Blog posts

VCs should adopt features of the Hollywood studio system

Originally published 2023.06.25

(Based on a Twitter exchange I had recently with Emmett Shear.)

VCs should have methods of reducing their risk other than limiting the amount of money they give, and the Hollywood studio system shows ways to do this.

In the movie industry you don't have production companies saying to directors, "I like your idea for the next big Star Wars movie, here's $100,000 to make a low-budget version on your own, let's see if you can sell that; oh, and you have to find all your own people to help you to make it".

Instead, there's an entire ecosystem around "idea guys" (script writers) who sell their ideas (scripts) to be made into actual products (movies) by a separate specialty class of "doers" (directors), who are in turn able to be supported by specialists in the various things that need to get done (e.g. the cinematographer), and the production company already has ongoing relationships with these specialists to go work on the ideas they're backing. There's a lot more specialization and a lot more "process" in place than in the VC industry.

Granted, the amount that the directors/scriptwriters are getting in the movie industry is much less than what founders stand to make: there's less risk and less reward. But I think a lot of people would take that trade: a reduction in their potential payout in exchange for a better support structure that reduces their risk.

"Sprint" is a good book that gives an example of one such method at Google Ventures: help the founders do one or more one-week sprints that allow them to test a mock-up in front of real users.

How I think about morality (having the benefit of my undergraduate degree in psychology and philosophy)

Originally published 2022.07.10

A friend on Facebook asked:

• In what circumstances is censorship the right answer, or is it /ever/ the right answer?

• Are there specific types of content that ought to be limited/ censored?

• Do the morals (or other qualities) of the director/producer/writer/actor/artist etc. warrant censorship?

I responded:

I spent many years thinking about philosophical topics like this and my conclusion was that there are certain areas in life where people tend to ask subtly malformed questions. For example, it seems to me that censorship isn't a matter of "right" and "wrong", it's simply a matter of a power struggle between different groups. It's like asking "in what circumstances is a lion catching a zebra the right answer?". I think "what is the meaning of life?" is another example of a malformed question. A math example would be something like "what color is the number five?".
Is–ought problem
https://plato.stanford.edu/entries/hume-moral/#io

She responded:

is this the same as moral relativism? For societies to function, there must be some moral 'absolutes' though, right? Otherwise there is no law… But how does one agree on those, and to what extent should deviations from those laws be screened from the public eye? As mentioned in the response above, I'm -in general- opposed to censorship, but am trying to find instances where censorship is the only recourse, let's say for society to adequately function.

I responded:

"For societies to function, there must be some moral 'absolutes' though, right?"

I'm not totally sure I understand what you mean by "absolutes", but: after a lot of classes and a lot of thinking about it, my understanding is that there's a part of our brain (the prefrontal cortex) that is inclined to make consequentialist judgements, and another part of our brain (I forget which) that is inclined to make deontological judgements. The trolley problem is a great example of when these two parts of our brain are in conflict: the utilitarian part wants to pull the switch so that the trolley kills fewer people, while the deontological part doesn't want to pull the switch because it would mean that you're "responsible" for the death of the person the trolley ends up killing.

So, people do have these parts of our brain that tend to think in these ways, but beyond those broad strokes, the details seem to be determined by the environment that a person is raised in. So those could function as the "moral absolutes" you're talking about: the shared brain structure and the shared environment / learned rules of behavior.

But there can be a lot of disagreement because people don't all have the same environment growing up ("My parents taught me to do X in this situation!" "Well, my parents taught me to Y!"), and beyond that, a lot of people's behavior is just determined by "what's good for me?" or by whatever pressures they're subject to. So you can end up in situations where, for example, people vote for some policy (like censorship) not because they've made some judgement about the morality of it but instead because they were paid to vote that way, or they were urged to vote that way by their pastor, or whatever.

So when I think of these kinds of questions (like the censorship one), I think not so much in terms of "What ought to happen?" and more in terms of "What will happen?", like if it was a physics problem. I actually imagine billiard balls bouncing around, where the question is trying to predict the future state of things.

Another way of putting it: when you're asking an "ought" question, you need to be very precise and explicit about what you're trying to optimize for / what the "ought" is trying to optimize for. Are you trying to maximize your own average daily reported wellbeing over the course of your life? Are you trying to maximize the duration for which a nation known as the "United States" continues to exist? And it's probably not possible to optimize for everything that you'd like to optimize for. But if you don't have a precise goal you could end up feeling lost, like you can't figure out a satisfying answer and you can't understand why.

I highly recommend checking out Joshua Greene (psych professor at Harvard) and his dissertation:

The Terrible, Horrible, No Good, Very Bad Truth About Morality and What to Do About It

To answer the last part of your question:

"As mentioned in the response above, I'm -in general- opposed to censorship, but am trying to find instances where censorship is the only recourse, let's say for society to adequately function."

Censorship is very common during times of war / life-or-death struggle, when the very survival of (you/your family/your nation) is at stake. Censorship was widespread during WW2. IMO that's a good example of how these moral judgements ("censorship is wrong!") can be subordinate to a judgement of "What's good for (me/my family/the nation)?" They're just a set of rules that people agree to abide by as a compromise with each other, but when things get crazy, the rules go out the window. There's nothing objective about them.

Why I didn't go to law school

Originally published 2019.03.11

I wrote up the explanation below a few months ago when someone sent me an email thanking me for my LSAT guide and asking me why I didn't go to law school.  I ended up sending the guy a one-sentence explanation instead, but I figured the full explanation would make a good blog post, and I'm just now getting around to posting it:

It was for a bunch of reasons.

Before I got my LSAT score I was dead-set on going to Harvard. It was Harvard or bust. But once I got my score I realized that I not only had a shot at Harvard, but also a shot at the even-more-prestigious Yale, and I started to wonder if I should try to make a serious attempt to go to Yale instead. I knew my LSAT score was valid for 5 years, and so getting a perfect score got me to start thinking about waiting a few years before applying so that I could improve my resume and increase my chances of getting into Yale. I decided to move to Washington, DC for that reason (improve my resume).

Also, now that I was sure I could go to Harvard, it lost a lot of the mystique that it'd had for me, and I started to wonder if I really should go to Harvard (it's like that Groucho Marx quote: "I don't care to belong to any club that will have me as a member."). ~2-3 months after getting my LSAT score I went to visit Harvard Law School, sat in on some classes, and talked to some students, and I left a little unimpressed; I got the impression that most of the students were smart people who didn't have a strong sense of what they should do with their lives, and so they had just followed the beaten path to become a lawyer; that also described me. I got a bad feeling from this, because it seemed like a bad way to make such an important decision (the decision of what skills to learn), especially given the news that was widespread at the time that the country was producing too many lawyers. For many, many years (at least far back as high school, if not earlier) I'd been mystified at how people end up in different careers: how did people decide that they should dedicate their life to X instead of Y? I'd always been more of a dabbler and dilettante, and I couldn't understand how people could do the same thing for years without getting bored. So in high school, I had no idea what career I should go into, and that continued into college. Now I was at a point where I had to make a decision (about what career to go into), and so I made the decision that I wanted to spend the time necessary to arrive at a crystal-clear idea of how to make the decision of what career to go into, even if that meant spending years thinking about it. This train of thought ("I need to figure out how to make this decision") was something that had probably started soon after I graduated from college and started reading autobiographies.

So to summarize the above, the decision to wait a few years to apply to law school was a result of a combination of 1) me deciding that it couldn't hurt to try to improve my resume, and 2) me deciding that I should really do research and develop a crystal-clear understanding of how to choose a career.

Once I was in DC I interned for a Congressman, worked a cubicle job at Thomson Reuters, and temped at a few different law firms, and I eventually realized I would be miserable if I had to wear a suit every day, stay in an office (I hate offices), be around people I don't really care about, etc. I realized that the marginal benefit of the money I could get from being a lawyer wasn't worth the marginal cost in terms of my freedom.

From doing a bunch of reading, I realized (like many people who go into law seem to realize, such as Peter Thiel) that if you are willing to pay the price (in terms of lifestyle) that you need to pay to become a lawyer, you can have a lot more upside if you go into finance instead of law (e.g. become an investment banker / go work at Goldman Sachs). Peter Thiel did this exact thing: he started in biglaw, then switched into finance because he could make more money there. And then in the summer of 2011 I learned about Michael Burry, who runs a hedge fund from his house, and I thought, "Wow, I'd love to be able to make a lot of money while living and working from anywhere". So for a while I was thinking I would try to become a hedge fund manager a la Michael Burry.

After learning more about Peter Thiel and how he then switched from finance into tech, I realized that tech (internet entrepreneurship) had some characteristics that I really liked: 1) it was a younger industry, and so it seemed like there might be more opportunities than in the further-developed hedge-fund industry, 2) it seemed much easier to be able to get a job where you're working remotely, and so even if I never made a ton of money, I knew I would be much happier if I could work remotely, so my downside was better protected, 3) it seemed possible to make a huge amount of money if I hit upon the right idea, so there was the same potential for upside that I would have gotten if I'd become a hedge fund manager. Nowadays I think of entrepreneurship as just being an extension of finance; an entrepreneur is like being an extremely-hands-on activist investor.

I also quickly realized that I didn't want to suffer in the short term for some longer-term benefit, so I wanted to try to maximize my happiness ASAP. So in early 2012 I set a five-year goal for myself and a 10-year goal for myself: the 5-year goal was to have a job / career where I could work remotely from anywhere in the world, and the 10-year goal was to start my own tech company or hedge fund. I've achieved my 5-year goal and I'm now working on my 10-year goal. There's one guy in particular that I'm interested in emulating for the next few years, his name is Pieter Levels (check out http://makebook.io ).

My impression is that many people who contact me about the LSAT are very entrenched in their social group (their family, their friends, etc.) and they don't feel comfortable doing something radically different from what their peers / family are doing. For example, they want to get their career going in X number of years so that they can get married and have kids within Y number of years. I didn't want to live that way, but many people do want to live that way. I think it's probably the #1 factor that determines whether the non-traditional path I chose is a good fit for a person or not.

Also, as a post-script: I have a friend who went to Stanford for computer science and had started several (failed) internet companies, and after he graduated he wanted to go to Harvard Law School and become an IP lawyer instead. I told him not to do it, that he should just keep trying to start new companies instead. He ignored my advice, got rejected from Harvard Law, accepted to Georgetown Law, worked his ass off his 1L year to be at the top of his class, successfully transferred to Harvard...and then realized he hated it and ended up skipping classes so that he could start new internet companies instead. Now, I do suspect the credential is providing him some benefit, but he apparently has well-to-do parents who could help him with the debt payments if everything went south. In my case I would've had to pay off the HLS school debt myself, and I just didn't think the benefit (in terms of prestige) that it would provide would be worth the cost if I knew that I wanted to start an internet company and not become a lawyer.

You can find more of my thoughts on this topic here: My Thoughts on Various Topics - Whether to Go to Law School

Some thinking I've done on how to do hiring

Originally published 2018.01.01

Below is an email I sent out to everyone at Infer in August of 2016. I've made some changes to clarify terms that the general public might not understand. Read only the bold sentences if you want a summary.

A little over a year ago Yang [one of the cofounders at Infer] asked me my opinion on the hiring process for the Data Analyst position, and I didn't have a thorough answer at the time, but I said I would think about it. I have a tendency to take a long time to think something through, but when I'm done I usually feel like I have a crystal-­clear answer. And I think that is what has happened in this case: I feel like I have a crystal­-clear answer on how to think about hiring.

First, some background information:

In the process of organizing the wiki, I have noticed that there is a branching structure that a company will have in its wiki as it grows.

At the top­ level of a wiki, you specify general responsibilities and general tools. These are things that you expect everyone at the company to know. At Infer, tools that everyone has to know how to use include Gmail, Google docs, meetings, lightning talks, and the espresso machine. A responsibility for everyone at Infer would include something simple like locking up if you're the last person to leave the office, or adapting to "the culture", which is just the collection of all the unwritten rules that people follow. For example, one of these unwritten rules is that when someone on the sales team closes a deal and hits the gong, you stop what you're doing (if your work isn't urgent), applaud them for their accomplishment, and then listen to the announcement. If someone continually didn't do that, they would be breaking an unwritten rule. (Tangent: Right now, as the name "unwritten rule" suggests, these rules aren't written down, but I think they should be, because it can make it easier to debate their merits. For example, my working from a conference room seems to be breaking an unwritten rule to only use the conference rooms for meetings and phone calls, but that's an unwritten rule I think should be debated. Right now we don't have an organized way of debating these rules; writing them down on the wiki would give us a place to debate them.)

Aside from these general tools and general responsibilities, the wiki / company also starts to branch into different teams as the company grows; so, at Infer we could divide the teams up in different ways, but one way to describe the teams would be to say we have an Eng [Software Engineering] team, a DA [Data Analyst] team, a Sales team, etc. And like a branching tree, each of these teams' wikis has the same structure that the wiki section for the company­-as-­a-­whole had: each of these teams has responsibilities that everyone on the team needs to be able to do, and there are tools that everyone on that team needs to know how to use. For example, on the Data Analyst team, the machine-learning-model-­building part of our internal web app is a tool that everyone on the Data Analyst team needs to know how to use. But it's also a tool that's kind of specific to the Data Analysts: SEs [customer-onboarding specialists] and engineers and SDRs [sales-meeting getters] are not messing around with that part of our internal web app as much. As for responsibilities: it is a Data Analyst responsibility to build a [machine learning] model. That is a responsibility that's kind of specific to the Data Analysts; that doesn't mean that no one else can build a model; an Account Manager could jump into our internal web app and build a model, or one of the engineers could jump in and build a model, but it's kind of specific to the Data Analyst team.

The same pattern of "Responsibilities / Tools / Teams" continues as the company grows, with each of these teams branching into further sub­teams. For example, I think that Eng has started already to break into separate teams, so (as far as I can tell) we have the Machine Learning team, which has its own tools and responsibilities that apply to it; we have the Dev Ops team, and they have their own tools and responsibilities that apply to them, we have a Front-end team, and they have their own tools and responsibilities that apply to them. And similarly on the sales side, it quickly branched into the SDR team, the AE [sale-closer] team, and perhaps also the SE team and AM [customer-subscription-renewal specialist] team.

</background information>

Why am I bringing all this up? Because it applies to how you do hiring. When you're hiring, you need to figure out which responsibilities and which tools you need that person to take on. And you can use the wiki (or at least this way of thinking about responsibilities / tools) to get a comprehensive list of what this person needs to be able to do and what you should measure them on.

So, for example, let's let's say we have a very specialized position, like being on the ML team on Eng. If you're hiring someone for the ML [Machine Learning] team on Eng, you could compile a list of all of the responsibilities and tools of the ML team, and all the responsibilities and tools for Eng in general, and all the responsibilities and tools for the company in general, and that's your list of requirements. Those are the things that the person needs to be able to do. And this includes cultural responsibilities like being proactive, communicating well, etc. All of those cultural responsibilities can (and I think should) be specified as responsibilities on the wiki.

Once you have that list of requirements then you simply devise tests for each of those requirements; you can you come up with some series of tests that measure that person's proficiency on all of those different responsibilities and tools to see if the person will perform well or not. And again, when you hear "responsibility", you might be thinking of things like "How good is this SDR applicant at cold calling?", but it also involves cultural responsibilities like "How quickly can this person learn new tools?", or "How much will this person encourage his coworkers?". (Tangent: Again, all of those things can (and IMO should) be specified. It makes it easier for employees to figure out how they should be behaving to get promoted, it makes it easier to gauge when someone should be promoted, and it makes it easier to figure out who should get hired and where they should be slotted.)

So, that's it. This seems to me to be a very clear and systematic method. I felt confused about hiring in the past, but this way of thinking about it makes perfect sense to me, and it makes it very clear to me what I would need to do, step-­by-step, if I was going to be doing hiring. I think that if you document responsibilities and tools in this way, and if you come up with tests for measuring all of those things, then that seems like a method for hiring that will work pretty reliably.

PS:

It seems to me that this way of looking at hiring lines up very nicely with the hiring advice that Ben Horowitz gives in his book, "The Hard Thing About Hard Things". He says that before you hire for an executive position you should do the job yourself for a while, because then you'll be in a better position to understand what the person must be able to do, and you'll know what kind of test to use to gauge their ability to take on that responsibility. Once you know what you need the person to do, you then develop questions to test for those abilities. I think what I described above is just a more­-explicit version of the requirements-gathering part of the hiring process he describes.

STEP 1: KNOW WHAT YOU WANT

Step 1 is definitely the most important step in the process and also the one that gets skipped most often. As the great self­-help coach Tony Robbins says, “If you don’t know what you want, the chances that you’ll get it are extremely low.” If you have never done the job, how do you know what to want?

(...)

The very best way to know what you want is to act in the role. Not just in title, but in real action. In my career, I’ve been acting VP of HR, CFO, and VP of sales. Often CEOs resist acting in functional roles, because they worry that they lack the appropriate knowledge. This worry is precisely why you should act— to get the appropriate knowledge. Indeed, acting is really the only way to get all the knowledge that you need to make the hire, because you are looking for the right executive for your company today, not a generic executive.

Advice for those interested in learning more about mathematics

Originally published 2017.09.09

A few months ago an acquaintance of mine (Cornell / Harvard Law School, so a smart guy) emailed me and asked if I had any advice for how to explore mathematics further:

Wanted to see if you'd recommend any online math courses (available free) lots of MIT stuff available and the like but wasn't sure which one was good.

My response (edited for clarity):

I wrote the advice below and then remembered that you'd done math competitions in high school, so I'm not sure how obvious some of it may seem to you:

1. What you should spend your time on really depends on what your goals are. For example, when I was studying math I really was only interested in how I might use math to help me in money-making pursuits, and as a result I was interested in getting a good high-level understanding of how the field worked, what mathematicians were up to, what the different advanced fields were about, and why I'd seen so many mathematicians who were good at learning other things. My thinking was that once I understood the field at a high-level, I would know where to dive deep (if necessary). On the other hand, if you're just studying math for the fun of it, then you might want to just jump into a type of math that sounds interesting and start learning about it.

1. I have a section on my website where I've gathered some of the stuff I've run across:  

NW's Thoughts - Mathematics

1. For a high-level understanding of what's going on in the field of math:

It's not free, but Timothy Gowers is a great writer and I remember finding this (short) book very helpful:
Mathematics: A Very Short Introduction

I also found this (short) book very helpful:
Letters to a Young Mathematician

1. For learning the general ability of problem-solving that makes mathematicians tend to be pretty good at tackling other subjects:

How to Solve It
I was at one point trying to create a summary of the book here: NW's Thoughts - How to Solve It

The movie "Hard Problems" was at one point on YouTube.  I found it helpful in at least three ways: 1) I found it fun to watch, and I think it's good to have fun while spending time thinking about mathematics, 2) it made the 'genius' math kids seem less mysterious, and 3) there was a very memorable scene where I saw one of these 'genius' math kids going through a tough problem and he is speaking out-loud about his approach to solve it, and the big trick is to just solve a simpler version of the problem first and only then start to solve more-complicated versions of the problem that are closer to what your final goal is. Hard Problems

1. As for learning actual math, the conclusion I came to after a lot of research was that the most important fields to be familiar with were 1) probability, 2) combinatorics, and 3) linear algebra (with the order of importance depending on your situation).

KhanAcademy has videos on all of those.

This video series on Linear Algebra was highly recommended by someone, but I haven't watched it myself yet: YouTube - The Essence of Linear Algebra

If you're not already aware of it, I also highly recommend checking out the 'Art of Problem Solving' website, as far as I know it's the premier hangout spot for school-aged math competitors, and there are a lot of friendly people there of all ages.

Problems that prevent routine white-collar tasks from being automated as a service

Originally published 2017.08.31

Recently I had a serial-entrepreneur acquaintance send me a form email for a new company he'd started with six others that aims to automate all routine tasks that white-collar workers have:

I have a quick question: When you hear the phrase: "Annoying business processes", what comes to mind for you?

My company automates these processes. We can handle anything (anything!) that can be turned into a set of instructions. Even if the instructions contain complex logic -- we're totally custom. Even if the labor required would normally take hundreds of hours -- it's the same price.

I responded:

I've had this same idea. To be helpful, just in case these are things you haven't encountered yet, some tricky things I'm aware of are:

1. dealing with edge cases. It's very hard to get a customer to be able to produce all of the edge cases up-front.

2. dealing with secret/proprietary information. It can be hard to convince a company to delegate / outsource something that requires that the employee know some kind of secret (often customer-related) information.

3. documenting a process in the depth necessary to delegate it can take such a large up-front cost that people may not be willing to pay it. Video recordings are a good way to get around this.

4. the employees whose processes are the easiest to automate are often the most resistant to having them automated, because it'll either threaten their job security or it'll require them to learn some new, often more-difficult process, often at the same pay

5. interfacing with intranets / private software / email can be extremely difficult.

Good luck, I think it's a really good idea!

His response:

Yup, you pretty much nailed it with all of these.  Extremely difficult up front, but they all make this extremely defensible after we solve it. And we do conclude, ultimately, that these are all solve-able. Would you disagree?

My response:

I think it's probably solvable for some tasks and not others, and so the trick is to figure out what kinds of tasks it's solvable for and how to efficiently identify which companies have such tasks.

I think #5 may be the hardest problem to solve. The only way to get around it that I could think of is to set up the work environment on a cloud machine and then use autohotkey-like software, but that's extremely fragile and would need to be updated every time the UI of a needed intranet / web app changes.

Another problem I just remembered is that because not everything can be easily automated, and because non programmers generally can't tell easily which tasks can and can't be automated, you can't just rely on asking an employee or manager which tasks to automate. You actually need to end up watching everything they do like a consultant and pick out the parts that can be automated.

To be honest I think it's a great idea but too hard, and so what you see successful companies doing instead is finding a single expensive process that many companies have and automating it. That's what Parker Conrad did at zenefits (dealing with employee healthcare) and that's what he's doing at his new company (employee onboarding and offboarding).

Oh yeah and another problem I just remembered is needing to interface with information contained in written English (emails, intranet comments) to automate the task.

His response:

Again, I'm impressed with your analysis. Very good. Companies certainly are specializing on discrete problems, as opposed to the general problem.  That creates a coordination problem: startups struggle to use all of the services available to them.

Am CCing two other partners on the team so they can read your analysis. These 6 problems you've identified are all ones we've identified too. We've already got some clarity on each one,  but it is good to bring them into relief again and reconsider our answers

My response:

Another competitor to think about is Zapier. They seem to be going after the segment of the problem for which every piece of software needed to complete the task is online and can be communicated with via an API. So there may be an opportunity to go after tasks that can't be totally automated via APIs.

(Just an idea:) Maybe you could explore making it possible to have an API for online or offline software that currently doesn't have an API: you'd have a cloud machine that would install the software if necessary (if it wasn't an online service), log in if necessary, and then scrape the page for whatever information was necessary, and provide that information in API form. You could sell it to Zapier, and they could charge the customer a premium to interface with those services. You could also sell it to software developers who want to build on top of those services but for whom creating their own fake API would be too expensive. Your job would be to watch the non-API software and update the scraper whenever the website / offline software changes. Again, just an idea.

Update 2024.01.20 - I've just found out about Appian, a service that makes this kind of automation possible. Their approach is to not automate your processes themselves, but rather give you the tools you would need to automate it, which bypasses the issues of data security and accessing intranets:

https://www.youtube.com/watch?v=urIFm6V7xyw

How to create a new Python project using PythonAnywhere, Bitbucket, SourceTree, and PyCharm

Originally published 2016.07.02

1. I often find it difficult to start new projects on PythonAnywhere because setting them up properly requires a lot of steps.

   • (By "properly" I mean "using tools that will help you avoid headaches down the road".)

2. To help my future self get past that initial roadblock, I recently documented the steps involved when I start a new project.

3. You can find it at the link below.

   • How to create a new Python project using PythonAnywhere, Bitbucket, SourceTree, and PyCharm

An email to a friend with some ideas for creating a repeatable sales/growth process

Originally published 2016.01.23

Here's an email I recently sent to a Stanford/HLS-grad friend of mine who started a delivery company that sells to individuals and businesses. He's currently trying to figure out a repeatable sales/growth process. Because his product involves physical delivery, he's aiming to target groups of businesses that are within close proximity of each other (eg in a strip mall / office park). As with Box/Dropbox, he can sign up one person at a company and then try to get that person to sign up others. I've made some clarifying edits and removed some identifying information.

Some ideas for what you should try next (viral growth)

I spend a lot of time thinking and reading about viral growth, and so I figured I'd write up some of my thoughts of how I might be inclined to go about setting up a viral-ish method for growth. None of these are things I came up with out of nowhere, they're all from having read various books / articles or from having worked at Infer. Feel free to ignore it / disagree, I'm just trying to be helpful because if you succeed, I'm more likely to succeed:

• I think you should look very closely at any instances you've seen of having an entire company adopt the product, or anything close to that, and try to figure out how you could make that an easily-repeatable process.

  • Analogy: I think the best method for creating a repeatable sales / growth process may be a lot like the best method for the creation of the product itself:

    • If you read stuff like the Lean Startup or read about successful companies like PayPal or Instagram or Twitter or Blogger, you see the same story over and over:

      1. the founders created a bunch of features and got some early users,

      2. they paid very close attention to what the users liked the most / what was working the best, and

      3. they then focused 100% on that feature / use-case.

    • Similarly, the way to create a repeatable sales process may be to try a bunch of different things and see which seems to work.

      • When PayPal was starting out, Peter Thiel:

        1. tried a bunch of different things

           • he authorized his marketing guy to spend $1 million "as quickly as possible"; the guy ended up spending it on ads that didn't have much of an effect

        2. one of the things they tried did work quite well: offering a $20 bonus for signing up yourself, and a $20 bonus for signing up someone else.

        3. So they then focused on that method.

  • Example: I suspect that one case you saw of the one guy who was able to sign up a bunch of his colleagues is a big opportunity. If I were you, I would spend a lot of time talking to that guy, thinking about that situation, and thinking about how you could make it happen more. If you can come up with a way to make that repeatable, you can raise a ton of money like PayPal did and just feed the raised money into this viral-growth-machine and end up with your service all over the country in no time at all.

• I highly recommend watching this talk by Parker Conrad, since you seem to be targeting roughly the same demographic as him:

  • Hypergrowth in the Enterprise | Parker Conrad - YouTube

• I would try to talk to salespeople for other businesses that sell to the companies you're trying to sell to and maybe even pay them as consultants to give you tips / pointers on the best way to break in.

  • For example, the sales rep for Rochester-Midland Corporation for the SF Bay area recommended that she stop by to talk to our HR person and the building management show off their bathroom-deodorizer product. This may not be the best strategy in your case, but if I were you I would want to know about that strategy and all the other strategies being used by companies selling to the same customers.

Here are some particular viral-ish growth strategies I'd be inclined to try:

• If you get someone at a new company to sign up and you aren't sure if they're influential or not, offer them $20 to find out who the most talkative / outgoing / influential person is at the company.

• Alternatively, try to figure out on your own who the most influential person at a company is.

  • Look at their job title

    • Sales guys - If this one guy who signed up everybody was a sales guy, that may indicate that sales guys at most companies are likely to be the most-talkative and outgoing.

      • At my company, the most-outgoing person I can think of is a Sales Engineer.

        • [Link to his LinkedIn profile redacted]

        • Just from looking at his profile briefly, I came up with some ideas (listed below).

    • More-senior guys

      • At my company, the people who have been around longer / have higher titles tend to be more influential.

  • Look at their online presence - You may be able to get a sense of who is most likely to be an influencer by looking at their LinkedIn / Facebook / Twitter / etc. profiles.

    • Scrape LinkedIn profiles and look for people who look the happiest / have the biggest smiles.

      • Microsoft Oxford has an API that you can use to gauge how happy the person is in their photo.

        • Just from briefly looking at some of the photos of the people at my company, I can say that it seems to correlate pretty closely: the people with bigger / more-genuine smiles are the ones I would call the more-influential employees among their peers. They're the ones who like to spend time with others / talk with others.

    • Look for people who have exclamation points in their profile.

      • [Example from my company redacted.]

    • Look for the most-attractive people.

      • Unfortunately there isn't an API for this yet (that I'm aware of), but I know that on Instagram the female models are heavily-pursued by advertisers. If you could find the most-attractive men / women at a company via their LinkedIn profile pic, you could have a similar set-up except you wouldn't be competing against any other advertisers.

    • Look for people with the most connections.

      • I have 255 connections on LinkedIn, [my outgoing coworker] has "500+".

  • Look at their car - You may be able to look at what car everyone drives in a parking lot and go after the people with the nicest cars, since [that may correlate with how willing they are to try the product]. [Edited to remove identifying information.]

    • You could leave a flattering, fancy-looking flier underneath their windshields offering them a special deal because they have such a nice car.

      • Flattery goes a long way.

      • The more you can hone in on the people you want to reach, the more money you can spend on the promotional material.

• I would give the influencers a message that flatters them and is honest about why you need their help in particular:

  •  "We have this great new product but it's a little unusual, and so people may be reluctant to try it. We're looking for the most influential people within each company to lead the way, and we're willing to compensate them for their time. X told us you were the one to talk to / We determined you were likely to be an influencer based on X."

• Offer free samples and / or a money-back guarantee: If you're really convinced that the product is so good that once people try it they won't want to stop, then it may be a good idea to take on the risk that the customer feels about trying something new by offering a free sample / money-back guarantee when you first try it.

  • Examples

    • Red Bull loves to give away free samples of its product because it knows the product is so good people will keep coming back.

    • 5-Hour Energy loves to give away free samples.

    • Cigarette companies love to give away free samples.

    • PythonAnywhere offers a money-back guarantee for the first 30 days after you sign up for a new plan.

  • On the other hand, the companies I list in the examples seem to have higher margins and cheaper products, so it may require some fiddling to get it to work for your business.

  • If you're not confident in the product enough to offer a money-back guarantee, then that may indicate a product issue that needs to get fixed.

• Offer bonuses for being the first (few) person(s) at your company to sign up.

  • Like I said [in conversation] before, it seems that signing up a new company is hardest at the beginning, and then gets easier as more and more people sign up.

  • Thus it seems you should front-load your bonuses to offer bigger rewards to early adopters and fewer / no rewards to the laggards.

• Offer bonuses for all employees of the first company in an office park to sign up.

  • This is the same idea as above, except inserting "company" where before there was "individual". I'd bet that as more and more companies in an office-building sign up, it'll get easier to sign up the other companies.

  • So you may want to front-load your bonuses towards the early-adopting companies, and offer fewer / no rewards to the laggard companies.

  • Examples:

    • When Mark Cuban was starting his first company, he did his first job for free in exchange for being able to use the company as a reference.

• Set up some kind of PayPal / affiliate-marketing-type program so that the people who end up being very good at signing up new customers can keep making money by doing it for other companies.

  • If this one sales guy was able to sign up his entire company, you may want to try to incentivize him to sign up other companies in the same area.

  • Examples:

    • PayPal offered a $20 bonus for signing up yourself and a $20 bonus for signing up a someone else. They had people who were making $10,000 by referring other people to sign up.

    • Amazon has had an affiliate marketing program since its early days in which bloggers can link to a product on Amazon's website and get paid for the referral. Bloggers love it and I'm sure it's helped Amazon.

I have made a bet that Trump will win the Republican nomination for President

Originally published 2015.12.24

Update 2017.08.15: I've added an image of the webpage where I placed the bet, which obviously doesn't perfectly reflect the way it looked when I wrote this post.

In 2012 I wrote a post in which I suggested betting against people who believed that the world was coming to an end. Unfortunately I didn't know how to find such people.

I recently discovered that you can currently go onto betting sites and get ~3:1 odds on a bet that Trump will be the Republican nominee (so if he's the nominee, you get $3; if he isn't, you lose $1). I bet some money on PredictIt.org on December 16 (about a week ago) at 3:1 odds that he'd do it, not because I'm totally convinced that he will, but because:

1. Regardless of who the candidate is, those odds seem suspicious to me for someone who appears to be so far ahead in the polls,

2. I don't find it that hard to believe that the party that had Sarah Palin on the ticket as the VP nominee might nominate Trump for President,

3. I find it harder to believe that Republican delegates would vote against Trump even if he won their state and was leading in the national polls,

4. The betting market for this election appears to be relatively small compared to the financial markets and so it seems there's more likely to be exploitable inefficiencies in the way people are making bets,

5. The media has been repeatedly wrong about predicting Trump's fall in the polls,

6. I've read that many successful investors / entrepreneurs got experience picking stocks / making bets when they were younger, and I'm trying to copy the habits of successful people, and

7. I have an interest in finance / trading / value investing but have avoided dabbling in the financial markets because there's so much competition that I doubt I would have a competitive advantage, and smaller betting markets seem like a way to get experience in a less-competitive environment.

So I'm treating this primarily as a learning experience.

How I could be wrong

1. A persuasive counter-argument I've heard is that it's still early enough that anything could happen. I agree with that. I just don't think that the odds should be 3:1 against someone with the kind of lead Trump appears to have.

2. Another point I've heard is that Trump's support is largely among people who have been less likely to vote in past elections, and that this low turnout is likely continue into this election.

3. Another point I've heard is that Trump's get-out-the-vote efforts are weaker than those of his opponents, and that this could lead to a difference between his polling numbers and the actual vote tallies.

4. I think a danger for Trump that the media doesn't seem to be talking about is character assassination; Trump seems to be filling the role of the "straight-talking, no-nonsense" candidate that Chris Christie was supposed to be playing in this election, until Chris Christie's reputation was severely damaged by the George Washington Bridge lane closure scandal ("Bridgegate"). It seems plausible that what happened to Christie could happen to Trump.

   • The problem with waiting for a Trump-character-assassination before the Republican nomination is that:

     1. Democrats don't seem to feel an incentive to use character assassination at the moment because they have reason to believe that they'll have an easier time winning the general election if Trump wins the Republican nomination.

     2. The other Republican candidates seem to not want to risk being associated with going after Trump because they're worried about ending up on the bad side of Trump's supporters.

   • I think the best way for other Republican candidates to go after Trump would be to paint him as a liar. They could find people from Trump's past to tell Trump's supporters that Trump has historically loved to make promises that he can't--and has no intention to--deliver. This way the other Republican candidates can avoid criticizing the controversial proposals that Trump's supporters seem to love. However, to execute this strategy would require the candidates to be very aggressive about researching Trump's past in a very short amount of time.

Ultimately I don't like this way of placing bets; I'd much rather be Ed Thorp with his mathematically-proven blackjack system than someone who bets off a hunch. But this exercise is turning out to be a very useful way to motivate myself to get more practice at evaluating bets; I'm already thinking about using web-scrapers and machine learning to get an edge that I could be more confident about.

I'm keeping more-detailed notes here.

Alcohol may reduce anxiety while programming

Originally published 2015.12.01

On Sunday I had my first-ever alcoholic drink on my own initiative, without being prompted by someone to drink (it was a Stella Artois in the fridge at work).

I decided to try it as an experiment to see if it would take away this anxious feeling I've been getting for years when I'm trying to program and I'm running into one error after another; the feeling ends up making me procrastinate really heavily, to the point where I'll go an entire day without getting anything done.

I got the idea from reading Mark Zuckerberg's blog posts from when he was creating Face Mash, in which he seems to describe drinking to take away the anxiety of not having your code working: "It’s taking a few tries to compile the script…another Beck’s is in order.".

Well, it seemed to work. I'm not sure if it was just the placebo effect or not. I didn't finish the beer. I'll probably keep experimenting with it. Obviously it's something to be careful with.

Update 2023: I've discovered that stimulants in general seem capable of dulling that feeling of anxiety that comes from having things not working.  I've recently been experimenting with caffeine (tea, coffee) and found it to have a similar effect.  It's not clear to me yet if it's something that subsides as you develop a tolerance for the stimulant.  Based on the number of successful indie hackers I see drinking coffee every day, my guess is that the effect stays strong enough to make it worth continuing to consume caffeine for that benefit, even after developing a tolerance for it.

How to Read Business Biographies

Originally published 2015.03.07

Here's an email I recently sent to someone who is interested in starting a business within the next few years; he had asked me if I had any advice from having read a bunch of business (auto)biographies:

I've attached a list of some of the books I've read since graduating college. Note: I haven't read all of those books cover-to-cover, but I only included books on that list where I read the "meat" of the book. I'll explain:

I recommend you think about reading books like eating steaks at a restaurant or in your home. Sometimes you'll get a steak served to you that has a little fat around the edges, and what I do is cut that part out and not eat it. Sometimes you'll have a lot of fat, and you only end up eating half of what was given to you by weight. Books work the same way.

I recommend that if you decide to read business biographies, you focus on the first few chapters, where it is explained how the person went from being like anybody else to being more successful than other people. If those chapters make up only 1/10 of the book, I think it's smarter to read those chapters 10 times than to read the entire book once.

Another analogy I use to explain how to read is Ikea furniture: when I'm assembling Ikea furniture, I don't read the instruction manual cover-to-cover before I start. I just read the first step and do that step, and then I read the next step. Business biographies work the same way: most biographies spend the bulk of their pages talking about the subject once he/she was already successful, but that isn't as useful to people like us because we haven't yet gotten to the same level of success.

So instead of reading these books cover-to-cover, you'd be better off focusing on the beginning, which is more relevant to your current situation. If you follow that strategy you'll also be able to get through more books, and you'll develop a better sense of what other successful people have done early in their careers.

Wailes, Nathan - Continued Study

Update: December 2014 & How I Got My New Job

Originally published 2014.12.17

I have moved from Washington, D.C. to Palo Alto, California for a new job. On December 8 I began working at a start-up called "Infer", which is using machine learning to help companies focus their salesmen on the potential customers who are most likely to buy. I got the job offer after I created a Python script that automatically applied to jobs on HackerNews' monthly "Who's Hiring?" thread.

One thing I found interesting was that I actually applied to Infer twice: once on August 1st, and again on September 10th. The first time I applied I never heard from them, but the second time I was offered an interview. I thought that was interesting because I had made several changes to my application materials between the first and second time I applied.

Below is the resume I sent them on August 1st:

Wailes, Nathan - Resume

And below are the attachments I sent them the second time:

Wailes, Nathan - Continued Study Wailes, Nathan - Portfolio Wailes, Nathan - References Wailes, Nathan - Resume

From this whole experience I came to a few conclusions that I could use in a future job search. I then tried out these ideas on Tinder and OkCupid and they really seem to work:

Don't spend much/any time learning about a company / woman until they express some interest in your candidacy. Learning about a company / person takes time. If you do that for every potential company / date, you are going to lose a lot of time to that process. Also, by doing it you risk developing feelings for them and getting hurt if they don't reciprocate. Not having imagined yourself being happy with them also helps you relax in the interview. It took me a long time to learn this lesson.

It's very useful to have a way to quickly send your application to a lot of potential employers / potential dates. This is helpful because 1. Your odds of getting any particular job / date are low (so if you're slow at applying, it'll take a lot longer to find a job), and 2. It's demotivating to spend a lot of time applying to jobs without hearing back (and so you risk giving up altogether). A general rule to follow is: Try to figure out how to limit the time you spend applying to a particular job / sending a message to a potential date to the amount of time that that person will spend on their initial evaluation of your candidacy. So if they're going to spend 10-20 seconds looking at your resume, try to figure out how to apply to each job in 10-20 seconds or less. My Python script for HackerNews takes less than 5 minutes to set up and it applies to ~100-170 jobs. There's a new HackerNews "Who's Hiring" thread every month, so I can repeat this every month. That's an extreme example. A first step is to stop customizing your application to the employer; I wasted a lot of time doing that when I first started applying to jobs straight out of college.

Even if the employer / female says they have specific requirements, apply anyway. Many will be interested in talking to a candidate who has an interesting application that is close to being what they want, even if that candidate wasn't what they initially had in mind. The job position at Infer that they had posted on HackerNews--the one I applied to--was one I was not at all qualified for. The smarter employers will be on the lookout for talented people; it's one of the big pieces of advice given in the well-known business book "Good to Great". Regarding dating, I have heard back from many women who said in their profile that they had particular requirements (which I didn't meet). But this bit of advice ("apply anyway") is only really useful after you figure out how to apply to a lot of jobs quickly, because you are probably going to have a lower chance of getting jobs you don't fully qualify for, and so if it takes you 30 minutes to apply to each job you'd be better off focusing on the ones you are best qualified for.

Try to make the potential employer's / date's  evaluation job easier, even if that means doing things they haven't explicitly requested. For example, most job postings on HackerNews will tell you something like, "Send your resume to bob@companyname.com". But that doesn't mean you have to only send your resume. Between August and September I changed my application to not only include my resume, but also a portfolio of projects, references, and books I had read. It made it easier for the potential employer to get to know me. I also made my initial email short (just 1 sentence), whereas many people will have long multi-paragraph "cover letter"-ish emails, which just make the potential employer's job harder when they have a hundred applications to consider. In my resume I underlined the most important points I wanted the potential employer to notice / remember, which makes it easier for people to skim. I have spent a lot of time working on my website / YouTube / LinkedIn / Facebook and have them all public, so potential employers / dates who want to learn about me have a much easier time doing so. On Tinder / OKCupid, I immediately cut to the chase and invite potential dates to dinner, saving us both time (the normal etiquette is to trade messages with the woman to help her decide if she wants to meet in person, but in my opinion that's not really a very helpful way for the woman to decide). To help them decide whether to accept my invitation to dinner, I give them my full name and invite them to Google me. I have had at least one young woman say I was the only person to ever do that, and she seemed very pleasantly surprised.

Invest time in making yourself a candidate potential employers / dates will be interested in. If you spend all of your free time watching TV shows and eating ice cream, you may have trouble competing against people who spend their free time working out, studying new job skills, finding nice clothing, or otherwise grooming themselves. I have met a lot of people and seen a lot of resumes from people who--even if they followed all of the advice above--would still have trouble finding a good job because they just don't have a lot to offer, and they don't have a lot to offer because of the way they are spending their free time.

Update: September 2014

Originally published 2014.09.16

Below is an email I sent to a friend of mine who is coaching me through the process of how to break into the programming world. He went to Princeton, worked at Google for a year or two, and has now been working at hedge funds for the past few years, so I trust his advice.

-------------

I made good progress this month: I made changes to my job application email (see below), my Python script sent out ~130 emails about a week ago to people posting in the HackerNews "Who's Hiring" thread, and my inbox is now full of ~30-40 responses. I've had ~5 phone interviews in the past few days, three or four of which were with people who went to Princeton / Harvard / Yale / MIT, and I have a bunch of people I need to get back to via email.

I think I've pretty much figured out how to get an interview. Now I need to practice getting past the interview. For the interviews I've had I can tell I have consistently underwhelmed my interviewer when they give me a coding challenge. For the next month I'm going to focus on HackerRank and Cracking the Coding Interview.

What I did this past month:

1. In August I only attached my resume to my emails. This time I attached my resume, portfolio, references, and a list of books I've read since I graduated college. (I've attached everything I used just in case you're curious.) I think this was a big part of why I got so many more responses this time, especially the portfolio.

1. I made the changes you recommended to my resume. When I first read what you wrote I was defensive, thinking, "I don't want to work with squares!", but after thinking about it longer I agree with you: a certain level of formality can inspire confidence in a potential-employer when you're first getting to know them; it helps show that you know how to get down to business.

1. I "rebranded" some of my previous jobs: for any job where I used programming to get my job done, I described myself as being a Developer of some sort, as opposed to being an Administrative Assistant who uses programming to get his job done faster.

1. I ran an A/B test on my resume: for half the emails I sent out I used a resume that said I had just graduated from DevBootcamp, and for the other half of the emails my resume said I had been accepted and was deciding whether to go. Surprisingly (to me), I actually got more responses to the resume that said I 

hadn't gone to DevBootcamp. All but one of my phone interviews were for applications that said I hadn't gone. Today I had an interview with an MIT PhD who said he found it very impressive that I was self-taught. So I think I'm going to pass on going to a bootcamp and just focus on being able to pass an interview (big-O notation, data structures, algorithms, coding challenges).

1. I added a second page to my portfolio with some more-minor projects.

1. Last month I applied the day that the HN thread was posted. This time I procrastinated and sent the emails out about a week later. Because of the large response I got, I suspect that might be a better way to do it in the future.

---------Unrelated to my job applications-------------

1. I started reading Calculus For Dummies, Linear Algebra For Dummies, Sams Teach Yourself SQL In 10 Minutes, and Beginning Programming With Java For Dummies. I'm trying to do a chapter a day for each book as much as possible. I can read the SQL / Java book at work without having to worry about getting in trouble, and I read the Calc/Lin. Alg. book at home in bed when I'm getting ready to go to sleep.

Update: July 2014

Originally published 2014.07.30

Since September 2013 I have been working at a DC startup called "Hitch", which coordinated the installation and management of credit card machines in the back of DC taxis. The installations were generally finished by December 2013, at which point the company shifted to troubleshooting problems with the systems (drivers not being paid, machines breaking down). The CEO is now hoping to use this position to create a taxi dispatch app that would compete with Uber.

While working there I have continued to study programming and business. Over the past few months I have been developing a computer program that automatically applies to jobs on Craigslist, and after getting it to work for myself I put in a few dozen more hours to get it working for other people as well. At the moment I have it running for 5 people, and I'm beginning to cold-call people to find more testers.

Update: July 2013

Originally published 2013.07.16

Here's a message I sent to a math/programming-genius friend of mine who's helping me stay motivated while I study math/programming:

I've been procrastinating about sending you an update, so this'll be a partial list of things I've been thinking about:

1. If you find YouTube ads annoying, Google "adblockplus". It installs in 1 click for Chrome and disables ALL ads. It's amazing, and I can't believe I've been dealing with ads for years when they're so easy to get rid of.

 

1. Check out "Dimmer" for dimming your screen, if that's annoying and you don't already have a solution.

1. Check out AutoHotkey for Windows (it lets you move the mouse around, click, press key combinations). It's great as a way of creating a macro that works across different programs, eg you want to copy something from Excel, switch to Chrome, and paste it in Google Search.

1. I started working on a Venn Diagram Teacher for my LSAT students but lost interest when it came time to make a menu:

http://www.nathanwailes.com/games/venn-diagram-teacher/

1. I found these cool explanations of how you can use linear algebra to analyze Chutes & Ladders and Candyland:

http://datagenetics.com/blog/november12011/
Candyland®

1. I've started a serious effort to try to be active in the AoPS community, use their free Alcumus service, and do a page from the AoPS Vol. 1 book every day. I've started with the topics I found most interesting (probability, counting, proofs), and I'm doing the practice problems at the end of each chapter every 2-3 days so that I get spaced-out practice.

6a. I helped Richard Rusczyk solve a problem he was having by using AutoHotkey, he responded "Wow; thanks. You really crushed that.":http://www.artofproblemsolving.com/Forum/viewtopic.php?f=133&t=542039

1. I spent a lot of time learning about AoPS, MATHCOUNTS, the AMCs, USAMO, and the IMO. I honestly had only the vaguest conception of what any of those things were when you would talk about them during high school. I watched the documentary "Hard Problems" on YouTube:

http://www.youtube.com/watch?v=B-nw75pNQgc

1. I've continued to make slow-and-steady progress studying for a perfect score on the SAT. I'm going through "Dr. John Chung's SAT Math", making flashcards of every problem that I i) get wrong, ii) find hard, or iii) solved in a different way than the solution given in the book. I can tell that I'm slowly getting better.

1. I highly, highly recommend that you check out Anki. It has really been amazing at helping me learn Spanish, Russian, German, American Sign Language, math, programming, people's names, the main ideas from books, etc. I know you said that you found spaced repetition to not be efficient without understanding WHY some mathematical thing was true (eg constantly rereading books), but I think that you can incorporate your idea of picking-stuff-apart into Anki. For example, you can add a card that says "This theorem says that these two things are necessary. Why is the first thing necessary? What happens if you don't have it?" and then on the back of the card it can be the lesson that you learned from working it out. Here's a useful reference of how to make effective cards:

Effective learning: Twenty rules of formulating knowledge - SuperMemo

10. I've been slowly making Anki flashcards for Teach Yourself Mathematical Groups, Introductory Graph Theory, and some other things. My goal is to make the cards work so well that a person can learn the material without reading the book. I'm starting to suspect that books are going to become outdated technology once people figure out how to use computers to help people learn faster. You can find my flashcards here: GitHub - NathanWailes/math-flashcards: Cards for various books / videos

11. The more math I do, the more I'm enjoying it for its own sake.

12. I highly, highly recommend that you learn about the mathematicians Ed Thorp and Jim Simons if you don't already know about them.
Jim Simons talking about his career:
https://www.youtube.com/watch?v=SVdTF4_QrTM

Very, very interesting articles/chapters (re: Jim Thorp, Michael Dell, etc.):
https://www.dropbox.com/sh/r7xu3ieionsv6iw/7ko6agatjG

13. After I took the LSAT I wrote up a lengthy guide on how I did it, which is now considered by many to be one of the best out there:http://www.top-law-schools.com/forums/viewtopic.php?f=6&t=120471
Would you consider writing up a short guide on how you got to where you are? I remember asking you about this in the past but I still only have a vague idea of what your path was (all I remember is you saying something like, "I was doing math on my own when I was like 4"). I'd be interested in things like, Did your parents read to you when you were growing up? Did you participate in MATHCOUNTS? What books did you use when you were in elementary / middle school? Were you in math clubs in elementary / middle school? What was the exact chronology of your learning to program? Was there a club at your elementary/middle school for programming? Did you start learning once you got to BCA, or did you know how to do it beforehand? Basically, I'm interested in having a very nitty-gritty understanding of how you got to be so good.

Monthly Update: January 2013

Originally published 2013.01.31

What I've been up to since my last post (June 2012):

• I started reading some books about programming / tech: Coders At Work, Founders At Work, the first few chapters of C++ Primer Plus (Prata), The Quants, started Dark Pools, most of The Facebook Effect, A Field Guide to Genetic Programming, most of Tesla's autobiography, most of Googled by Ken Auletta (the story of Google), Automate This, started The Singularity Is Near, started Rainbow's End, read most of On Intelligence, started some intro to Arduino books, tried to start "Concrete Mathematics: A Foundation for Computer Science" (it's very slow going).  I also did some USA Computing Olympiad (USACO) puzzles on their website.

• I started reading some books about mathematics: The Quants, Mathematics: A Very Short Introduction, Letters to a Young Mathematician, The Cartoon Guide to Statistics, started A Prime Obsession, Logicomix, started Everything And More by David Foster Wallace, a biography of Einstein, Innumeracy, started The Colossal Book of Mathematics, read about half of Introduction to Mathematical Thinking (Keith Devlin).  I also watched a lot of KhanAcademy videos on math.

• read some (auto)biographies: Money Mavericks (re: being a hedge fund manager), started Ben Graham's autobiography, read about half of Pour Your Heart Into It (re: Starbucks), started King of Capital (re: Steve Schwarzman), 'Surely You're Joking, Mr. Feynman!' (great book), started The House of Morgan, The Education of Millionaires (easy read w/ good info and good anecdotes), started the bio of Einstein by Isaacson (mentioned above), bio of Steve Jobs by Isaacson

• misc books: The Cartoon Introduction to Economics (Vols 1+2), started The Great Hangover (re: Vanity Fair articles about the '08 financial crisis)

My Portfolio: June 2012

Originally published 2012.06.01

Quick post:

What I'm thinking at the moment: For the past few months I'd been a bit suspicious of the rising market, and after thinking about it a while I started to suspect that it was a self-reinforcing rise that resulted at least partly from the collective pressure that lots of fund managers feel to get a certain amount of yield per year.  People were looking for good news because they needed something to go up to justify their (managers') existence.  When I left Treasuries back in February(?) I put my money in a Vanguard money market fund until I could figure out a better place for it.  I didn't get to participate in the rising market after then, but I also didn't take on as much risk of a catastrophic loss if some big scary event cause a panic or forced sell-off.  However, in '08 different MMAs saw losses, so even being in a Vanguard MMA isn't totally safe.

Re: the global outlook

What I'm thinking at the moment: I've been slowly making my way through Reinhart's "This Time Is Different" as well as a Peterson Institute report called "The Global Debt Outlook Over the Next 25 Years".  And I've also just been reading lots of miscellaneous articles.  The general feeling I'm getting is that there were a lot of bad loans made by banks and pensions in the past few years, and someone is going to have to eat those losses.  Because the problem is so big it seems like governments are going to try to spread the pain as much as possible, and they'll try to prevent another collapse like Lehman Brothers, so it doesn't look like the next big shock to the market will come from a bank's collapse.  Instead it looks like it will come when people try to flee the debt of particular countries.  That seems like it will be the point at which countries will no longer be able to "play pretend" about the reality of the situation.  I still don't know where money could go to escape; I think the US has the most capital denominated in its currency and so if countries need to inflate away debts then the US may be able to do it in the least disturbing way.

My Portfolio: April 2012

Originally published 2012.04.01

Quick post:

1. I pulled out of Treasuries about a week after my last post, when there was the first inkling of a sell-off.  I hadn't realized that if bonds are really popular now and become less popular in the future, you'll lose money if you're invested in a bond ETF.  I

think that's how that works.

1. I also learned that the borrowing cost of Groupon is so high (~100% of the stock's value per year) that you can't just hold it for 5 years and wait for the company to go bankrupt.

At the moment I'm just doing a lot of macro-level reading, trying to figure out what's going to happen in the market over the next few years.  I'm worried that there'll be another crash like in '08 because there's so much bad debt out there that could knock over a bank and cause another forced sell-off.  I want to learn more about exactly why the '08 crash happened.

My Portfolio: February 2012

Originally published 2012.02.14

I've started making trades in the Investopedia Stock Market Simulator; my user name is Nathan_Wailes, and I'm playing in the "2012 No End" game.  I started with $100,000.

I don't expect to be successful at first, but I figure I'll learn more by trying things and making mistakes than by only reading.

Actions taken:

• Bought ~$90,000 of Treasury ETFs (VGLT, VGSH, TLT) at the market

• Bought ~$5,000 of the biggest S&P500 ETF (SPY) at the market

• Sold Short ~$5,000 of Groupon (GRPN) at the market

I've moved my money into Treasury ETFs except for a bet that Groupon's value will diminish relative to the market.  Right now I have it as a spread trade: $5,000 shorting Groupon, and $5,000 long on an S&P500 ETF.  This basically allows my bet to control for the overall level of the market, so if people start putting a ton of money into the stock market and the general level of the market goes up, I should still be able to make money as long as Groupon's value relative to the rest of the market goes down.  The downside of this style of betting is that your reduced risk comes at the cost of reduced profits, since half of your betting money is being used to control for the general level of the market.

Reasoning for My Actions:

The Treasuries bet:

• There is a lot of uncertainty now about what is going to happen in the EU and the US because of this huge credit bubble stacked on top of an enormous debt problem.

• If growth for the next 10 years looks like it's going to be flat, stock PEs may gradually go down as the consensus on stocks changes (ie over 5-8 years of zero growth people will stop looking at stocks as being worth the PEs they're currently trading at).

• I want to avoid another crash like '08, and some intelligent investors have been saying that the danger to the banks that we're seeing now is just as much worth worrying about as the danger to banks in '08.

• Kyle Bass has said (and I think I've heard this elsewhere as well) that investors see Treasuries as being the "high ground" in this disaster: the asset least likely to take massive losses in value if everything goes haywire.

• Timeline for this bet: I'm going to keep the money in Treasuries until I find other uses for it.  So the money could be there for years if I don't find any other way to use it.  More likely is that it will be there a few months.

The Groupon bet:

• I've gotten the vibe that Groupon was a money-making scheme for its creator, and that's not a good omen for its future.

• The practices I've heard them engaging in don't sound like they can go on for years.

• At a $12 billion market cap, it's worth more than Office Depot, Safeway, Suzuki, Sunoco, and Quest Diagnostics (individually, not combined).  It seems very unlikely to me that Groupon is that valuable.

• Jim Rogers has said that he is short US technology stocks; given his opinion on the Facebook IPO (he recommends against buying Facebook shares), I'd be willing to be Groupon is one of the tech stocks he's shorting.

• Timeline for this bet: Based on the fate of many late '90s internet companies, I would guess that Groupon would lose value within the next 5 years.

Things that could make my decisions turn out to be bad ones:

The Treasuries bet:

• Interest rates go up - Treasuries could lose value if interest rates go up.  Interest rates will go up if the Fed raises them.  The Fed will raise interest rates to avoid high levels of inflation.  High levels of inflation may arise if the amount of money circulating in the economy starts to rise by a lot.  The amount of money circulating in the economy would rise if lots of everyday people went into more debt.

• The global outlook gets rosier - I'm not exactly how Treasury ETFs work, but it looks like if demand for equities goes up, and then demand for Treasuries goes down as a result, I could lose money.

Further Questions Regarding this Bet:

1. How much more money, exactly, would need to start circulating in the economy to cause X, Y, or Z levels of inflation?

2. How much inflation would need to exist to cause the Fed to raise interest rates?

3. How much money would need to flow out of Treasuries and into the stock market to cause an X% decrease in the value of my holdings?

The Groupon Bet:

• Groupon purchases or creates a new, exciting, and/or profitable business - Google has bought and created lots of businesses over the years, and my uneducated guess is that that type of growth has played a big role in its huge valuation.

• Keeping it as a spread trade isn't worth it - The reduced risk that I get from having it as a spread may not be worth the reduced return I will get from cutting my bet in half.

Investment Idea: Bet Against Apocalypse Believers

Originally published 2012.02.08

“What’s the idea?”

Find people who genuinely believe the world will end in 2012 and enter into contracts with them in which you give them the cash-equivalent of a fraction of the value of their property now (February 2012) and you get all of their property on January 1, 2013.

“A 2012 apocalypse? Huh?”

The idea of a 2012 apocalypse has been hyped up by the news for the past year or two.  I remember hearing about it being based on the Mayan calendar, but I also remember hearing about Christians talking about it, so honestly I don't know what the whole story is.  I'm honestly not even sure how many people genuinely believe in it and how much of it is media hype, but I do remember the hype around Y2K and I'm pretty sure a bunch of people thought that Y2K would be the end.

“Why would a hedge fund / individual want to do this?”

Well, it might be too-small and too-expensive an idea for a big hedge fund, but for a smaller group or individual investor it might be profitable.

To understand the idea you need to understand that one of the main ways that people make money in a capitalist system is by removing inefficiencies from the system.  So, for example, if the owner of a painting thinks that his painting is worth $100, and you know that that painting would be valued by someone else at $100,000, it is inefficient for the current owner to have the painting.  You can make money by removing this inefficiency, that is, by buying the painting from the first guy and selling it to the guy who values it more.

In the case of the 2012 apocalypse believers, it is inefficient (from their perspective) to have rights to things that they will not need; that is, it is inefficient for them to have the rights to their property from 2013 on forward.  And from your perspective, these apocalypse believers have a mistaken view of the world that is a source of inefficiency in the economy: if someone believes in Y2K and loads up on food and equipment to prepare for the end, and then the end doesn't come, that food may spoil and the equipment may never be used, and thus all the effort that went into creating those supplies has been wasted.

The real source of the profit here is the discount you might be able to get on the value of the property you'd be buying: since you'd be making an unusual offer, and since there probably aren't other buyers out there who are making offers like this, you might be able to get extremely favorable terms with all of these sellers.  Since they think the world is going to end before 2013 anyway, what difference does it make to them whether they get 100 cents on the dollar for the value of their home or only 50 cents?  Or only 25 cents?  Their time is limited, and it doesn't seem worth spending a significant portion of it trying to get the best possible price.

“How are you going to get these people to enter into the contracts?”

This is where some salesmanship may be required.  One thought that I've had is that you could find one believer who is willing to deal with you and use that person to convince the rest of them.  He could say something like the following: "Hey, I'm preparing for the 2012 apocalypse myself, and in order to spread the word to as many people as possible I've sold the rights to my property after 2012 to some foolish non-believers.  I'm then using that money to purchase advertisements on billboards and on Google to try to save as many people as possible before it's too late."

Investment Idea: CDS ETFs

Originally published 2012.01.31

"What should I have read to understand what you're about to talk about?"

As far as I can tell, this idea is mainly based on three things: The Big Short by Michael Lewis (from which I learned how profitable CDS can be), Kyle Bass's 2010/2011 AmeriCatalyst videos (from which I learned how profitable a bet against indebted countries could be), and my CFA Level 1 Alternative Investment textbook's explanation of what ETFs are (from which I learned how useful ETFs can be).  Google them to find them.

"What's the idea?"

Create an exchange-traded fund (ETF) that is completely invested in a certain credit-default swap (CDS).  Basically, it'll be a way to let the little guys out there with not that much money place bets on whether certain countries/companies will default on their debt.

"ETF? CDS? Huh?"

An exchange-traded fund (ETF) is a fund that is bought and sold on a stock market like any other company.  ETFs make it easier to quickly buy and sell small pieces of things that would otherwise be difficult to buy and sell.  For example, the most popular ETFs nowadays allow people to buy and sell packages of stocks; there are S&P 500 ETFs, tech stock ETFs, etc.

A credit-default-swap (CDS) is basically a life-insurance policy on bonds sold by a particular company or country; it protects people in case the country/company is unable to pay back the bond.  CDS became famous after a bunch of hedge fund managers (the first being Michael Burry) used them in 2008 to make a killing off the subprime mortgage collapse, and for some unknown reason banks are still selling CDS at very low prices.  Warren Buffett has said that he would charge much, much higher prices for CDS.

"Why would a hedge fund want to do this?"

It'll be a way for a hedge fund to profit from the large body of people out there who are nervous about the state of the world right now.  On 2012/01/30 the WSJ had an article that said you could pay $155,000 for a CDS on $10 million worth of Japanese government debt.  If I understand this stuff correctly, that's like a bet where you lose $1.50 if you're wrong and gain up to $100 if you're right; or to use a more pertinent example (based on the dollar amounts individual investors typically trade), a bet where you lose $150 if you're wrong and make up to $10,000 if you're right.  I would bet that a LOT of small investors would want to get in on that bet if they had the chance, but the problem is that CDS are typically only negotiated between the big guys in the market (big banks and hedge funds with hundreds of millions of dollars under management).  So if you could buy CDS from a big bank at a low price and sell shares of that bet to smaller investors at a higher price, you could make a profit regardless of how the bet eventually turned out.

"Why are you so confident people would want shares of the ETF?"

Well...1) The Big Short was a very, very popular book, and CDS was the centerpiece of the entire thing.  2) In just 2-3 months after Kyle Bass recommended the stock of a particular mortgage insurance company (MTG) at AmeriCatalyst 2011, the value of the stock almost doubled, which suggests that a lot of people are paying attention to Bass's ideas.  If word spread that people could get in on a Kyle-Bass-style CDS bet against Japan, I think it would have the potential to go viral.  The sheer novelty of the idea could make headlines.

"What's the downside risk here?"

Well...one risk is that no one will buy shares in the ETF.  If that happened you'd be stuck with the CDS unless you could re-sell it.  But the CDS doesn't seem to be such a bad bet; I spoke to one guy who works at a fund-of-funds who said that his fund would have invested with Kyle Bass if they knew that his Japan bet would pay off this year (his Japan bet is to buy CDS on Japanese gov debt).  Their (fund-of-funds') problem was that they couldn't afford to have their money sitting around for 3 years waiting for a big payoff, because the people who have invested with them follow the fund's yearly returns and will pull their (investors') money out if they (the investors) don't see results every single year.  Re-selling the CDS via an ETF could solve their problem because they could make an immediate profit from arbitraging* the difference in price between what the banks are willing to sell at and what individual investors are willing to buy at.

"Would this even be allowed by the SEC?"

I'm honestly not sure.  My guess is that the big hurdle here would be the fact that only accredited investors are allowed to invest in hedge funds, and CDS is something that is typically done only by hedge funds.  On the other hand, though, since little guys are allowed to invest in mutual funds / pensions that invest in hedge funds, and little guys are allowed to invest in ETFs, it seems plausible to me that the ETF would be allowed if structured the right way.  And from the SEC's perspective, this could actually be a better way of pricing CDS: one of the touted benefits of a futures exchange is its ability to help find more efficient prices for goods, and so if there are tons of little guys out there bidding up the price of CDS, regulators might not have to worry as much about making sure banks are shooting themselves in the foot by selling the stuff too cheap.  That's a thought, anyway; I'm new to this stuff, so I could be way off here.

*when I say "arbitrage" I'm using it in the colloquial sense, not the strict sense of "riskless profit".