Continuations

Unix has some very powerful commands available. For instance, the simple combination of “rm -rf” if invoked by a superuser at the root would pretty much delete every file and directory on the system.  Because such power is available systems are often configured to provide for some measure of protection from accidental disaster (e.g. thinking that you are in some directory way down the tree when you invoke rm when in fact you are in the root directory).

This week I encountered two situations where I would have loved to have a bit more protection from operator error.  The first one involved tungle.me.  I was trying to use tungle.me to set up individual phone screens for 25 applicants, where each of the applicants would be able to pick from the same set of available times.  What I was doing instead was setting up a single meeting with all 25 applicants and the invitation was showing everyone’s email address!  Would have been great if the system had prompted me *before* sending the emails to double check what I want to do.  This could be a one time check with a box that says “Don’t ask me again.”  I recognized my mistake within seconds but it was already too late.

The second case involved Jobscore.  I was using the system to send emails to a large number of candidates to let them know that we were not going to phone screen them.  On the plus side, Jobscore makes it very easy to send notes to a lot of people in one go.  On the minus side, there is no check if you just sent a person the same note a short time ago.  I wound up — by operator error — sending several people the notice twice within minutes.  Again there are clearly scenarios where you do want to send the notice again (e.g. you have updated it), but at least the first time you do this, the system should ask whether you really meant to. 

At some level, neither of these is on the order of a disastrously misplaced rm, but the issue here is that I was causing harm/distress to third parties and not just shooting myself in the foot.  Also unlike rm where you at least might have a backup from which you can restore, there was no “undo” operation available here as the emails had already gone out.

Anyone who does mechanical work appreciates the impact the right tool can have on a job.  For instance, the other day I needed to open up a cover that was held down by six screws.  I first tried a screw driver.  The screws were so tight, they didn’t even budge.  But because the screws had shaped heads I was able to use a wrench instead.  Minutes later the cover was off and I was examining the impeller for a seawater pump.

Well, the same lesson of course applies much more broadly including processes at work.  Yet, we often make do with abusing Excel (or now Google Docs) and/or email resulting in far more effort and breakage (stripped screw heads?) than would be required.  That’s why we chose to go with an applicant tracking system (in our case Jobscore) for our current recruiting effort at USV.  Without it I doubt that we would have been able to handle the volume.

Continuing in the same vein, I wanted to use a tool to help me schedule the roughly 25 phone screens that I will be conducting.  I settled on tungle.me which seemed like a good fit.  Then I promptly screwed up big time by not understanding how the tool works.  My error resulted in exposing some email addresses of applicants to each other.  I have already apologized to those involved directly, but it was a clear case of where I should have spent more time learning and testing the tool first.

Having now done so, I have actually figured out a way to use tungle.me for this purpose (I will write a separate post on that because my use case isn’t supported perfectly).  Once again the net effect was fantastic.  Instead of a lot of of back and forth, I was able to schedule the bulk of the screens with a single email.  Each successfully scheduled appointment results in a confirmation message that with one click transfers the information to my calendar.  I am not sure yet whether I will use tungle.me more broadly, but for scheduling a large number of calls over a condensed period of time it beat email hands down.

This week my blogging is likely to be lighter than usual and my email responses slower than usual.  The reason is that we have move to the phone screen stage of finding candidates for the two positions at USV.  If you are interested / have applied, you can read a more detailed update on the hiring process.

At USV, we have long believed in “Social Objects.”  In the real world when people talk they often form a connection around objects, such as sharing a common interest in a movie.  That was our thesis behind investing in GetGlue.  Alex and the team there have done an amazing job delivering on that idea.  When you have the Glue extension installed in your browser, you can connect with friends and discover new people around “social objects,” such as movies, books, etc.  The power of Glue derives from the fact that it recognizes these objects and so it doesn’t matter on which web page people interact, as long as the page is about that object!

A while back GetGlue started promoting the AB Meta markup as a way for pages to indicate what they are about.  If you look at Facebook’s OpenGraph markup, you will find quite a bit of similarity.  As Alex points out in his great post today over at RWW, both formats RDFa based, are easily human readable and have fairly minimal requirements making them easy to implement.  With Facebook putting the weight of 400 million users behind this initiative, it is likely that a lot more pages will start to add this information.  That’s good because it will bring us closer to the ideas behind the semantic web.

But there are a couple of important issues here.  First, the standard for marking pages up semantically should not be controlled by one company.  It would be great to see Facebook bring others into the process for evolving the OpenGraph protocol.  Second, there is a big agency problem around people marking up their own pages as soon as there is money (or nearly equivalently traffic) associated with it.  There is an incentive to claim that a page is about say a popular movie when it isn’t.  So what is really needed is a way for third parties to express an opinion as to which social object a page is about.

GetGlue actually has done the hard work of recognizing many pages around the web already.  If you want to make your application aware of the interactions around social objects, the Facebook API is not your only choice.  GetGlue also has an easy-to-implement API.

I have spent most of this morning reading reactions and thoughts about Facebook’s “Open Graph” via Techmeme, Tumblr, Twitter and comments to my post and Fred’s post.  All of it has me convinced that the power of the web will be stronger than any one company.  That is with one important caveat: what we can do on the web has to stay as open as it is today!  This is why I believe the real fight is not with or between Google or Facebook or even Apple.  The real fight is for net neutrality and against endeavors such as the finally officially released ACTA (International) and the recently passed Digital Economy Bill (UK).  We all need to let our representatives know that maintaining the freedoms on the net is central to our future.

Hats off to Mark Zuckerberg and the entire team at Facebook.  They are managing that most impressive feat of innovating at scale.  They are also incredibly ambitious in what they want to accomplish.  The goal seems nothing short of one identity and one graph to “rule them all.”  With over 400 million users worldwide and a sign on system that is being widely adopted this ambition doesn’t seem crazy.  Especially when you layer on top of this the possibility that soon many of these users might have Facebook currency that could be used by sites to implement 1-click purchasing (and by venues to enable RFID based payments via Facebook presence).

But I see at least one flaw with this plan for domination.  I simply don’t believe that there is a single social graph that makes sense.  I may very well follow someone’s booksmarks on del.icio.us that I don’t want to have any other relationship with.  Or take the group of people that I feel comfortable sharing my foursquare checkins with — these are all people I trust and would enjoy if they showed up right there and then.  That group in turn is different from the people I work with on Google docs for various projects which is why I would be nervous about using the Microsoft docs connected to Facebook.  Trying to shoe-horn all of these into a single graph is unlikely to work well.

As a little historical aside.  There is a bit of a personal irony in the huge noise around the Facebook “Like” button.  Yahoo could have had this in 2005 following their acquisition of del.icio.us if they had started to promote it to their users and to content sites!

A good friend of mine was one of the early practitioners in Europe of complex financial engineering for large corporate clients.  I vividly remember a conversation with him on a vacation some 20 years ago, when he told me about a deal that was meant to remove energy cost risk from a company and sell that risk off separately.  At the time he made the argument that this allowed for better hedging on part of the company and allowed investors to take a position in this risk alone.  I took the position that any time you do this you are likely to run into huge incentive problems.  A company that thinks they have sold off all their energy risk has no incentive to try to improve their energy efficiency.  A lack of such improvements is likely to bring with it many other things that don’t get improved (e.g., production processes) that will ultimately come back to bite the company!

I have thought back to this conversation many times over the unfolding of the financial crisis.  Given my involvement with innovative companies as a VC, I am generally excited about innovation and have an open mind for financial products innovation.  But it continues to strike me how much of financial products innovation suffers from severe incentive problems.  The latest case in point are the revelations around the creation of CDOs by Goldman Sachs and others for Paulson to bet against.  It will be up to courts to determine whether there was any punishable offense here, but it is clear that there is a profound incentive problem when one party pays a fee to another party to help create and market a product that the initiating and paying party wants to fail!

I don’t know how this can be fixed other than requiring that folks retain a certain amount of risk.  For instance, mortgage securitization itself wouldn’t have gone so horribly wrong if the original issuers had to have retained meaningful default risk.

So I know that there are lots of problems with the Google book settlement. But yesterday I got to experience the power of digitized books firsthand and it was awesome to behold. My son had to do a project for school where he described five different professions from Colonial times. There were quite a few resources available on the web, but many of them lacked any kind of source attribution for their claims about what Colonial life was like. Now instead of going to the library, I pointed my son at http://books.google.com and it was amazing. For instance, one of the professions he was researching was “farmer.” Within minutes he found a book that had an incredibly well researched chapter about Colonial farming. Now here is the really cool part: the book was from 1916! I fully appreciate that the rights situation is a tricky one. But it would be a shame for humanity if this kind of access to knowledge would instead disappear behind paywalls and fractured into many different places. We have to find a solution that keeps open access available!

I have both a Kindle and an iPad and don’t see the latter replacing the former. They solve different problems for me, at least so far. I use the iPad primarily for browsing the web. Around the house (eg breakfast table, in bed) it is much nicer for that than having a laptop open and the Kindle is a non-starter for browsing. When it comes to reading books though the Kindle will remain my weapon of choice. It is lighter and easier to hold (and carry along in addition to my laptop). It has great battery life. And the e-ink display is wonderful for reading longer stretches. On top of it all I can use the Kindle apps on all my devices for moments of reading when I don’t have my Kindle with me. So my Kindle has nothing to worry about from my iPad.

It turns out that seeking advice is one of the hardest things to do - at least it was for me for the longest time. Here is what I have learned from my own mistakes.

First, it is critical to have some people you really trust. People who have good judgment. Who understand you as person and who care about your emotional and spiritual well being (not just your commercial success). There are a great many people who will want to give you advice and you can learn something from most folks, but when it comes to crunch time and actually making a decision, not every piece of advice should be weighted the same.

Second, you have to be willing to provide all the relevant information to your trusted advisors. It is impossible for anyone to give good advice if they don’t have all the information. Holding back, especially bad news, is not a good strategy for seeking advice. One way you can tell if someone is to be trusted as an advisor (see above) is if they actively seek information that’s relevant. And then actually process that information.

Third, and probably hardest, is to actually listen to the advice you are getting. It is well known from cognitive research that we tend to hear what we want to hear. This is very different from saying you have to take the advice. You have to make your own decisions, but if you don’t pay extra strong attention to what you hear, then the effort on #1 and #2 above is wasted.

I have learned all of these the hard way and wrote this post mostly because I need to remind myself of these lessons. And I try to act accordingly as a board member, with the big caveat that there are times when economic interests can be significantly at odds. I will post another time on how I try to address that.

We just announced on the Union Square Ventures blog that we are hiring for two positions: an Analyst and a General Manager of the USV Network.

Last week we took a family vacation in Rio de Janeiro!  We went to visit friends who had lived just up the street from us here for the past eight years but moved back to Rio (where they are from) a few months ago.  I was excited because it was my first time in the southern hemisphere, my first time in South America, and my first BRIC country visit (Brazil, Russia, India, China — planning to visit all of them over the next couple of years).

The most notable impressions from Rio were the palpable sense of economic growth and my high perceived level of safety.  Due to the city’s fairly unique geography of being bordered by water and mountains, expansion is taking place primarily further out in the Barra area where we stayed.  There one can see tons of new high rise buildings being constructed.  The result is a bit of a traffic nightmare.  From our friends house to Ipanema is 20 minutes without traffic and 60 - 90 minutes during rush hours!

Given the extensive warnings we had gotten before leaving, I was a bit surprised that I felt completely safe the entire time, including going out at night, getting lost and asking for directions.  I have spent a little time searching online, trying to find a break down of crime statistics for Rio between the favelas and the rest of the city but without luck.  Based on appearances alone (and hence not to be trusted), it seems as if much of the violent crime takes place inside the favelas.  Incidentally, the housing in the favelas was less rickety than I had imagined it — construction is crazy tight and without any building codes, but one can also see tons of air conditioners and satellite dishes sticking out and Internet access is apparently widely available.

Postscript: Sadly, Rio was hit by torrential rains right after we left and close to 200 people have died, mostly from mudslides and mostly in favelas (which at their edges tend to push up against ever steeper territory).

The big news of course isn’t multitasking, or unified inbox, but iAds and GameCenter.  These are aimed squarely at making the iPhone (and iPad) a monetizable and viral platform to rival Facebook.   Still this strategy seems flawed by being tied entirely to Apple Hardware.  What percentage penetration of the handset market can Apple possibly achieve?  Facebook on the other hand can reach any user on any device.  Well … any device other than the iPhones (and iPads), where Apple can of course exercise control through the app store.  Interesting times!

I have blogged in the past about threats to the open web and what to do about them.  It is easy to be dismissive of concerns about Apple’s philosophy of control which necessitates a partially closed system.  But there is a real danger in being so seduced by the sleekness of a device and its user experiece that one does not reflect on whether the trade-offs in terms of restrictions are actually worth it.  That becomes all the more relevant as Apple is beginning to use control of the playing field to tilt it to its favor, as evidenced by Marco’s post yesterday about the use of private APIs in the iBooks apps. In the case of Apple, I am glad that the market is at work thanks to Google’s Android push (that push itself is not without problems due to Google’s ability to cross-subsidize).

Also yesterday we saw a major setback with respect to net neutrality as a court ruled that the FCC lacks sufficient authority to regulate broadband providers.  Again there is a real risk here to be swayed by claims of cost and capacity limits to accept restrictions on the flows of information that could be far more severe than any perceived benefit.  Here there is much less of a market force at work as a potential corrective because in many local markets there is only a single broadband provider available and at best most markets have a duopoly. 

I do believe that the price of Internet freedom too is eternal vigilance, which is why I keep raising concerns like these on this blog and support others who do.  At the end of the day if all the concerns wind up being unwarranted, no harm will have been caused by having raised them.  But if we were to find ourselves in a world where a small number of companies controls what applications go onto our devices and how those devices can access information we will have suffered great harm.

We were on family vacation through Sunday, so when our iPad arrived via UPS on Saturday nobody was home, which meant that we had to wait until Monday for it to be re-delivered.  I wish I could share first impressions this morning, such as Bijan did this morning, but we frankly didn’t get that far.  The iPad appears as a beautiful physical object, but when you turn it on it greets you in a very disappointing way: an icon asking you to connect it to iTunes via a cable!  What a contrast with my experience with Google’s Nexus One phone.  On that, all I had to do to get started was — well nothing.  I was able to turn the phone on and start using it right away.  When I wanted to sync it up to gmail, gcal, etc. all I had to do was enter my google account info once!  Yes, Apple has MobileMe — but it is a paid service and as far as I can tell covers a small fraction of all Apple users.  I believe that for a slate to be truly useful, syncing to the cloud is a critical requirement.