Seattle Cloud Camp 2009

First of all, the view here is awesome



A few big names here. I just talked to the guy who created slice host. Hopefully I'll get to chat with Dan Kaminsky. Uninspiring organization, but that's to be expected from an "un-conference".

Update: Talked to Rackspace. They have no idea about Cell/CUDA. Talked to Jeff Barr from Amazon - they are starting to play around with it. Nothing public.



Update: By far, the best talk given that day (that I attended) was the one about security by Dan Kaminsky.

I did learn a few things in the talk, related to security, mostly dealing with how practical day-to-day operations affect security practices. It was interesting that essentially the message of the talk is that there is a really low bar for top-of-the-line security so moving to the cloud isn't that big of a deal.

The thing that really impressed me was the super simple language that was used. It almost reminded me of listening to the Feynman lectures on physics. With such simple language, the transmission of meaning becomes digital - either you get it or you don't. There weren't these shades of understanding that so often haunt most communication.

The other thing that I really appreciated was that it was a talk - Kaminsky talked for more than 50 minutes and the entire 50 minutes was informative. All in all, terribly amazing.

Recession Links

Hat Tip: Cassandra Does Tokyo

An excellent discussion on the solvency of the US banking system focusing on five different ways that banks could be solvent and examining each individually: Bank solvency and the "Geithner Plan"

Some criticism of the previous post's numbers, one of the key points being that never in recent history has risk premia between AAA and junk been 15%: Prospects for the U.S. Banking System

All Soul's College Enterance Exam

Quite possibly one of the most awesome things I've run across lately is the entrance exams for All Soul's College here: All Souls: The toughest test you’ll ever take

The really neat thing about this test is how open ended the questions are - they really give people an opportunity to weave a compelling case.

The format of the test is quite simple. There are five papers people are required to write - 2 from the general list, 2 from the person's area of study (listed below) and 1 from the general list whose entire subject is a single word (past examples: "value", "water").

I'd like to imagine that this sort of test is the most difficult to fake, but I guess I'd have to see examples of previous years' entries. I'm definitely going to take a stab at taking the test using previous years' questions.

The article links to some of the past examination questions which have since been removed from the site. The Wayback Machine comes to the rescue, though. Questions from 2002-2006:

General Classics Economics English History Law Philosophy Politics

Update: The History and Law files seem to not be available. I have backed up the others in case they become unavailable as well.

New Thoughts on Recession

I've been thinking a bit about recessions and why they occur, and I think I've narrowed things down even further than before.

The crux of the problem is the cost of capital and the need for profitability. During a boom, money is cheap because everyone is trying to spend it to buy the boom commodity. During a recession, money becomes expensive - most people need money to cover the loans they took out during the boom period.

The big problem is all of this is that the interest rate, or the cost of capital goes up in a recession.

An example: say a business owner secures a 10 year loan at 6% interest to finance his business. In a boom, there might be as much as 3% inflation, making the loan a 3% loan in real terms. The business will be profitable if it generates a net profit over 3% + other costs. In a recession, however, the 3% inflation turns into 5% deflation. Now the loan costs 11% in real terms. If the business is profitable at a 3% real cost of capital but not profitable at 11%, it will go out of business, and rightly so.

This is one reason why Paul Krugman is an idiot.

If the real cost of capital is 3% for an extended period of time, there will be plenty of businesses that are just eking out an existence(there are plenty which fail to meet any threshold set). One of the big drivers of unemployment is the rising cost of capital. Businesses that are unprofitable at the new, higher cost of capital have to either go out of business or become profitable by firing their least productive employees.

This too, is the reason why it takes so long to alleviate the unemployment in a recession - it takes a great deal of ingenuity and opportunity to come up with ways to be profitable at the new, higher cost of capital, particularly if you are being crowded out of the capital market by grossly unprofitable government stimulus.

Update: Krugman certainly isn't an idiot. From his work he is obviously intelligent. That being said, he is also quite obviously more interested in influencing public policy than understanding economics.

Good News for Google App Engine

I'm in the process of crunching through the "news" section of my feed reader and came across this blog post on the Google App Engine Blog: The sky's (almost) the limit! "High CPU" is no more.

The short story is that there is no more "High CPU limit" - the one major error that Concurr and many others were constantly bumping into.

There are also some additional features - longer timeouts, etc. Looks like the constant development on the App Engine platform is starting to make it into an actually viable platform for developing stuff. I'm sure there are a few other deal-breaking pet peeves out there, but if the GAE team keeps on cranking away at them at the current rate, the platform will be quite polished a year from now.

Nationalization or Subsidies: You Decide

I used to be of the mindset that there should be no subsidies, tariffs, etc. Companies should rise and fall on their own merits. If a country wanted a national industry for military reasons, it should just nationalize it outright. I'm not so sure anymore.

Some of the upsides of nationalization vs subsidies:

• Complete control of policy by the government
  
• Impossible for the company to appear to be failing
  
• Maintains the purity of the market. This provides a higher quality information source for investors
  
• A clean separation between government and consumer goods means other corporations could (depending on the details) provide the same good as the nationalized company to the civilian sector.
  

Some of the downsides of nationalization vs subsidies:

• Government pay scale - possibly difficult to attract and keep top talent.
  
• Also, the tendency to attract people more interested in power than money can degrade the functionality of the government in general.
  
• Market pressures and a wider customer base drive more extensive cost cutting and product innovation
  
• It is easier to see how much the (quasi) nationalization is costing the government
  

I can see why a democracy generally prefer nationalization and an autocracy would generally prefer subsidies - many of the benefits of subsidies I have listed are actually liabilities for political actors in a democracy. That being said, I am actually leaning slightly towards subsidies the preferred means of nationalization.

MBA Training ground?

I've been thinking a bit about the relatively high quality leadership that seems to be present in some of the MMO organizations. Perhaps that is a place to think about recruiting leaders from for real world corporations?

Of course there are aspects to the real world that such people are not used to dealing with, but my guess is that the very best would be quite adaptable. Perhaps this is equivalent to the hire-open-source-programmers strategy for leadership?

Superbowl Madness

Reflecting a bit on the SuperBowl, I am forced to admit that the Referees were somewhat biased in the execution of their duties. While that fact is not entirely surprising, being that they are human beings, on reflection, the equity holders for the NFL, at some level, have to be in favor of this.

It would be simplistic (and possibly correct) to claim that this is a bad decision; I was wondering what some of the reasons this might be a good decision are.

Perhaps fans have a short enough memory that biased judging doesn't degrade the reputation of the league in a long-term manner.
Perhaps enough fans will leave a star team if it "goes cold" that a short term perspective is super-important.
Perhaps there is some sort of referee's union and the equity holders can't exert meaningful pressure on the referees.

I'm afraid I'm at a loss - these three reasons seem relatively suspect to me. I've seen with my own eyes (GM, Ford, etc.) shareholders commit long-term suicide, so it certainly does happen. I'd just like to think that it isn't true in this case.

Greed and War in the Virtual Age

I just read a fantastic article at Shacknews: EVE Online: Spy Game.

There aren't really any more grand wars going on. Nukes and MADD pretty much put a stop to that sort of thing, if ideology wouldn't have prevented it anyhow. On EVE, though, massive galaxy-wide wars rage all the time - and they aren't just military. There is a neat synthesis of military and economic warfare along with infiltration and espionage. The neat thing is that since everything is done virtually, records are much better - someone knows (after the fact) pretty much what happened.

If you want to know how the best in the business do their well... business, this article is certainly a good start.

The Stimulus Package

In the spirit of Ryan and Scott, I thought I'd say something about the Stimulus.

I don't think that half-way measures will work, particularly in this case. The bill is either too large by far to accomplish its task, or it is too insignificantly tiny to accomplish its task.

As a side note, I'm somewhat surprised that the silver lining hasn't lasted longer - a choice quote from one of the articles Scott linked to:

The last thing is just about the stimulus bills as it stands. Two things here. One thing is what do you think about the ratio of spending to tax relief in the bill. And the second is, if you judge it by Larry Summers standard -- that stimulus be temporary, timely and targeted -- does it clear the bar?

This is probably the worst bill that has been put forward since the 1930s. I don't know what to say. I mean it's wasting a tremendous amount of money. It has some simplistic theory that I don't think will work, so I don't think the expenditure stuff is going to have the intended effect. I don't think it will expand the economy. And the tax cutting isn't really geared toward incentives. It's not really geared to lowering tax rates; it's more along the lines of throwing money at people. On both sides I think it's garbage. So in terms of balance between the two it doesn't really matter that much.

Wondering about the Froth over the Community Reinvestment Act?

I just read an amazing article over at Going Private: Dare Ye Inquire Concerning Such a Wretch?

Anyone who is interested in a high quality intro into the free market (or not) nature of the Real Estate market of late (particularly in regards to the Community Reinvestment Act) could do much worse than to peruse what looks like the finale post of an amazing writer.