Abtin Forouzandeh

Six Million Lines of Code

Abtin Forouzandeh — Fri, 13 Jul 2012 11:30:00 EDT

Lets say you have one thought per minute and that each thought represents one line of code. Lets say you work ten hour days. This means you have the capacity to write 600 lines of code per day, 3000 lines of code per week, or around six million lines of code over a 40 year career.

“Wait a minute,” you exclaim, “I write way more than a line of code per minute when I really get going!” Sure, I do too. But lets be honest: you are not going to write six million lines of code in your lifetime. Think of the numerous factors that conspire to reduce our programming output: meetings, status updates, unit tests, documentation, hacker news, problems at home. They take a toll. A massive one. So lets be generous and say you are an extraordinarily brilliant programmer, operating at 50% efficiency - better than most natural systems.

You have three million lines of code in you. What are you going to do with them?

The possibilities are limited by one key factor: those three million lines of code also must not only provide for you and your family, but also for your retirement. So if you earn anything less than $1 per line of code (total compensation), or about $30/hr, you are in serious trouble. And if you live in silicon valley, well, you know that you need to earn quite a bit more than that to get by. And don’t forget retirement! Given improving lifespans, its not unreasonable to assume you will live 40 years beyond retirement.

If you want to get good, don't gamify - workify!

Abtin Forouzandeh — Thu, 12 Jul 2012 11:30:00 EDT

Recently, while watching the original The Karate Kid, I was struck by how the Mr. Miyagi’s teaching process contradicts modern conventional wisdom.

The Kid wants to learn karate. How does Mr. Miyagi teach him? Does he develop a gamified point system to keep The Kid’s interest? No. Does he break the process down into a series of small goals to create reinforcing patterns and build confidence? No.

Mr. Miyagi tells The Kid to wax a bunch of Mr. Miyagi’s old cars. And then sand Mr. Miyagi’s huge patio and fence. And then paint Mr. Miyago’s fence. And then paint Mr. Miyagi’s house. Each time, The Kid is told to perform the task with a specific hand motion.

We’ll avoid discussing leftist queasiness with unpaid internships.

Instead of focusing on the direct fun thing (learning karate), which is the easy gamified approach, Mr. Miyagi “workifies” the whole process. He makes it harder and not fun. Because The Kid needs to develop so many other skills to properly master karate - build muscle, build endurance, build calluses, build muscle memory, and so forth.

And I couldn’t agree more with this approach. If you really want to learn something through and through, you can’t take the easy way. You need workify it so that you build all the ancillary skills required.

This is why I am not at all interested in efforts like Codeacademy. It takes time to master a new skill, and directly learning that skill is just about the least effective way to do so (Consider Peter Norvig’s famous essay Teach Yourself Programming in Ten Years). You need to constantly practice and apply the skill, because there is no other way to discovering and mastering all the ancillary skills required.

If you really want to master programming, do it the hard way so that you get exposed to all the fundamental concepts. For example, if you want to build a website, do it in straight c using the mongoose library. It will be hard and you will struggle. But you will come out of with a solid understanding of pointers, memory management, maybe even recursion.

A Startup is like Fishing in an Empty Pond

Abtin Forouzandeh — Wed, 11 Jul 2012 21:30:00 EDT

A few months ago, I went fishing in a remote Mississippi pond.

It was my first time fishing.

I was told told that I probably wouldn’t catch any fish because the pond hadn’t been restocked in a long time.

But that didn’t matter, I wanted to learn how to cast a fishing pole.

I was told that if the pond contained some fish, it would a kind that could be caught with certain type of lure that floats near the surface.

I was told that it was about the perfect time to catch this certain fish, around noon.

I spent about two hours casting the fishing pole, to no avail.

Of course, I got better at casting. Soon, I was able to pick a spot to cast to and the line stopped getting tangled.

I was told it was time to go. The fish weren’t biting that day.

Then I felt a tug on the line. It felt heavier than what I was told it would feel like.

I reeled it in.

And it was a catfish.

Catfish swim near the ground, so I had been using the wrong kind of lure.

Catfish are nocturnal, so it had been the wrong time of day.

Catfish weren’t supposed to be in this pond, it had never been stocked with them.

Against all odds.

It doesn’t matter what the environment is. It doesn’t matter what the market says. It doesn’t matter what feedback you get.

You learn. You work. You endure. You persevere. And with a modicum of luck, you succeed.

The Genius of the Facebook IPO

Abtin Forouzandeh — Tue, 22 May 2012 13:30:00 EDT

There is a lot of misguided chatter about how the FB IPO has gone horribly wrong. The situation is in fact the opposite: FB IPO is the most successful IPO in history.

Contrary to popular myth, the purpose of an IPO is not to enrich wall street investment banks. The purpose of an IPO is to raise capital for the company, fueling rapid expansion without the downsides of incurring debt. The success of an IPO is determined by how well it meets this purpose.

Unfortunately, most companies kowtow to wall street, reducing the amount of capital raised, for various reasons:

They company is unable to generate sufficient demand and so is forced to lower its wholesale share price. This price is what the investment banks pay for the stock, and they profit by selling it at a higher price to the public. Of course, they are performing a vital function, but they often have enormous leverage in negotiating the wholesale/retail spread. Note that as the spread widens, the company raises less capital and the benefits of IPO accrue to the investment banks in increasing proportion.
The founders/internal shareholders want to cash out as much as anyone. They may agree to a wider spread in order to win terms that enable them to make more money at launch.
Founders/Internal shareholders need the price of the stock to jump so they can keep their jobs. Typically, by the time of IPO, founders hold less than 20% of corporate stock.

Facebook didn’t face any of these problems, and so they were able to raise the maximum amount of capital to fuel growth - capital that, in a standard IPO, would be used for investment banker bonuses:

You would be insane to bet against Facebook. They have employed the very best individuals to build their future, and they are starting from a base of close to one billion users. One Billion Users.
Mark Zuckerberg has repeatedly stated that he takes the long view and is uninterested in short term payouts. And thanks to secondary markets and tertiary investment rounds, internals shareholders have already been afforded the opportunity to cash out to whatever degree they desire.
Mark Zuckerberg holds well over half the voting rights in FB, meaning that investors have no option to fire him, short of an extremely expensive civil case.

I expect there to be a lot of fallout from this IPO.

Wall street executives have extensive connections in the political world. They will almost certainly seek new regulatory rules that will protect their profits, for example by mandating a minimum spread between wholesale/retail prices. Of course, they will lobby in the name of protecting ordinary investors, even though only large institutions capable of participating in the pre-IPO process will benefit.
Wall street will lobby to adopt European-style rules where multiple-class stock structures are illegal. The FB IPO absolutely could not have happened if Mark Zuckerberg didn’t possess super-voting founder stock.
Founder’s will do their best to push for similar IPO structures, because the benefits to the company and employees vastly outweigh any perception problems. Sure, the stock may suffer a little bit because retail investors have a herd-like mentality and want to invest in winners. But, the additional capital should help the company grow better than it otherwise could have, meaning the perception problems will resolve themselves in the medium term.

Capital Gains Tax Rates and Antisemitism

Abtin Forouzandeh — Tue, 17 Apr 2012 10:55:00 EDT

It started when a friend of mine asked the question:

Question for those more financially savvy than I am: what is the economic logic behind taxing capital gains at a lower rate than regular income? It seems like the lower tax rate is intended to foster economic growth by encouraging investors to hold onto the asset(s) in question for longer, thus reducing market churn and encouraging long-term investment. Is that the sole reason?

What disturbed me about the ensuing discussion is a sentiment that comes up frequently when discussing Wall Street, “speculation”, high-frequency trading, banking, and just about any other finance related activity, industry, and profession.

One person expressed it this way:

Today [the capital gains tax rate] is also applied to short sells and other casino-like behavior where the asset is bought and resold in a matter of days or even hours and there is absolutely no justification for that.

Another person said it like this:

If you take money you have saved and invest it in a new company, like a VC investing in a new technology startup, this actually creates new jobs. This money should be taxed differently than plunking money down long term on something like…betting oil futures rise in the next year.

The problem is that our tax law treats the two the same. One reinvests capital into the economy to create jobs (good) the other shuffles it around (usually with insider information) to make sure that gains outpace infaltion. The other side effect of investing like this is that it ties capital up in a way that ensures no new jobs get created (bad).

For thousands of years, people have displayed open hostility to those who make money through finance. This has most frequently manifested itself through antisemitism and marxist anticapitalism, resulting in the murder of tens of millions across the globe.

Murder is out of fashion now, so it’s popular to advocate that certain undesirable individuals/professions be taxed to death. But the thinking is the same: if we could just stop certain people from trading in a certain way, all our economic woes would go away. History shows that this has never worked. And there is no reason to believe it ever will work.

Just because you don’t understand something doesn’t mean it doesn’t provide a ton of value. And just because you can’t see the effects of a policy or course of action doesn’t mean there aren’t any.

Anyway, here is my response to the original question:

Encourage investment. The lower the tax rate, the lower the returns need to be for a given investment to make sense. Warren Buffet has talked extensively about this.

Thus the ideal capital gains tax rate is zero, which would lead to significant job creation over time (with the caveat of point number 5 below). This means that it is entirely plausible that a zero capital gains rate would actually increase total government tax receipts as a result of income taxes on new jobs.
Avoid multiple taxation. In order to make investment income, all of these things must happens: a person must make money (which is taxed), invest that money in a corporation, the corporation must make money (which is taxed), the corporation must issue dividends to the individual (which is taxed). Thus, capital gains taxes actually represent triple taxation. Some will say this is fair because a small portion of people in this country have never worked and should pay through the nose. But it’s unfair to the overwhelming majority of people - think anyone who is on the verge of retirement and has dutifully contributed to retirement accounts for decades.

In a very narrow set of circumstances, the second tax (corporate tax) can be avoided, but that is very uncommon and not relevant to the original question.
Inflation. Lets say you invested $1,000 in Microsoft 10 years ago at about ~$25/share and you sold it today ~$31 for a total of $1240. Unfortunately, you just lost money because of inflation - according to the BLS , $1000 ten years ago is the same as $1275 today, resulting in a de fact loss of $35. Even more unfortunate, the federal government doesn’t care - you pay 15% tax on the raw $240 “profit”. This results in increasing your loss from $35 to $71.

It is easy to imagine situations where the capital gains tax could turn even a modest real gain into a real loss.
Capital gains disproportionately hurts seniors. Young people typically have very little investment income, but they can work long and hard and make money. Old people can’t and have to rely on whatever provision they made earlier in their lives.
All of that said, it doesn’t really matter in terms of institutional investment. There is an enormous body of work showing that temporary, whether actually temporary or perceived transience due to political volatility, tax rate cuts have almost no impact on invest decision making. Our current political/economic environment, where raising the capital gains tax is even merely discussed, is enough to massively deter investment. The only way to stimulate investment is through credibly permanent tax cuts.

Do Not Use Go for 32bit Development

Abtin Forouzandeh — Sun, 08 Apr 2012 10:48:00 EDT

I’m sitting here rewriting a ton of Go code in C. And I can’t bill for the time. As the principle programmer, the buck stops here. I talked to my client and took responsibility. My mistake was trusting Go. I should have remembered my motto, stolen from the Royal Society: nullius in verba.

Despite the 1.0 release announcement of Go, I must emphatically state that if 32bit support is a hard requirement for your development, Go is not a viable option.

There are a number of bugs in 32bit Go, but the real show stopper is this:

At runtime init, Go programs attempt to reserve 512mb of virtual address space. If they can’t do this, they crash. Go needs the space because it is garbage collected - presumably, all GC languages, or any program that takes a similar approach to memory management, will be susceptible to the same problem. On 32bit Windows machines, processes only have access to 2GB user space memory. This user space can become fragmented (say, a linked DLL is loaded by the system before the 512mb allocation occurs), preventing the Go runtime from reserving its arena.

Go expects one large allocated block. I assume that other GC runtimes implement an algorithm to deal with this problem by requesting smaller chunks of memory.

The problem is hard to replicate and test - most of the time, Go programs run perfectly fine. I speculate that rebooting the machine might help solve the problem. I also speculate that only Windows is susceptible - I’ve had no bug reports about the linux version of my code. The Go devs are aware of the issue, but have designated it as low priority.

I love programming in Go. And I agree with Rob Pike that programmers complain too much when they should be fixing the problem instead. Unfortunately, I am not familiar enough with the Go runtime code to dive and make the required changes with confidence that I have both actually fixed the problem and not introduced bugs into the memory management code. I don’t have time before my deadline to look into the issue either. I will certainly look into fixing the problem once I have some spare time.

Edit: I would like to point out that the Go devs have been very helpful in diagnosing the issue. Unfortunately, their resolution always boils down to “Switch to 64bit”.

Edit: The original article states that Go tries to reserve 512mb of RAM. This is incorrect: it tries to reserve 512mb of virtual address space.

Edit: More comments on hacker news

Economics for Programmers - Backward Bending Supply Curve of Labor

Abtin Forouzandeh — Tue, 03 Jan 2012 15:58:00 EST

Lets take a look at the backward bending supply curve of labor from a programmer’s perspective. For individuals, this is one of the single most important concepts in microeconomics. It explains why people are willing to work insane hours in the early part of their lives, but relatively little as they grow older. If you ever wondered why Fortune 500 CEOs make so much money, this will explain most of it.

The Algorithm

Here is how I would illustrate the backward bending supply curve of labor with code:

//starting conditions
time = START_OF_ADULT_LIFE
net_worth = BROKE
productivity = INCOMPETENT

//the days of our lives
while(time++ < LIFE_SPAN)
	income = CurrentMarketPriceForLabor(productivity)
	if(MarginalUtility(income) > MarginalUtility(personal_time))
		DoWork()
		net_worth += income
		productivity += AdjustProducitivityForIncreasingSkill()
	else if(OfferedAboveMarketIncome()) 
		if(MarginalUtility(above_market_income) >
			MarginalUtility(personal_time))
			DoWork()
			net_worth += income
			productivity +=
				AdjustProductivityForIncreasingSkill()
	else
		net_worth -= DoPersonalTime()

//rational thoughts
MarginalUtility(income) =
	Utility(net_worth + income) - Utility(net_worth)
MarginalUtility(personal_time)
	Utility(net_worth + DollarValue(personal_time)) -
	Utility(net_worth)
DollarValue(personal_time) =
	//this should probably be further adjusted to value personal
	//time over work as time approaches LIFE_SPAN
	net_worth / DESIRED_NET_WORTH * MAXIMUM_PERSONAL_TIME_VALUE

What it means

We start life with low skill, worth, and too much free time. So we are willing to trade that free time, which might have been spent playing Call of Duty, in order to work and put food on the table.
As we work, our productivity improves because we gain experience and skill.
As our productivity improves, we make more money.
Making more money is great. But like all things, the more money we have, the less we need more of it. The first $40 million is great, but maybe the next $40 million should be ignored in favor of playing more Call of Duty.
The value of MAXIMUM_PERSONAL_TIME_VALUE is extremely high. So as we approach the end of our work life, when our net worth approaches its peak, the value of personal time is vastly greater than when we started life.
It is worth working hard early in life because the sooner we approach our desired net worth, the faster our DollarValue(personal_time) approaches its peak. The longer we can live with that value maximized, the more happiness we derive from life.
The only reason we work late in life, assuming we have steadily increased our net worth, is if we have a massive possible reward. This could be both in the form of money, as in tons of stock options and bonuses, or perhaps some non-pecuniary reward, like the chance to create the iPhone and change the world. Without such a reward, we would derive much more happiness be enjoying our personal time instead of working.
A common complain about public-sector workers is the obscenely early retirement age most of them enjoy - usually after just 25 years of work. Once you have mastered the backward bending supply curve, it becomes obvious why. As a result of their political influence, they are able to exact high compensation and retirement benefits. Naturally, once these high benefits are secured, the next objective is to lobby for an earlier retirement date so that they can be enjoyed for as long as possible. For readers outside the US: American government employees are unique in that they are the only segment of the population to enjoy retirements where their income only goes down a small fraction and is guaranteed to be protected from inflation through yearly cost of living adjustments.

Notes

Astute readers of my previous article will observe that this flat out contradicts the advice to pay programmers $200/hr, since the programmer will more quickly hit their desired lifetime income and retire.
Clearly, this ignores that our desire to have free time is cyclical. In other words, it ignores our need to have weekends in order to explore the larger concept of the role of work over a whole lifetime.
Certainly, some people enjoy their work for its own sake. In that case, this algorithm would be tweaked so that net_worth is not discounted by DoPersonalTime(). But the common view of work is something that we would generally not do unless we are paid. So as a programmer, you might enjoy coding up some utilities for personal use, but you probably wouldn’t be writing internal business software for the umpteenth time. So even though you enjoy programming, it is still work if you are doing it for someone else’s benefit.

Pay Your Programmers $200/hour

Abtin Forouzandeh — Fri, 30 Dec 2011 17:42:00 EST

Update: more comments on reddit and hacker news.

If you are hiring programmers, you should pay them $200/hr. This breaks through otherwise impenetrable psychological barriers, helps solve the agency problem, and ensures you are only hiring programmers when you really need them.

Expectations

$200/hr is an open declaration of your expectations. It states that you are expecting to hire both a professional and an expert in the field.

As a programmer, I treat contracts that pay me $200/hr differently from jobs/contracts that pay me $65/hr equivalent salary in two key ways. First, I treat issues and requests with far greater urgency. I’ve been on vacation when a call comes in, the customer requesting a new feature. I will go back to my hotel room immediately and build it; the high rate means its easy to justify this to my spouse - in fact, she totally supports it. I would simply never consider behaving this way under normal job conditions, because its not worth it. The worst that might happen is I lose my job, but there are a glut of $65/hr jobs available.

Second, and really this is a huge benefit to the employer, I generally do not charge for what I consider to be “learning” time. If you pay me $200/hr, I will only charge for the amount of hours that I believe an expert in the field might charge. I expect, with some variation in degrees, most programmers will behave this way. For example, I was given a project that involved coding a ton of low level C. I hadn’t coded any C in a decade. I spent two 16 hour days getting my c legs back, but I did not charge the client at all for this. I also did not charge for any bugs that I felt an experienced systems programmer would not have made. Overall, this had the effect of roughly halving the number of hours I billed.

Loyalty

But guess what? I didn’t resent the client at all. In fact, I was delighted and so were they. I got to ramp up on some rusty skills and know they value my work, and they got a product that met all their needs, on-time and on-budget.

And the next time my client needs something? They shoot right to the top of my priority list. And they benefit from an even faster and higher quality turn around, because I only need to brush up on C once, I don’t repeat the same mistakes, and I learn more about their business and the types of features they need but may not know to ask for.

You also never need to worry about the programmer leaving you. If you have a ton of work to send their way, they will do it and pocket huge surpluses on their labor. If you can only send them a few hours per week, they will still be able to live comfortably and enjoy their free time. Its win-win for contractor and client.

On a political note, high pay correlating with high expertise is more true among programmers than other professions, like lawyers and doctors, because the only reason other professions routinely charge higher rates is because they have artificially constricted supply through licensure requirements - you are effectively coerced into paying a high rate. But when you pay a programmer a high rate, there is no element of coercion whatsoever. This recognition of value is big loyalty booster all by itself.

Trust

For the employer, trust is less of a problem when you pay a high rate on contract. If the programmer doesn’t deliver, fire them. If they deliver low quality goods and don’t fix it, fire them. If you are unhappy for any reason at all, fire them.

With a lower paid programmer, its much harder to justify rapid firing. You might be tempted to do a cost/benefit analysis - maybe you suspect that programmer is slacking, but you still feel you get a good bit of work out of them. This is a fatal error. First, there is no other market on the planet where the information asymmetry between vendor and client is greater. The programmer can spin any tale they want, and you will have absolutely no way to dispute their claims. Second, ineffective or disaffected programmers are not just zeros, they are are negatives. Really big negatives that damage your business.

Optimal Resource Allocation

If you cannot afford $200/hr, you should not be hiring programmers because you are working on a problem that addresses an insufficiently large market.

Which situation would your rather be in: risking $100,000 investment for a $200,000 return, or risking $1,000,000 investment for a $50,000,000 return? This might seem like a self-answering question, but virtually all programming hires fall into the first category. Nearly every IT department and most businesses are focused on the wrong class of problems. And the problem is that this is a gross misallocation of resources, because there are tons of opportunities in the second category.

When you focus on a problem in the second category, a lot of silly business problems go away. Super tight controls on payroll and expenses are no longer needed. Schedule slips are less of a problem and less likely to occur. Morale improves from working on a big, important problem.

On the other hand, when you are in situation where paying a programmer $75 instead of $65 would destroy the economic justification for the hire, you are in an extremely fragile state. All the problems above are exacerbated and you are one micro-economic shift away from bankruptcy or irrelevance.

I Just Ordered a Ladle from Amazon

Abtin Forouzandeh — Thu, 29 Dec 2011 23:02:00 EST

I just ordered a ladle from amazon.

I looked at a few different ladles.

And started reading ladle reviews.

Actual reviews that real people wrote about ladles.

And thats when I realized the internet really has changed everything.

And then I stopped reading the reviews and just ordered the least expensive ladle.

Because what the hell is wrong with me that I am reading ladle reviews?

Update: more comments on hacker news

Why I Will Miss Steve Jobs

Abtin Forouzandeh — Thu, 01 Dec 2011 17:57:00 EST

Take a look at the screenshot below, from the iPhone iOS5 weather app:

See that word “hourly”? It has meaning. Touching it, or anywhere on the screen, results the display of hourly weather data:

False Affordances

So whats wrong with this?

The word is grayed out, which in any other setting would mean “nothing to see here, this feature is disabled.”
“Hourly” is an adjective describing your current view: when it is gray, you are in the daily view; when it is white, you are in the hourly view. This is stupid - it is utterly obvious to the user which mode they are in.
“Hourly” should be a verb or an image (say of a clock) because pressing on it causes a change in the UI. If it followed the convention of the rest of the default applications, like the traffic function in maps, it would be “Show Hourly Forecast” and “Hide Hourly Forecast”.

Fixing these sorts of problems prior to release requires one individual to have total ownership of the product. With Steve gone, an no apparent iPhone czar replacement, the number of similar paper cuts in iOS will grow over time.

And this is why I will miss Steve Jobs: he was perhaps the greatest example of our time of the meaning of “ownership”.

What is Ownership?

I once worked on a leaderless software team. I approached my manager and explained that no one on the team felt a sense of ownership over the product. He has no idea what I was talking about and started rambling something about how “everyone is working really hard and if you don’t like it, leave.” I took his advice and, shortly thereafter, the project failed.

There are 3 elements of product ownership.

Caring About the Product

Paul Graham has described this as the thing you think about while taking a shower.

The same applies in marriage. There is an episode of House in which a man turned orange and mentions that his wife hadn’t noticed the change. House diagnoses the problem as drinking too much carrot juice (delish!), and further declares that the patients wife must be cheating on him. This turns out to be true. When you don’t care, or are distracted, you don’t notice the little problems. You are thinking about the wrong thing when taking your shower.

Caring about the product enables you to notice every little problem with it. When you hear some anecdote about Steve Jobs chewing out an engineer because something is 5 pixels off, that is caring about the product. Apathy dulls perception.

Enjoying the benefits of success

This is where every large company falls flat on its face. In medium or large sized companies, the benefits of a product’s success do not accrue to the responsible team, but to the company.

Because corporate owners want to earn as much profit as possible, they often attempt to persuade employees that this element doesn’t matter. Joel Spolsky is notorious for this - he has written at length about why compensation and bonuses are bad ideas - that employees should be driven by “intrinsic” motivation. What a bunch of communist garbage.

I have intrinsic motivation to master the skills that I care about or to build my career - programming, business, marketing, and playing Call of Duty are all things I intrinsically care about. However, working on a specific product is never intrinsically motivated. When someone tells you that asking for more money is wrong, that you should be “intrinsically motivated”, what they are really asking you to do is to abdicate reason. If you are working on a project, there must be a reason, and if that reason is anything other than a desire to make money, or otherwise improve your (or your loved one’s) material well-being, your reasoning assuredly faulty.

Of course, the route to making money doesn’t have to be direct. You might work an unpaid internship in order to build skills in order to eventually get a good paying job. But this is obviously different from developing skill mastery that you care about. You don’t have to pay me money to get me to learn a new programming language (or to play call of duty), I love doing it. But you do have to pay me to use those skills to build your product.

Suffering the costs of failure

Enjoying the benefits of success is the carrot, suffering the costs of failure is the stick. And once again, most medium and large corporations get this totally wrong. They try to shield employees from failure by merging failing teams into the rest of the company rather than firing everyone.

Suffering the costs of failures enables better judgement of risk. Sure, you could decide to stop all work on your year-old project to write a new c compiler, or design a doohickey, but it will dramatically increase your team’s odds of failure. Programmers are notorious for this sort of thing, but this model makes it perfectly clear why. Programmers almost never enjoy the benefits of success (beyond a modest salary). Thus, any programmer that actually cares about programming will seek to improve their intrinsic skills, even if its is detrimental to the success of the project. Put another way, because the programmer’s compensation is low, their cost of failure is low as well, and so they do not assign high risk to project failure - instead, they assign high risk to failing to develop their own skills. After all, it is trivial to find a crap paying programming job - engineers never experienced the great recession. The only way out of crap paying programming jobs is through further skill development.

If project failure does not result in a greater loss to you than failure to develop your intrinsic values, you do not own the product.

What it means

If you are building a startup, you need to own it.

You have to care about the product, so it needs to align with your skills and interests. Trying to build a startup in an area you don’t care about means you will fail to address problems and to seize opportunities that your more caring competitors will exploit.

You have to benefit from you success, so excessive outside investment or too many co-founders diluting your stake will destroy your ability to stay focused.

You have to suffer from failure, so you can’t keep your day job. Perhaps you can build a prototype while employed, but, if you want to take it to the next level, you must quit your job and embrace the startup.

Steve Jobs showed the success possible by embracing these values, and he changed the world in the process.

Oil Companies and LIFO

Abtin Forouzandeh — Fri, 14 Oct 2011 17:20:00 EDT

Recently, in the WSJ and other places, it has been proposed that oil companies are escaping their taxes through LIFO accounting. This makes no sense at all and is obviously being proposed by folks with little accounting knowledge. At least, less knowledge than me, and I’ve only taken Accounting 101.

For the purposes of calculating taxes and sec reporting obligations, companies have to pick one of three inventory accounting methods: LIFO, FIFO, or exact matching. The choice is basically arbitrary, unless you are in a business that can use exact matching (doesn’t apply to oil). Or you are a publicly traded company, in which case LIFO is virtually mandatory.

Neither LIFO nor FIFO automatically confer an advantage to profitability.

FIFO is great for companies where the product can be easily identified into lots and is subject to deterioration. For example, it would work great for a hypothetical fresh fruit stand, where the product frequently goes bad. Think of it this way: you can’t sell fruit that has already spoiled, so clearly you must be selling the most recent batch you purchased.

LIFO is great for when you are constantly buying at different prices. It is also better at representing the replacement cost of your product - in other words, if you want to continue operations, what would be the cost of doing so? Clearly, for oil companies, the cost of replacing the oil they sell is quite high.

Aside from actual accounting concerns, the folks making the proposal are ignoring IRS and SEC rules. In order to deter the sort profit manipulation the oil companies are accused of, regulations requires that companies wishing to change their inventory accounting method go back and restate income over three years (maybe longer, i don’t recall exactly). They also have to go through an onerous approval process through the IRS. This is an extremely costly procedure, which is undertaken only if there is a very good reason. Without careful analysis, it can easily turn previously profitable years into losses, and vice versa. Same for future years. This has tax consequences which might actually nullify any potential new taxes. Without an excruciatingly detailed analysis of oil company records, it would be impossible to determine the actual net change in taxes.

While improbable, it is possible that switching from LIFO might actually result in a net decrease in tax liability over the last several years, resulting in the government to issuing a refund.

Thus, assuming a company will be an ongoing operating concern in the foreseeable future, the choice of inventory accounting method will have zero net impact on reported profits and taxes due.

And all of this doesn’t even get into the constitutionality of the issue. Congress setup a regulatory framework for businesses to operate under. Businesses made investment decisions based on the law. It may be unconstitutional to suddenly change the law in order to punish an industry’s success. More likely than not, a mandatory changing of accounting standards would require that historical investment decisions be exempt from the new law.

How iPad can Dominate the PC Market

Abtin Forouzandeh — Fri, 09 Sep 2011 08:57:00 EDT

Last year, Paul Graham observed that, through software, tablets are capable on taking on an increasing number of tasks. Never one for idle conversation, he put his money where his mouth is, looking for startups that can expand the iPad “just works” philosophy into more markets.

Certainly, pg is correct that software will be a big part of this expansion. But Apple must to lead the way in hardware for the revolution to really happen.

Apple needs a low-priced laptop competitor, not because that is what other companies are doing, but simply because the current $1000+ entry point for a laptop is way too high. The traditional wisdom has been that it should release a low-cost netbook. But that runs counter to the entire Apple design philosophy and would destroy the Apple brand. And Apple has never held to the philosophy of “if you can’t beat’em, join’em.” Apple’s motto might as well be, “other companies are irrelevant; redefine the market.” *

Yesterday, a friend of mine called me up, wanted advice on purchasing a new laptop. His old macbook was falling apart, but a new mac would’ve blown his budget - he was looking at $400 Acer laptops. I warned him that transitioning back to a PC would be difficult, but that didn’t matter. At most, he was willing to spend $500-600. And it is hard to justify spending much more when your main use cases are browsing the web and storing media. So how can Apple address this part of the market?

By adding just enough features to the iPad to allow it to do the core of what netbooks do:

Make it way more obvious that photos and videos can by copied to the iPad simply by plugging in a camera or iPhone.
Enable iPhone to treat the iPad as its host PC for synchronization purposes.
Enable the iPad to access some kind of mass storage. Photos and videos can add up pretty quick and internal flash memory is not sufficiently cost effective. This could be as simple as allowing it to use a USB hard drive or flash drive. Or better yet, allow it to tap into time capsule or other network drive.
Create a new iPad cover that can house both a the tiny Apple wireless keyboard and iPad. It also needs to open in such a way that the the iPad is vertical and the keyboard is horizontal, so that it can be used like a laptop.
Create an iPad Bundle package, that packages the new cover and a keyboard along with the iPad.
Price the bundle at $500, which would mean dropping the price of the iPad by $100.

And that’s it. Instant domination of the low-price PC market. Heck, the upcomming iCloud is already set to address some of the points above.

With such a bundle, there would simply be no reason to purchase a low end laptop, which are generally only used for light computing tasks. $500 is not quite inexpensive enough to compete in the $250 netbook space, but that can addressed as tablet prices fall over time.

* This attitude doesn’t even acknowledge the existence of competitors (”I’m a Mac/I’m a PC” commercials being a rare exception); instead, it is the view that each of us ought to strive for the best within us, thereby creating products and services that transform the market. As a result of this, in a future post I’ll demonstrate how and why Steve Jobs is a Randian Hero.

Why I'm both an Anarcho-Capitalist and a Minarchist

Abtin Forouzandeh — Thu, 08 Sep 2011 19:01:00 EDT

In libertarian circles, there is often an argument about whether a state is necessary at all.

There are those who claim that the best, or most practical, form of government is a minarchy, in which the sole role of the government is national defense, law enforcement, and dispute resolution. There is some debate over whether such a government ought to tax its citizens. Some assert that taxation is an unjustified use of force, while other say it is necessary for the state to function, and is therefore justified. I’m not aware of any good solutions for how to fund a minarchy’s operations other than through taxation.

Then there are those who take the core premise behind libertarian philosophy, that the initiation of force is wrong, to its logical end. Anarcho-capitalists, because they all rely on force, all forms of governments are immoral; therefore, there should be no government. Further, they assert that laissez-faire capitalism (i.e. a system of voluntary exchange free from coercion) is the only proper way for individuals to interact with each other. Under anarcho-capitalism, all the functions of government are taken over by private entities, including law enforcement, defense, and the courts. For example, individuals would sign up for services with a private defense agency, which agrees to protect that person from harm. Individuals would be free to choose which private defense agency they subscribe to. Seeking profit, the PDAs would seek to minimize the amount of force they must use to protect and retaliate against other PDAs. Thus, they might sign agreements outlining how to resolve disputes between citizens, compensate victims, and so forth.

In my view, anarcho-capitalism is not a normative assertion about the ideal form of government. Rather, it is a purely descriptive theory of how the real world operates. The governments of the world exhibit all the qualities of private defense agencies. When citizens of different governments have disputes, the governments (generally) seek to delay hostilities as long as possible. And they have all sorts of agreements in place to deal with such problems, like extradition treaties and tax treaties and intellectual property treaties.

Lets say we lived in the anarcho-capitalist ideal of a stateless world. In that case, we would still have to figure out how to structure the internal rules of our PDAs - before we even start contemplating how PDAs should interact. What sorts of behavior does the PDA prohibit because the costs to other subscribers might be too high? What sorts of punishments are justified when one subscriber harms another?

And the moment that discussion starts, we return to the original libertarian dispute. Some folks will argue for minimum intervention. Others will argue for ever smaller PDAs.

The argument over the structure of the internal PDA argument is what the practical libertarians are focused on. Should the PDA operate under communism, fascism, or some form of liberalism? How involved in the individual’s life should the PDA be?

Compared to other political philosophies, minarchism is most closely aligned with all the examples of how to create a prosperous society. Its time to end the theoretical arguments over whether we should have a government at all, and recognize that we (classical liberals) are all anarcho-capitalistic minarchists.

Open is Overrated

Abtin Forouzandeh — Thu, 08 Sep 2011 10:32:00 EDT

My previous post spurred some good discussions, both at HN and in the comments.

One line of criticism I noticed was repeated by a number of folks goes something like this. The stack powering AWS is closed. Google might not show as much developer love, but at least their code is open. This is illustrated by a comment on the post:

Google not only creates interesting products, services, and APIs all the time, but they tend to make them open and available.

Now I love open source and open platforms as much as the next engineer. Being able to debug your application down into the internals of the platform is hugely valuable. But that is a second order problem. The first problem an engineer needs to solve is how to build the damn thing at all.

It’s not that Google can’t help in this area. My startup, Wompt, is built on top of V8/Node, uses Google Analytics, and uses Google Apps for email. We also owe a lot to Chrome and its phenomenal debugging capabilities.

But there is a qualitative difference between these tools and what AWS provides. Each of the Google tools solves a core problem, but they don’t amplify the capabilities of my startup. The AWS stack acts as a huge multiple to my startup’s capacity and engineering velocity. AWS also eliminates all operational capital costs. This is hugely important for a startup. Because all of the operational costs are variable with growth, you are free to scale with reduced worry about running out of money. In fact, if you have a business model such that you can start out with revenue exceeding costs, then there is literally no limit to your potential.

Just consider Simple Queue Service. Could I just go grab something like RabbitMQ and setup a queue? Sure. But it would take time. Time that I could spend creating value in my core product. And my own implementation of the queue probably wont be nearly as reliable or performant as Amazon’s. And I would have to pay for a server.

Some folks point out that AWS isn’t free - you need to pay to use the compute power. Well duh. Even if everything in the platform was open source, you would still have to pay. Open source can’t make operational costs dissappear (yet?). I trust Amazon to do a much better job than me at ensuring basic reliablity, like keeping the power on. For the nominal cost of using their services, I get an enormous amount of value.

That fact that Amazon’s services aren’t open, as in open source, increases operational risk, but it actually only has a very minor impact on the value I recieve. And because Amazon’s services operate in a highly competitive environment, I know that if it really becomes necessary, I can move everything to another provider. By comparison, something like gmail or even Google Analytics have very high lock-in.

And truly, which is more closed: a platform that you can pay to use or a platform that, for all anyone outside the company knows, is just make believe. In that sense, petascale MapReduce is like a Unicorn, a mythical creature capable of amazing feats, but existing as nothing more than words on paper.

Amazon is More Interesting than Google

Abtin Forouzandeh — Wed, 07 Sep 2011 22:15:00 EDT

Via Google’s research blog:

Almost three years ago we announced results of the first ever “petasort” (sorting a petabyte-worth of 100-byte records, following the Sort Benchmark rules). It completed in just over six hours on 4000 computers. Recently we repeated the experiment using 8000 computers. The execution time was 33 minutes, an order of magnitude improvement.

…

What would it take to scale MapReduce by further orders of magnitude and make processing of such large data sets efficient and easy? One way to find out is to join Google’s systems infrastructure team. If you have a passion for distributed computing, are an expert or plan to become one, and feel excited about the challenges of exascale then definitely consider applying for a software engineering position with Google.

As an engineer, I hate this sort of announcement. Google has invented something very cool, but they won’t show it to you unless they think you are among the world’s elite engineers. They decide that you are an elite engineer primarily based on whether you went to an elite school. Elite code. Elite engineers. Elite schools. Elite company. How very elitist.

Google has been doing these sort of blog posts for years. Some engineer writes up an entry about how they are doing research using terabytes or petabytes of data. And then they end by saying you should work at Google. So nowadays, I don’t care about any of what Google does. Google+? Meh, they’ll probably discontinue it after 6 months because it doesn’t instantly win. MapReduce? Great, they’ve been sitting on this technology for a decade. Good for them. It doesn’t matter to me.

But the world has changed, and Google can’t seem to keep up. Amazon has become the polar opposite of Google, empowering every developer on the planet to make incredible technology. Want MapReduce? Amazon has you covered. Want to play with terabytes of data like it ain’t no thing? Check. Want to launch thousands of servers to handle a tough computation? Check, check, and check. Want to launch thousands of human brains to solve otherwise unassailable problems? No problem. Heck, want to simply send email to your users? They have that too.

Google offers none of it. Its like they are stoned and watching the world pass them by in a daze. Compared to Amazon Web Services, or even RackSpace, Google is in the stone age.

Perhaps I am biased by my spirit of entrepreneurship. I want to create great technology, but I want the benefits of my effort to accrue to me, not an elitist corporation.

Google is a cocktease.

Amazon finds new ways to be nice to me every other week.

Determinism & Free Will

Abtin Forouzandeh — Wed, 07 Sep 2011 02:00:00 EDT

Determinism asserts that every event must have a cause. This is more than merely a natural way to think of the physical world; every human interaction with the world relies on this principle. It would be inconceviable for an event in the world to not have a cause. However, determinism is controversial when applied to the human mind, because it conflicts with experiential evidence that humans make choices in their affairs. Under determinism, the totality of preceding causes result in an environment where only one outcome is possible, including the decision by a human mind to act in a certain way.

Free will is the idea that humans are capable of making causeless choices. Under free will, a human mind must be able to select among available choices without influence or restrictions from the physical world.

The two viewpoints are in deep conflict. If free will holds true, what types of “stuff” in the real world have the quality of excercising free will? For example, animals seem to not posess it, or at least they are unable to tell us about it. Do other people have free will or is it just the self? We often attempt to ascribe causes to the actions of others, but, under free will, these causes shouldn’t matter and may not even exist.

It is also the case that the only evidence for free will is the self and the self’s internal thought process - every other piece of experiential evidence reveals a universe that operates deterministically. In other words, if humans have free will, then humans are the only source of randomness in the universe. It would be difficult to reconcile a deterministic universe with a random component. If the universe has a random component, then the universe is random. If it does not have a random component, then the human mind must operate deterministically.

However, in one sense, the two positions need not be in conflict. It could be the case that the world operates deterministically in such a way that humans have the illusion of free will. As we cannot tell this illusion apart from reality, we would still be responsible for our actions (unable to blame them on the starting conditions of the universe).

Government & Natural Disasters

Abtin Forouzandeh — Wed, 31 Aug 2011 13:00:00 EDT

A friend of my shared a call-to-action from a progressive group angry that Eric Cantor wants “cuts to vital government programs” as part of providing hurricane relief:

It’s the role of government to respond to the needs of its people in the wake of disaster. Republicans like Eric Cantor need to get out of the way and let the federal government do its job – help the states aid those in immediate danger and as quickly as possible begin the essential work of repairing the damage done.

…

It’s the role of government to respond to the needs of its people in the wake of disaster.

I’ve always been a bit baffled by why its the government’s responsibility to step in after every so-called “disaster” and provide so much assistance.

Don’t get me wrong: if there is a serious threat to life, the government may have a role regardless of whether the situation was foreseeable or preventable. That might include providing evacuations, providing emergency supplies, or whatever else is needed to safeguard life.

But for something to be a disaster, doesn’t it have to include an element of being an accident or unforeseeable? Otherwise, it’s just negligence to prepare on the part of the victims. Hurricanes happen with extreme regularity in certain areas. People know this, especially those living in such areas. At some point, doesn’t this personal choice become tantamount to giving society the finger?

Again, sure, a government providing immediate relief in order to safeguard life can make a lot of sense. That part of the article is non-controversial except among those opposed to all government.

But providing money for rebuilding? Maybe that forces folks to stay in an area where they are prone to be victims? That is neither a good use of money, nor is it fair to the victims. How about the alternatives? Just to name one, how about giving folks assistance to leave the area permanently?

Same goes for a wide variety of other regular environmental issues, like the yearly flooding along the Mississippi or snow storms in New England.

The article also mentions tornadoes, which I find to be particularly problematic. That is a type of disaster where the effects can be mitigated through insurance (because a tornado doesn’t wipe out all the policy holders at once, the way a flood might, the risk can be distributed across the subscriber base). It seems to me that local governments can address the immediate effects, while homeowner’s insurance ought to cover the rest. How is there any role for the federal government at all when it comes to tornadoes?