uneventful so I am not going to do a typical full writeup. For those are still intrigued, here is the URL:

Link to Webcast

And a nice wordle.net tag cloud of the notes

Hope if helps

*************************************************
****** START OF RAW SCRIBBLE TAKEN WHILE RUNNING ******
*************************************************
*************************************************

http://www.fastcompany.tv/video/microsoft-search-will-it-use-mahalo-techniques-compete-with-google
Brad goldberg (GM of MSFT Search Business - Aka Windows Live Search)
……………..○ 6/16/2008, 5:37 AM
……………..○ Big announcement on cash back last week
……………..○ Query stats
…………………………….§ 40% queries unanswered
…………………………….§ 50% refined
……………..○ Targeted pages for key words
……………..○ Content … that is what they do well
……………..○ MSFT has ability - software, software @ scale, long term investment (ie servers)
……………..○ 6/16/2008, 5:40 AM
……………..○ Status
…………………………….§ 1st Goal … get relevant
…………………………….§ Quality of search - Believe it same across yahoo, google, msft
…………………………….§ Key goal how to get relevance high % on 1st page
…………………………….§ Up till now focused on basics of relevance
…………………………….§ Now cash base
……………………………………………□ 4 tasks in search
…………………………………………………………..® Entertain, buy, reference, navigation
……………………………………………□ Commercial focus … task basis
……………………………………………□ Niche focus
……………..○ 6/16/2008, 5:43 AM
……………..○ Other fundamentals - Speed, design, etc
…………………………….§ In the US, On par w/ relevance … in US … not outside of US
…………………………….§ UX - Biggest change in next 5 years (no change in last 5 years in industry)
……………………………………………□ Blended search
……………………………………………□ Universal search
……………………………………………□ Their advantage … not driven by AD $ … so they can take more risks
…………………………………………………………..® Bauer Comment - Google does limited A/B testing … not sure if that is accurate
……………..○ 6/16/2008, 5:46 AM
……………..○ Mobile, Local
…………………………….§ Two goals
……………………………………………□ PC Search
……………………………………………□ Mobile Search
…………………………….§ Bauer Comment - I am not sure if in the 5 year view this is a wise play
…………………………….§ Gas finder / prices
…………………………….§ Tell me - voice based search
…………………………….§ Two main mobile use cases
……………………………………………□ Mapping
……………………………………………□ Business information
……………..○ 6/16/2008, 5:47 AM
……………..○ How change the game? w/ Google.
…………………………….§ A few things
……………………………………………□ Great relevance
……………………………………………□ Commerce (i.e. cash back … reward for search behavior)
…………………………………………………………..® Bauer Comment - better use of their 3B $ they offered Yahoo … buy the eyeballs … a lot of market share is comfort … have to break that
……………………………………………□ Advertisers
…………………………………………………………..® Cost per action (defined by merchant) focus is ADS
…………………………………………………………..® Afraid of monopoloy … no choice (just google) … europe examples
……………..○ 6/16/2008, 5:51 AM
……………..○ Brand
…………………………….§ Took his dad 2 years (scoble) to switch from alta vista to MSFT
…………………………….§ Split search query types
……………………………………………□ Users are not aware (via toolbar embedded in PC or site) of engine
…………………………………………………………..® So they need a better search engine
…………………………………………………………..® HP deal last week as an example
…………………………………………………………..® Last few years not out there
……………………………………………□ Users are aware of search engine
…………………………………………………………..® Users only aware of google
…………………………………………………………..® Really they use 4 major engines
…………………………………………………………..® How create awareness
………………………………………………………………………….◊ Commercial search, cash back … provide reasons to try it on an ongoing business
………………………………………………………………………….◊ Then drive awareness (homepage, OEM, marketing)
………………………………………………………………………….◊ How much engagement (return usage)
…………………………….§ Bauer Comment - They need to run a lotto … ticket per search … gaming it would be a problem
……………..○ 6/16/2008, 5:54 AM
……………..○ People search
…………………………….§ How to integrate to search experience as social networks emerge and people want to find people
…………………………….§ Most of experience on people search is hard (per him)
……………………………………………□ Bauer Comment - Disagree … if you know 1-3 things about the person … you have a pretty good lock in on google
…………………………….§ Facebook … walled garden … does MSFT’s 10% stake give them special access?
……………………………………………□ Bauer Comment there is a ok result on facebook profiles
……………..○ 6/16/2008, 5:57 AM
……………..○ Media search … hard to get results that are media
…………………………….§ Two pieces to that
……………………………………………□ Video search
…………………………………………………………..® Built a browser summary of a video … quickly navigate from a set of videos
…………………………………………………………..® How create traffic … search on videos link
………………………………………………………………………….◊ Bauer Comment - Really they need one search … that dynamically adjusts
………………………………………………………………………….◊ Scoble agrees w/ that point
…………………………………………………………..® At some point there will be break through on UX to get image search going
………………………………………………………………………….◊ Bauer Comment - agree that image search on google works for me
…………………………………………………………..® Not that exciting if you don’t change the UX based on content
……………………………………………□ Weather
…………………………………………………………..® Informational query (answer on 1st page)
…………………………………………………………..® Bauer Comment - Semantic search sounds probably here
……………..○ 6/16/2008, 6:01 AM
……………..○ Back to usability
……………………………………………□ All searches stop @ 1000 results (no matter how many hits they have
……………………………………………□ Over time … users will look on a task basis
……………..○ 6/16/2008, 6:01 AM
……………..○ Mahalo
…………………………….§ Whats experience been w/ that? (asking Scoble)
……………………………………………□ He answers Google, Yahoo, MSFT .. Not satisfying wants a portal
……………………………………………□ Better experience
……………………………………………□ Head oriented query
…………………………….§ Doesn’t scale even w/ crowdsourcing
…………………………….§ 2 Styles
……………………………………………□ Portals - Human programmed experience
……………………………………………□ Search - Algorithmic result
……………………………………………□ Bauer Comment - Good point … he knows the focus of his tool … MSN is the portal
……………………………………………□ Bauer Comment - Google’s style of search types might be the wisest … blog, image, news, etc … basically categorizing pages by style of author … you get a better sense of where to present that (or allow user to choose)
…………………………….§ Ideal model has mix of algorithmic + programmatic
……………..○ 6/16/2008, 6:05 AM
……………..○ Friendfeed push
…………………………….§ New data type … activity streams
…………………………….§ Friendfeed indexes real time
…………………………….§ Google, MSFT … indexes hourly … so cutting edge focus
…………………………….§ What was your experience w/ that (his ? To scoble)
……………………………………………□ Good. For real time.
…………………………….§ How much will users segment searches by search portal focus
……………………………………………□ Friend feed - current breaking news
……………………………………………□ Mahalo - Trips
……………………………………………□ MSFT -
……………………………………………□ Google
……………………………………………□ Paid views for search … increasing focus on navigation … not discovery … sites become primary way to find other sites
……………..○ 6/16/2008, 6:08 AM
** END RAW SCRIBBLE TAKEN WHILE RUNNING **

Didn’t even break a sweat on these two … too short … here was the skinny:

Interesting … but nothing groundbreaking … having a bit more time …. I wandered over to Google.

Another snoozer.

The biggest lesson from these two … I shouldn’t of strayed from my rule that it has to be a presentation or a high end interview. Adhoc interviews (both of these) are more ‘hit-n-miss’ in revelations or insights. People take more time preparing for conferences. Quality of content is better.

* START OF RAW SCRIBBLE TAKEN WHILE RUNNING **

http://www.techcrunch.com/2008/06/01/bezos-talks-web-services/
Also mentioned on Nick Carr
……………..○ 6/2/2008, 6:26 AM
……………..○ Complexity inside … network and apps
…………………………….§ Bulid interface between groups
……………..○ Services solution
……………..○ 4 years ago
……………..○ 2 years to build services
…………………………….§ Simple queue
…………………………….§ Storage
…………………………….§ Ec2
……………..○ People are using those services
……………..○ 6/2/2008, 6:27 AM
……………..○ Why book retailer innovating here
…………………………….§ Retail is a low margin business
…………………………….§ Efficiency is key to their low margin business
…………………………….§ So their culture might of drove it
……………..○ 6/2/2008, 6:28 AM
……………..○ Why do cloud computing … per wall street
…………………………….§ These services drive their businesss
…………………………….§ They are jut doing it in a manner to share
……………..○ 6/2/2008, 6:29 AM
……………..○ What is next
…………………………….§ Persistent storage in ec2
……………………………………………□ Limited beta
…………………………….§ Cloud computing webservices for startups .. Grateful
……………..
…………………………….
……………..
http://techbays.com/2008/05/29/google-io-android-qa/
6/2/2008, 6:30 AM
……………..○ 2nd half this year is release plan
……………..○ www.AndroidCommunity.com
……………..○ Buy device or retro fit?
…………………………….§ Software opensource …so anyone can use it
……………..○ High use of touch
…………………………….§ Platform is very generic to UI
…………………………….§ Showed a touch screen
…………………………….§ Could of done w/ trackball
……………..○ Touchscreen demo … detail
…………………………….§ Vendor did it … NDA
…………………………….§ Details … qualcomm processor
…………………………….§ Synaptics touchscreen
…………………………….§ Memory …128mb of ram … 256 on flash
……………..○ 6/2/2008, 6:34 AM
……………..○ What was shown first?
…………………………….§ Notifications
…………………………….§ Desktop launcher (spatial representation of workflow)
…………………………….§ Compass / GEO location
…………………………….§ Unlock via user gestures
…………………………….§ PacMan
……………..○ 6/2/2008, 6:34 AM
……………..○ Google will provide?
…………………………….§ Infrastructure to OEMs
…………………………….§ Developers will range from free to buy
……………..○ 6/2/2008, 6:36 AM
……………..○ Who was represented in team
…………………………….§ A lot of interest … globally
…………………………….§ Collaborative in many countries
……………..○ 6/2/2008, 6:36 AM
……………..○ Why in house till now? When open … what open/closed?
…………………………….§ Usually in the middle
…………………………….§ Now a corporate sponsor is different (usually organic)
…………………………….§ Important to have a open stack in industry in mobile
…………………………….§ When handset can run stack well … they will release
…………………………….§ What liscense?
……………………………………………□ Contribute intellectual
……………………………………………□ Apache liscense … use software … any pieces … contribute modifications back is optional
……………………………………………□ Linux stack is example
…………………………….§ Everything shown to date is open source
……………………………………………□ Exception … google services being used … not long term block … roadmap issue
……………..○ 6/2/2008, 6:39 AM
……………..○ Platform … Envision carriers to allow people to hack their phones? Pose problems for them (mgmt, reliability?) carriers or parents.
…………………………….§ His answer … why not?
…………………………….§ Current platform is 20 years old
……………………………………………□ Use lowercase l instead of 1’s
…………………………….§ Security gets better (in options) in platform
…………………………….§ Bauer comment — Still I think the point is valid. This will be highly hacked. Blocks on bundles getting broken (like AT&T and Usage of IP ) will be hard … just by # of developers

** END RAW SCRIBBLE TAKEN WHILE RUNNING **

As I watched Eric Schmidt, CEO of Google, speak on this webcast visions of Barney (the dinosaur version of Mr. Rogers) singing I Love You, You Love Me, We’re A Happy Family kept running through my head. So much stroking was going on between IBM and Google I couldn’t tell who was courting who.

Amused and curious on what this relationship was all about, I hit slapped on the headsets and started to listen.

I could go on but this webcast was full of fluffy thoughts that didn’t always jive so well together or were overly oiled. I was hoping for more of a detail analysis of their current product stack in the cloud space. Oh well, perhaps next time.

Would be curious on your thoughts … as always.

* START OF RAW SCRIBBLE TAKEN WHILE RUNNING **

• http://www.johnmwillis.com/other/eric-schmidt-ibm-and-clouds/
• Via RedMonk
• 5/1/2008
• IBM Business Partners Leadership Conf
• 5/21/2008, 6:00 AM
• Eric Schmidt
………………..○ Love in w/ the arch / tech team @ IBM
………………..○ Thoughts on distributed computing
………………..○ IBM is ahead of Google on Cloud issues
………………………………….§ Bauer thought - interesting statement … ponder that if they are ahead … why isn’t that shown in other ways
………………..○ 5/21/2008, 6:02 AM
………………..○ Models of Innovation (4 in the world)
………………………………….§ Europe, China, India, US
………………………………….§ All share success based on Internet
………………..○ RICE supply, Fish supply by mobile connections … short … due to pressing middlemen
………………………………….§ IBM figured out the platform of cloud
……………………………………………………□ Bauer Comment - isn’t IBM a middleman on their primary revenue (Services?)
………………..○ 5/21/2008, 6:05 AM
………………..○ Internet growth
………………..○ Moore’s law … creightors (sp?) law on disk capacity doubling every 12 months
………………………………….§ Bauer Comment - notice how he doesn’t acknowledge AMZN
………………..○ Open approach to internet … drives scaling (centralized)
………………………………….§ Bauer Comment - amusing … IBM is NOT open in many ways
………………..○ 3 tops books in Japan delivered 1st on mobile devices … then printed later
………………..○ 5/21/2008, 6:08 AM
………………..○ Shift from desktop to network centric (centralized)
………………………………….§ Bauer — Why should anyone be surprised by IBM backing centralized (that is their model - Big Iron). What is strange(but reality) is that Google is a centralized backer. Is the battle between peer-to-peer and centralized? Is MSFT the new robin hood w/ Mesh (allowing control to stay local)
………………..○ 5/21/2008, 6:09 AM
………………..○ Convergence isn’t about one device - its one server in client … access point by user can differ
………………………………….§ Bauer Comment - if computing becomes so critical and centralized in cloud … why wouldn’t you regulate and have the government take it on to insure cost control
………………..○ Big Switch Book reference (no reference directly)
………………..○ 5/21/2008, 6:12 AM
………………..○ Cost of Innovation is lower … Consumers drive this … Corporations not quick enough … caught
………………………………….§ IBM good job in app space doing this
………………..○ 70M blogs … average blog has one reader …
………………..○ 7M photos uploaded per day to Picaso
………………..○ 10 hours of video to YouTube every minute
………………..○ Drives integrated story telling
………………..○ 5/21/2008, 6:14 AM
………………..○ Cloud computing does matter to enterprise
………………………………….§ Virtualize
………………………………….§ VMWare good leader
………………………………….§ Email
………………..○ Google collaborative meetings … real time edits to shared doc
………………..○ You can ask customers thing you couldn’t before
………………………………….§ Trial a product
………………………………….§ Not a 3 yr / 1yr plan
………………………………….§ Changes on Fly
………………..○ 5/21/2008, 6:17 AM
………………..○ Talk of Google apps … Integration of Stuff into it
………………………………….§ Docs
………………………………….§ Calendars (cross link, search and cross update)
……………………………………………………□ Bauer Comment - he keeps saying ‘my point is’ but I think he is moving around w/o a central theme … more like a adhoc talk than a planned message
………………..○ 5/21/2008, 6:18 AM
………………..○ Talk of Google Android
………………………………….§ GPS, Phone, Camera, IM … those 4 apps intertwined
………………………………….§ Friend finder … helps you get to them quickly
………………………………….§ Example of cloud computing in summary
………………..○ IBM doing infra (linux / java)
………………………………….§ Got the headstart
………………………………….§ WPS … embed … into applications
……………………………………………………□ Business context of collaboration between google gadgets and IBM customers
………………..○ “The point here is”
………………………………….§ Challenge for developer
………………………………….§ Wifi enablement of devices
……………………………………………………□ Bauer Comment - consumer focus largest play
………………………………….§ Dev must anticipate large data sets … video etc
………………………………….§ Distributed data storage
………………..○ 5/21/2008, 6:23 AM
………………..○ Flip back to intro point … “Sam calling him”
………………………………….§ What Google and IBM needs to do
………………………………….§ No one understand cloud computing … google, IBM, university
………………………………….§ Build a version of this architecture … give to Universities
……………………………………………………□ Bauer Comment … seeding the next generation …
………………………………….§ Announced Open Source Cloud Computing for public in Oct 07
……………………………………………………□ Did @ Univ Wash, MIT, Berkley, Maryland, Bejing, Global, etc
……………………………………………………□ IBM gear, Tivoli, Google Apps
………………………………….§ 19 year old physics … modelled big bang on it
………………………………….§ European developed largest collection of prime #’s every
………………………………….§ Out of control … hand off to national science foundation on who gets CPU time
……………………………………………………□ Bauer Comment - making a big deal out of a capped infrastructure being donated to public … really many do this … probably at higher investment rates
………………..○ 5/21/2008, 6:26 AM
………………..○ “key Idea here”
………………..○ Computers are not alone, You can build these cloud platforms and get great value
………………..○ End of keynote … switch to panel Q&A
………………..○ 5/21/2008, 6:28 AM
………………………………….§ Panel Q&A - Eric / Sam
………………………………….§ How do Google / IBM collaborate given differences
……………………………………………………□ Sam - Business vs Research … debate on cloud (face to face, counter eric’s point of collaborative doc change)
……………………………………………………………………..® Bauer Comment - interesting .. I have always thought meetings are useless unless some asset is created or evolved (just talk … nets to little)
……………………………………………………□ Sam (still) - Who is the best partner .. .standards .. Aligned with strategy … culture of teams .. Committed to open source … google idea to go to academic community (eric said it was theirs)
……………………………………………………………………..® Bauer Comment - interesting … as I think going to the university can be a sticky wicket (bad PR) if perceived wrong
……………………………………………………□ 5/21/2008, 6:32 AM
……………………………………………………□ Eric - All innovation comes from bottom up … compatible culture
……………………………………………………………………..® Culture was similar
……………………………………………………………………..® Other companies were top down … too legal
……………………………………………………………………..® Sam - IBM has a lot of lawyers
………………………………….§ 5/21/2008, 6:35 AM
………………………………….§ Tangent - Love in On Why IBM is champion on focusing on University Arch Spend
……………………………………………………□ Eric - Always has built specialized gear / solutions for University
……………………………………………………………………..® No idea why good business to do so
……………………………………………………………………………………….◊ Bauer Comment - come on. No idea? Eric? No idea? Why don’t you ask Jobs why he focused on schools … easy … hook em early … you get em later
……………………………………………………□ Sam - Somewhat admitted true goal
………………………………….§ How you tactically make the two teams work
……………………………………………………□ Sam - Eric challenged two teams on mapping and shopping solutions. 8 weeks. Created prototypes.
……………………………………………………□ Eric - Show the future. Talking to people. People tend to agree but not back. Tries to use stuff even if he hates it. Immerse in it. Use the product. Sit next to it.
……………………………………………………………………..® Bauer Comment - Howard Hughes flash back.
………………………………….§ How does intersecting worlds of Enterprise and Consumer work
……………………………………………………□ Sam - Most things start in consumer space … lower risk of deployment. Tech view … is there a logical way to scale and get virtualization … while protecting the transaction … place to learn and then apply elsewhere
……………………………………………………………………..® Bauer Comment - Good point … shared gear … higher risk of breach in cloud
……………………………………………………□ Eric - Not much difference between consumer and enterprise … 1 key … enterprise will pay for some features … enterprise will pay for security and up time … traditional data centers offsite … IBM specialized … Google mini-me free for unsupported use
………………………………….§ 5/21/2008, 6:42 AM
………………………………….§ Local relevance … versus centralized
……………………………………………………□ Sam - Create cloud of business product … ERP .. HR .. Procurement … Not by region at IBM .. So they run in cloud … so it allows the business to focus on
……………………………………………………………………..® Bauer Comment - He isn’t answering how they stay locally relevant (sales relationships … delivery people on floor). I think this an even bigger problem for Google
……………………………………………………………………..® Eric - Doing Accntg and Payables (IBM is) for them
……………………………………………………□ Sam - Still - Have to deal w/ countries differently … China Out, MiddleEast Out, etc … basis of that … IBM tries to connect “IN” … start w/ societial model then embed in economic expansion … Leneuvo (sp?) deal in china … example
……………………………………………………□ Eric - Simple rule … people around the world want same stuff … family, material items, safety, healthy … only difference is language and culture (smaller factor)
……………………………………………………………………..® Bauer Comment - They also want to remove the middleman … which Google and IBM are in many situations
………………………………….§ 5/21/2008, 6:48 AM
………………………………….§ Social Network differences
……………………………………………………□ Eric - Impact business. Still too closed. Need to be more open.
……………………………………………………………………..® Bauer Comment - A ding on Facebook there
……………………………………………………□ Eric - Still - Real phenomena … this isn’t kids online in college … a large demographic is shaping relationships in this model
………………………………….§ Close

** END RAW SCRIBBLE TAKEN WHILE RUNNING **

For those not tracking the geek world, David Heinemeier Hansson is a rather successful coder that built some well adopted apps (e.g. Backpack) and a framework that got some people excited about its ease (Ruby on Rails).

O.k. So what …

Well, DHH went on a rant about how the new business model of FREE that everyone is jumping up and down about. Specifically saying how it was nuts. In his view, “real” business startups (not those looking to get bought by some big fish) should focus on bringing value that customers pay for … not how they can be free and gain adoption from that.

Novel thought … but why is he pointing this out?

As usual my raw notes are below, but here are a few key themes that jumped out me as I watched this webcast over lunch

In summary, gotta give hops to DHH for pointing out the frenzy in business today to launch businesses based on models that are not clear (just build membership) especially in a setting where VC’s are presenting before him. I think he is right to call for a refocus at the individual level especially as he talked about focusing on a niche and delivering to that niche. Much like the line Security through Obscurity … you can protect your business concept by starting small in niche markets. If you go after the mega models (like the next FriendFeed … you have big fish to swim with and will likely get eaten).

** START OF RAW SCRIBBLE TAKEN WHILE RUNNING **

http://omnisio.com/startupschool08/david-heinemeier-hansson-at-startup-school-084/21/2008
11:40 AM
……………○ 4/21/2008, 11:41 AM
…………………………§ 37 signals
………………………………………□ Not hiring
………………………………………□ Not looking for VC $
………………………………………□ Makes them different than most businesses
…………………………§ Their focus - how to be viable outside of big dogs getting bought out
…………………………§ New focus - Don’t focus on the $
……………○ 4/21/2008, 11:43 AM
……………○ A secret to making $ Online
…………………………§ Great application
…………………………§ ???
…………………………§ Profit
……………○ ??? = Price is key
……………○ Point is web 2.0 thinks everything is free … not true
…………………………§ BAUER COMMENT - BASECAMP and their model is paid … so his point is VC $ makes free ok
……………○ Simpler way to build a business
…………………………§ Have a price
…………………………§ Signup
…………………………§ Working 4-5 years … multi-million $ business
……………○ 4/21/2008, 11:45 AM
……………○ Shows campaign monitor (email service, track who clicks on what)
…………………………§ $0.01 per recipient
……………○ 4/21/2008, 11:46 AM
……………○ Shows FogBugz
…………………………§ On Demand
…………………………§ Buy the software
……………○ Show FaxItNice
…………………………§ Route around the fact that many businesses work w/ fax
…………………………§ Wrap it to a nice webservice
…………………………§ Pricing options
………………………………………□ $5 per fax,
………………………………………□ $20 retainer then a per send
………………………………………□ Send and receive faxes monthly
……………○ 4/21/2008, 11:47 AM
……………○ Not rocket surgery
…………………………§ Still hard but easier
…………………………§ Most businesses fail
………………………………………□ So why not try to be the next Google
…………………………§ Bad logic in his view
………………………………………□ Odds are not the same for barn burner … odds for a small niche product is better
…………………………§ Odds
………………………………………□ 1:10 is not facebook creation odds
………………………………………□ Example of relevance
……………………………………………………® 1:10 of making a million $
……………………………………………………® 1:10,000 for a billion $ business
…………………………§ Take better odds at smaller reward 1st … then go for the moon as the change in lifestyle from 0 to $1M and $1M to $1B is notable (the latter being small)
……………○ 4/21/2008, 11:52 AM
……………○ Example
…………………………§ 2000 customers
…………………………§ $40/month
…………………………§ 12 months
…………………………§ $1M year
…………………………§ 40,000 signups on trail … 5% conversion … 110 a day
…………………………§ …
…………………………§ What if you were happy @ $200k / yr = 400 customers @ $40/month
………………………………………□ BAUER COMMENT SO IT GETS EASER
……………○ 4/21/2008, 11:54 AM
……………○ Finding customers
…………………………§ Backpack
………………………………………□ Hard to get consumer to get people to pay
………………………………………□ Hard to establish business w/ that segment (consumer)
………………………………………□ Relaunch backpack 2 months ago … doubled the revenue by focusing on businesses
…………………………§ The Fortune 5,000,000
………………………………………□ The lifestyle business (mom & Pop store?)
………………………………………□ People make fun of that but he calls BS
………………………………………□ There is a lot of room in between
……………○ Craigslist
…………………………§ Without taking VC $ allows you to make the shots
…………………………§ Run @ your pace
…………………………§ $1M / yr … like no meetings
…………………………§ So its satisfying to not go for the brassring
…………………………§ Craig Newmark - $1B people are not happier
……………○ Good cause is incredibly hard and time consuming
…………………………§ Why flip business?
…………………………§ Enjoy it over 20 years
…………………………§ He questions the sell out startups … is it really the good life?
………………………………………□ Trade passion for meetings
……………○ 4/21/2008, 12:03 PM
……………○ This is not the movie industry
…………………………§ No need to dominate the box office
…………………………§ Solve small simple problems
…………………………§ Most of the great companies started small (200 companies)
…………………………§ Don’t Model after Facebook
……………○ VC Fuzz
…………………………§ Forget viral
…………………………§ Forget the overbaked ideas
…………………………§ Treat your customers nice and ask for $
……………○ Pressure to get started
…………………………§ People are selling stuff way before your idea
…………………………§ Sell it better
……………○ 4/21/2008, 12:06 PM
……………○ Basecamp
…………………………§ Developed w/ 3 people doing other stuff
…………………………§ Single server 1st year
…………………………§ Limited amount of time to work on something focuses your energy
…………………………§ He was contracting w/ 37 signals at this time … 10h/wk to develop basecamp
…………………………§ It makes the time really matter
…………………………§ Less time is a huge benefit
………………………………………□ BAUER COMMENT — Power comes from focus knowing value is each hour
…………………………§ Grew basecamp a year before you stopped doing consulting and just focused on basecamp
…………………………§ 7+ years for great businesses maybe longer
……………○ Don’t Overbuild
…………………………§ Deal with those issues later
……………○ Take it easy
…………………………§ There will never be less work
…………………………§ It will never be less
…………………………§ Amount of work in beginning creates more work
…………………………§ 14 hour days / 7 days a week … get out of that
…………………………§ Change the practices … or they will stick with you
……………○ 4/21/2008, 12:10 PM
……………○ Blame loss on someone else
……………○ 4/21/2008, 12:10 PM
……………○ Q&A
…………………………§ How does great achievement drive from small goals (i.e. how did Ruby on Rails happen)
………………………………………□ Solve simple problems. RoR was his effort to enjoy coding. Solve simple solution.
………………………………………□ Tons of VC come to them (RoR) .. .they say no … he uses RoR to build stuff … so he understands it …
………………………………………□ Solve simple problems that you are intimately involved with
…………………………§ What would you have done differently (BaseCamp, RoR)?
………………………………………□ Not much. Happy with results. Happy didn’t listen to give it away for free.
…………………………§ How many businesses are in the sweet spot of the fortune 5,000,000
………………………………………□ 500 to 1000 people … that is huge
………………………………………□ His focus
……………………………………………………® 3 people
……………………………………………………® 5 people
……………………………………………………® Guy doing something in his spare time
………………………………………□ Those people … want something that sounds like it works
…………………………§ The is no business too small for fortune 5,000,000
………………………………………□ Deal w/ people on monthly … you are not relying on 1 customer
………………………………………□ Big customers can yank your change
………………………………………□ Small customers but many you can control innovation
…………………………§ How stay focused
………………………………………□ How do you not get distracted … 10-14 hours a day
………………………………………□ Parent filter …
……………………………………………………® Joke
………………………………………□ Stop working in 14 hour a day … work 5 hours a day
………………………………………□ 37 signals is 4 day work weeks … 8 hrs a day … 2-3 hours productive
………………………………………□ Get 3 GREAT hours a day in

** END RAW SCRIBBLE TAKEN WHILE RUNNING **

Social networks want to portray themselves as your friend in ‘discovery’. Your wingman in serendipity. However, aren’t they really more like new age Drug Dealers? I say that because all of their models for making money from their service (ads) incent them to create addicts. So, in their view, a daily touch isn’t enough. Forget that. How about hourly? Minute to minute? Ok. Maybe I am being a bit harsh. I got this idea as I watched this presentation from the Graphing Social Patterns (GSP) conference in March. I initially was intrigued because it was a panel of faces I knew and some I didn’t on an area that ties to my thoughts as of late. As usual my raw notes are below, but here are the thoughts that jumped out at me as I ran (i.e. jogged) along.

Basically it boils down to insight that the social networks aren’t enabling true usability on average in the use case of searching for insight. They are really focused on a use case that maximizes views and interaction. At least most of them are. A weekly feed by topic, friend (trusted source) is what they need. ** START OF RAW SCRIBBLE TAKEN WHILE RUNNING **

A panel discussion at GSP West with Ian Kennedy (MyBlogLog), Bret Taylor (FriendFeed), Kevin Marks (Google) and David Recordon (Six Apart) on the rollout of shared activity streams as part of the latest revolution in social networks • 4/11/2008, 6:08 AM • Fred Wilson - Facilitator
…………..○ Single search box.
…………..○ Front page. Genius. Wallstreet newsfeed at facebook drove facebook success.
…………..○ Commentary feeds and aggregation is key • 4/11/2008, 6:09 AM Ian Kennedy - MyBlogLog
…………..○ Goto othersites and pull information • 4/11/2008, 6:10 AM Kevin Marks - Google Open Social
…………..○ Activity streams
…………..○ Bauer comment - Much more technical, buzz wordish
…………..○ Was at technorati for 3.5 yrs prior • 4/11/2008, 6:10 AM david recordion - six apart
…………..○ Api focus for him
…………..○ How interact social networks • 4/11/2008, 6:11 AM Bret Taylor - FriendFeed
…………..○ Share stuff from web
…………..○ UI of their system is key for their growth … info discovery tool • What is the benefit to early adopters
…………..○ Rebroadcast what you are doing
………………………..§ Flickr
………………………..§ Twitter
………………………..§ Mass of links
…………..○ Bauer — Each person you know has the ‘mass’ and ‘mess’
………………………..§ Make it dynamic
…………..○ DR - enable your posse to follow you in more richly
………………………..§ Digg, flickr, etc
…………..○ BT - Not lifestreaming … content discovery
………………………..§ Too much information on stuff
………………………..§ See all these angles on an event
………………………..§ Use people you know for a filter on an event
………………………..§ So they reverse to an event rollup … then your posse’ view on that event
………………………..§ Bauer comment - techmeme but filtered to your trust network
…………..○ KM - agrees w/ BT
………………………..§ River of news • 4/11/2008, 6:16 AM • Lifestream
. Public versus Private … commenting
…………..○ BT -
………………………..§ APIs, Feeds
. Give information that is public … too obscure to be relevant … flickr default is public for photos
. Most people don’t care to shut off
……………………………………..□ Bauer comment - security through obscurity is what he is saying goes away due to APIs
………………………..§ Risk of exposing friend private responses to a broader group … their view of your posse vs your view of the posse
…………..○ DR - Public to Not Public Hard
………………………..§ Doesn’t have a method to share Facebooks activity out
………………………..§ Vote on sharing activity stream on facebook
…………..○ BT - Moveable type plugin
………………………..§ His photos in facebook are blackhole for now but expect them to change
…………..○ IK - SSB
………………………..§ Aggregation layers was hard on security … change of relationship impacts what you will or wont share
………………………..§ So they simplified to public only
…………..○ DR - Feeds
………………………..§ Super secret URL feeds get posted anywhere so that doesn’t work
………………………..§ Don’t give other applications PW to your blog • 4/11/2008, 6:24 AM • Need for microformats on blog?
…………..○ KM - Microformats to mark up a feed (RDF knock off?)
…………..○ IK - Add feed from friendfeed
………………………..§ Creates duplications
………………………..§ Use microformat … get UID for event … source item by publisher so various feeds don’t replicate the same item by author • 4/11/2008, 6:25 AM • Duplication - Twitter is an example
…………..○ BT - Atom is a good fix
………………………..§ So they are working on it
………………………..§ Work around to hide 1/2 the twitters (days before mailing list)
…………..○ DR - De-Dup how presented to users, consistent appearance
………………………..§ Flickr feed, giant photo
………………………..§ Talks about aspects of aspects of a noun … blog, abstract, thumbnails
…………..○ BT -
………………………..§ GEO RSS is an example
………………………..§ ATOM Summary is another
………………………..§ Fragmented right now, become ubiqutos • 4/11/2008, 6:29 AM • How do you monetize … both agregators and publishers
…………..○ BT -
………………………..§ Focusing on user base
………………………..§ See a lot of ad $ potential with actual content and social context
…………..○ IK - Affiliate Links
………………………..§ Netflicks
………………………..§ BT - Pieces of content by users ok. But ad affliate links drive conflicts of interest. Want to make the view for the user decides … allow users to choose.
………………………..§ IK - 3rd party though
………………………..§ Bauer comment - nice push by google rebuffed by BT on ads pushed down versus bubble up
…………..○ IK - Sends traffic to core sites that hold people • 4/11/2008, 6:33 AM • What seeing when building open social apps, desire to be in activity streams
…………..○ KW - Very aware. Abstraction There is Important
………………………..§ Competition over next few years how to makes sense of flows of info
…………..○ IK - Drive back to creator site is key
………………………..§ Drive down traffic to creator site kills off creators (who live on ad $?)
………………………..§ Google moving to engagement not presentment
………………………..§ Move away from mashup to interaction to websites … so 10 people that put together the content on a page … get paid
…………..○ BT - Traffic They Use For Redirect
………………………..§ Do you provide enough value for a net win. Driving growth
………………………..§ Blog publisher wants reaction on blog BUT people want to talk wherever they go … so they go to a aggregator site … FF distributes link among friends and talks at friendfeed
………………………..§ Bauer Thought - It is key to be able to pull threads on your product if possible … question is would the conversation occur if they couldn’t move into a non-public blog forum
…………..○ IK -
………………………..§ Googlenews getting blocked back and forth so debate continues by audience. • 4/11/2008, 6:40 AM • Viral Loop To Generate Activity. Pushing Apps To Get Activity. Sponsorship opp?
…………..○ KW -
………………………..§ Risk of being seen as a spammer. Viral apps that is. Feeds like flickr etc is interesting.
………………………..§ Think about ad ad person … how get product awareness via an associated conversation around the product
……………………………………..□ Bauer comment - key point … how companies can market in this model … sponsoring social networks / conversation that lead to their product potentially.
…………..○ DR - Sharing my flight info / Beacon • 4/11/2008, 6:44 AM • Goto Audience
…………..○ When feeds from myspace / facebook
………………………..§ IK - Going to avail now … not invited to
………………………..§ KM - No clear answer. Activity streams are core. • BAUER COMMENT - Only one question - DOH! • 4/11/2008, 6:45 AM • Will the Social Networks Be The filters of attention streams or tools that create and filter (so what choice?)
…………..○ BT -
………………………..§ Facebook announced plan to pull in information
………………………..§ Key is they have a solid graph
………………………..§ Problem @ Facebook is the proper tagging (relationship) of network … dating focus of its past will be a struggle for them
…………..○ DR -
………………………..§ Task of filtering will be the next ‘war’ … how filtering / aggregation is done
………………………………………□ BAUER CoMMENT - vhs/BETAMAX wars
…………………………§ 12 companies are going after FF
………………………..§ Yahoo w/ myBlogLog
………………………..§ Google w/ OpenSocial / Orchid
………………………..§ Plaxo
………………………..§ Etc • 4/11/2008, 6:49 AM • Niche feeds. Wider set of people, different trust levels, only with people w/ this aspect (kids, bachelor, work skill xyz). May choose to ignore niche feed. How to filter.
…………..○ BAUER COMMENT - let the humans do it … weekly feed summary……………○ BAUER COMMENT - They (they players in this) are not incented to do this although feedburner says they are working on it

** END RAW SCRIBBLE TAKEN WHILE RUNNING **

On first blush Google App Engine (released to 10,000 developers yesterday) seems to be just another entry into scalable computing. But on further review there is quite a bit being introduced here that is notably different than the offerings of the early leaders of this space.

You can read the running thoughts at the bottom but here are my initial thoughts on what is unique on this offering when compared to other virtual computing plays.

So at first blush, for rapid prototyping or throw away apps this looks pretty good. But for large scale computing in the cloud I think I will still look for setups where I can host my images (like IBM, AMZN, and MSFT seem to be pondering).

** START OF RAW SCRIBBLE TAKEN WHILE SITTING **

Campfire One: Introducing Google App Engine
4/8/2008, 5:40 AM
…………○ Intro - http://youtube.com/watch?v=3Ztr-HhWX1c
……………………§ Welcome Tom Stocke
…………Campfire One - Intro Google App
…………http://youtube.com/watch?v=3Ztr-HhWX1c
…………duration 9:10
…………
……………………2:20 - Google IO in May
……………………2:35 - kevin gibb (lead for Google App Engine)
……………………–Scalable infra is goal
……………………–Hard to create / deploy
……………………–Typical tasks - Apache server, Config, Tables, Scripts, Code push model, monitoring solutions
……………………–Find machines is 2nd task
……………………4:23
……………………–$
……………………–3rd challenge coming up
……………………–machines break, config issues,
……………………–app grows
……………………–design goals
……………………–1–Easy to use
……………………–2–Easy to scale
……………………–3–Free to get started
……………………–BAUER COMMENT (Free is good)
……………………–We run web apps (not a grid, not a virtual machine)
……………………–They run applications
……………………–Provide for entire lifecycle of web app (request log, application logs, push new version, run db, integrate tasks together)
……………………7:12
……………………–Easier to deploy
……………………–Access to google scaling model … google account, gfs (google file system), big table (?)
……………………–BAUER COMMENT (similar to AWS)
……………………–supports python only right now
……………………–SDK
……………………–Web based admin console
……………………–datastore (big table is thier model)
…………
…………○ Demo … 9m … http://youtube.com/watch?v=tcbpTQXNwac
……………………§ 4/8/2008, 5:43 AM
………………………………□ Uses a Yaml file
………………………………□ create project
………………………………□ looks like a config file (version, api_version, runtime
………………………………□ Handlers .. all code maps to main python file
……………………§ Switch to terminal
………………………………□ Starts web server locally
………………………………□ Works. Build webapp.
………………………………□ Starter app … web app …
………………………………□ Import from google
………………………………□ Define handler
…………………………………………® Get
…………………………………………® Hello msg
………………………………□ Run python with standard python wsgi
……………………§ Run
………………………………□ Stack track from python … line 18
………………………………□ Now fixed
……………………§ Add templates
………………………………□ Library based on jango templates
………………………………□ Import template
………………………………□ Use module to load template
……………………§ Build an app now
………………………………□ Form
……………………§ Put data in data store
………………………………□ Import datastore
………………………………□ Use concept of modelclass
………………………………□ Model class (table) by fields
………………………………□ On post request … create new shout (data) using form
……………………§ Write a query to pull data out
………………………………□ Gql (google query lang?) to pull out data
………………………………□ Create a dictionary to pass values into template
………………………………□ Use jango templates to render
……………………§ Clear data store
……………………§ Putting stuff in now
……………………§ Add sort order to query
……………………§ Extend data model and associated template
………………………………□ From who field added
……………………§ Example of handling data exceptions … formatting to code driven values
……………………§ Now push live to web
………………………………□ Another tool appconfig … deploy immediately
………………………………□ Cmd line
………………………………□ Pushes to live app
………………………………□ From local
………………………………□ Real time data create (tables etc)
……………………
…………○ Part III … Back to Kevin … 13m .. http://youtube.com/watch?v=oG6Ac7d-Nx8
……………………§ Back to 5 components
………………………………1. Scalable servering infra
…………………………………………a) Distributed / low overhead
…………………………………………b) Auto fault tolerant
…………………………………………c) Connect request to running instance in code is their work
…………………………………………d) Fluid in scaling … in response to demand
………………………………2. Python runtime
…………………………………………a) Language support is modular (api, console, etc - all generic so new lang later)
…………………………………………b) No restriction on code (js, css, html, etc)
………………………………3. SDK
…………………………………………a) Written in linux, mac, windows
…………………………………………b) Code, compile, test cycle … faster .. Local
………………………………4. Admin Console
…………………………………………a) Simpler LAMP stack
…………………………………………b) Add users, version running, tools on request logs / app logs, data explorer (vs sql admin), hook up domain, stats near real time, errors, what urls
………………………………5. Scalable Datastore
…………………………………………a) Schemaless object store
…………………………………………b) Supports million entities
…………………………………………c) Not on DBMS … uses BIG TABLE (google gear)
…………………………………………d) Horizontally distributed fault tolerant system
…………………………………………e) New tools, new way … no SQL … may like the system better in long run is their view …since there is no schema …new properties, new table … just code
…………………………………………f) Removes DBMS
…………………………………………g) Datastore model class
……………………………………………………i)
…………………………………………h) Advanced and primitative types supported
…………………………………………i) Data time property
…………………………………………j) Rating property
……………………………………………………i) Gdata Type (system from google) … symantic type … store entity and work w/ it in a generic way via an XML namespace
…………………………………………k) Query power
……………………………………………………i) Sort (single or multiple)
……………………………………………………ii) Transactions
……………………………………………………iii) Batch
……………………………………………………iv) User settable primary keys
…………………………………………l) Gql is like sql … jquery and fbql
……………………………………………………i) Googles sql on big table
……………………………………………………ii)
……………………………………………………iii) More complex
……………………………………………………iv)
……………………………………………………v) Joins not supported
………………………………………………………………One. Joins don’t work well across machines
…………………………………………………………………………First. BAUER COMMENT (linked server?)
………………………………………………………………Two. Allows to scale
…………………………………………………………………………First. BAUER COMMENT (who needs this level of scale? Worth the loss)
…………………………………………m) Email API
……………………………………………………i) No SMTP required … just create mail object and go
……………………………………………………ii) Outgoing HTTP requests
………………………………………………………………One. Use to integrate w/ webservices
…………………………………………n) Auth w/ Google Accounts
……………………………………………………i) Handle users alread
…………………………………………o) Support Frameworks
……………&