Sensemaking

Program above and beyond your actual ability by using FreeMind

2012-09-27T09:25:08-04:00

Technology is complicated.

These days building even the simplest website requires knowing at least half a dozen separate technologies, almost double if you want your site to be scalable and modular. Learning to program has always been challenging for beginners, but these days there are so many disparate pieces that even the best programmers are regularly called upon to write code in languages that they don't fully understand.

The question, then, is how can we possibly produce high quality code?

That is, given that it's no longer feasible to be an expert in all the tools we're required to use on a daily basis, how can we possibly produce code that's not just good enough some of the time, but that's consistently well commented, easy to read, efficient, and reusable?

The most obvious solution is simply by programming less. After all, the most readable line of code is always the one that isn't there.¹ How can we do this? One strategy is just to use the built-in tools of the language whenever possible rather than writing our own code that duplicates functionality.

The problem is that when you're not an expert at a given language then it's not obvious what functionality is already built in. And for most relatively minor problems, it's usually faster to just write a new function rather than trying to find out if a better solution already exists.

Why is this?

First you need to do a Google search to find the right page in the documentation. That 'page' might actually be dozens or hundreds of pages long. Then you need to control-f to try to find a tool that does roughly what you're looking for. But often what you're looking for doesn't come up, or else what comes up is only one of the many functions that could potentially do what you want, and the first one you find isn't actually best option.

When the quick search fails you end up paging down through the list of functions, scanning over the names and maybe the first couple sentences of the descriptions, trying to find a good fit. But this is extremely slow and tedious, so most likely you'll just end up going with the first thing that's close enough. This is bad.

But the worst part is that you probably won't even remember the exact method and how it works the next time you need it, meaning that especially for somewhat obscure methods you'll need to repeat this whole process again and again, getting only marginally faster each time.

We can do better. Not just a little better, but vastly better.

How?

In one of his essays Paul Graham speculates that perhaps the best hackers are the ones with the largest working memory, "so that when they look at a large program they don't just see that line but the whole program around it." However, he goes on to suggest that fortunately this isn't entirely necessary because you can cheat by using a really dense programming language. ²

This is exactly the right idea, but we need to take it a step further. In the same way that one can become a better programmer just by using a denser language, one can similarly become a better programmer by using denser documentation.

So how do we do this?

By using a specific piece of mind mapping software called FreeMind. While there are many software programs for doing mind mapping, FreeMind is unique because it has what its authors refer to as "a strong emphasis on folding." This is a distinction that's critically important. However, rather than formally defining mind mapping and folding, I'll explain by way of example.

Here is a link to the (partial) documentation for my own stack: python.mm. You can get it by right clicking and saving as a .mm file. Download it now. The way you view it is by installing FreeMind.

Now let's say you want to quickly look up how to check whether a dictionary has a certain entry in Python. Just use your arrow keys to scroll through the different nodes, and use the spacebar to expand a given node: Python > Dictionaries > Check if a dictionary has a key > key in d. Here is what you should see:

(Click to embiggen)

Instead of having to do a Google search and then scrolling through tons of documentation, or else trying to guess the right method based on the autocomplete suggestions in your IDE, you can actually look up the real answer in less than 10 seconds. If this doesn't impress you, it should. But what's even more impressive is that faster information lookup is just a small fraction of the total benefits we get from working this way.

The big picture is this: using FreeMind lets us densify the documentation for our stack by a factor of at least 1,000. Why is this important? Because it allows a programmer with 1x ability to program with 3 - 5x efficiency, and if you're already a 10x developer it will basically turn you into a god.

I know this is a big claim, so let's go through all of the things that FreeMind is going to let us do point by point:

See both the forest and the trees. In the example above, you can see the syntax for how to check whether or not a dictionary has a given key, but at the same time you can see all the other dictionary functions. And not only that, but you can also see a list of all the other (most important) datatypes that Python supports.

Every time you refer to the mind map you get a visual reminder of how the specific tools you're using relate to the larger toolset. For example, every time you look up how to use a RegEx Match Object you're reminded about the relationship between Match Objects and Regular Expression Objects. Especially as a beginner, but probably at every level, having these visual reminders makes it easier to write not only good code, but also comments that are both articulate and thoughtful.

This is huge. Why? Because both creativity and good decision making come from being able to combine pre-existing ideas in intelligent new ways. This means that being able to visualize both the details as well as the larger context and the alternative options at the same time is essentially an IQ supercharger. It's not just a matter of being able to look up the documentation slightly faster; once you get used to working this way you'll come up with vastly better ideas and solutions than you ever would have been able to before. This is something you simply need to experience to believe. I know it's considered poor form to start flame wars online. But based on extensive personal experience, I can't help but thinking that if you're not working this way then you might as well be eating paint chips, at least as far as your effective IQ is concerned.
Group tools by functionality. FreeMind lets you group similar tools together based on what they do, rather then having to read through the entire list alphabetically each time. In Python there are four different ways to add values to a dictionary, and they're all conveniently grouped in the 'add values to dictionary' section. Rather than paging through dozens of pages of documentation and going with the first semi-plausible method you find good and then continuing to do it the wrong way for the next five years, you can actually choose the best tool for the job each and every time.

Life is about using the whole box of crayons, and organizing your documentation this way is a big step toward writing code that's concise, readable, and efficient.
Never make the same mistake twice. Every time I find a runtime error due to using some tool with the wrong syntax I immediately correct the mistake in the REPL, and then add that REPL example (as well as the initial error) to the MindMap. This way either I won't make the same mistakes again, or if I do then they'll be much easier to find and fix.

Similarly, when I'm using a new built-in method for the first time it often takes a few minutes playing around in the REPL to learn how that method actually works and how to use it correctly. This is fine, but it isn't a process you should ever have to repeat. Being able to cut-and-paste a couple of canonical REPL examples directly into your map is a godsend. Not only should you add one or two examples that show how the tool is used correctly, but often it's useful to add an example or two of ways you'd think it might work based on the documentation but that actually result in errors.

This is huge the next time you need to go back and reference a section. Reading one or two really good REPL examples is often a much faster and more effective way of understanding the tool than reading the official documentation. And even if those REPL examples aren't enough to convey the full functionality of the tool in its entirety, being able to glance at them first generally gives context to the documentation and makes it easier and faster to read. Rather than having to carefully read several paragraphs multiple times to start to understand what the method does and how it's supposed to work, you actually have full comprehension of the documentation the first time, and much faster to boot.

In construction there is a saying that you should measure twice and cut once. When coding it should be just the opposite: REPL once, cut twice.
Cast a wide net. In addition to storing snippets from the official documentation along with REPL examples, you can also store Stack Overflow posts, blog entries, quora threads, A List Apart articles, HN comments, etc. The reason that I'm only sharing a relatively small portion of my own personal map is that frankly it has more copyright infringement than a public performance of a pirated Girl Talk album. This is a good thing.

Rarely is the best explanation for complex functionality found in the official documentation. For example, in Python often the explanations for new features in the Python Enhancement Proposals (PEPs) are more clear and readable than the official documentation once they actually get added to the language. Between the tutorials, the documentation, and the PEPs, rarely is the best explanation for a given feature found in just one of these sources. Often each one has insights the others lack, so generally the best approach is to cut-and-paste a few sentences or paragraphs from each into your map.

Often I also like retyping sections from books into my maps. Not only is this much more efficient than looking through the book each time want to look something up, but the act of retyping the most important sections forces you to really focus on what you're reading and helps you to better understand and retain the information. Generally when I'm reading I'll pencil in a dash next to the most important ideas and information, and then at the end of the chapter I'll add these sections to the map. I do this for both technical and non-technical books, and I find that it makes a world of difference.
Take complete ownership of the text. Let's say you really like a Stack Overflow explanation, but one of the examples isn't quite as clear as it could be. No problem. Just edit the text to make to say whatever it should have said originally. You own the text now, so you get to do whatever you want with it without permission or attribution. (Just don't redistribute this publicly.)
Easy memorization. Because each node in FreeMind is folding, your map essentially doubles as a set of flashcards that can help you memorize how to use the core features you find yourself frequently using. In fact, every time you look something up you're essentially drilling yourself with flashcards whether you realize it or not.
Never get stale. The problem with programming is that if you take a few weeks off to work in another language or just do something else entirely, when you come back it takes several days to get back into the flow. This is especially frustrating for tools you use only intermittently. How often do you spend hours teaching yourself something, and then the next time you need it the only thing you can remember is that you used to know this, and now you're going to have to spend hours learning it all again? The essence of FreeMind is that it lets you permanently store the entire contents of your mind in RAM, rather than having to offload the vast majority to the hard disk or trash collection. This lets you get back up to speed on any aspect of your stack in just a few minutes or hours, rather than the few days that it would normally take.

The big picture is that to program well you need both information and context. And not just from one source, but from the best of every source available.

One of the biggest reasons why beginners write bad code isn't because they're lacking the ability to do better, but rather because it takes so much time just to figure out what they need to know to get the program running correctly that there isn't any time for them to think about the larger issues: using the best tool for the job, writing code that's concise, readable, and reusable, thinking about how different sections of code fit together, taking the time to periodically refactor, etc.

FreeMind makes it infinitely faster to access this information. And not only are you accessing it faster, but you're getting information that's better because it's synthesized from a variety of sources. And not only this, but it's an entirely qualitatively different experience because you're essentially able to see both the forest and the trees at the same time. This helps you write lines of code that are not only better on their own, but that are also better in the context of the larger program.

To use a gaming analogy, it's like getting the benefits of sniper-level accuracy even if you only have shotgun-level aiming ability.

So, how can one implement this in real life?

When I'm teaching myself a new piece of technology I generally like to start by reading a some combination of books and tutorials, or else taking a class or watching a lecture. The first thing I do is to copy down everything I learn into the mind map. The material from these sources tends to be kind of scattershot, so much of it won't survive once I read the actual documentation. However, it's still useful to get an idea of what the most important features of a language are so that you know where you should be focusing your initial efforts. At the same time, there are often useful warnings about what parts of the language not to use because they're poorly thought out or confusing. I tend to avoid copying these parts into the final mind map so that I won't be tempted to use them later, or else I make sure that they are heavily caveated.

The next step is going through the official documentation and copying the useful bits. It's at this stage that I like to try to get things working in the REPL, and then add in those examples as well. If I still don't understand something then I look it up on Stack Overflow or the mailing lists and then also assimilate the information from there.

Unless I'm working with a relatively simple piece of technology (e.g. Less) I generally don't add everything to the map, just the most important functionality. I then continue to add more and more as I use the language, encounter new blog posts, etc.

The ideal is for every individual to do this for themselves. Feel free to use whatever parts of my map are useful for you, but don't blindly copy. Why? For two reasons:

You won't understand what's in there or how it's organized. This will eliminate much of the usefulness.
What should or shouldn't be included and how it should be organized varies depending on your ability as programmer, your knowledge of each specific programming language, how your mind works, etc.

In addition, to all of the benefits for programmers, there are a number of enormously useful non-technical uses for FreeMind as well. Here is a video I made explaining how I use it to organize the most important insights and data from all of the book, articles, academic studies, and government reports that I read:

This is usually where I find the citations for my Hacker News comments. By organizing data this way I'm able to easily find any given fact from anything I've ever read, all cross-referenced and vetted by primary sources. And even though this document would be hundreds of pages long if fully expanded, I'm able to do all this virtually instantaneously.

Pretty neat, huh?

I'll be the first to admit that using these techniques won't make you a 10x programmer overnight. But what they can do is help you to efficiently build up a solid foundation of programming and CS knowledge. And they can help you organize what you learn so that you can actually use it, commit it to memory, and add new knowledge as needed. If that isn't the way to becoming a 10x, I don't know what is.

And if you still can't see the wisdom of this approach, all I can say is just download a copy of FreeMind anyway. Try it out for a few years, you might find that it grows on you.

Notes

[1] I'm talking here about eliminating code by not repeating functionality and by not adding functionality you don't need in the first place. The other way to eliminate code is by condensing the rest of your program until it reads like aramaic, which is obviously a bad idea.

[2] A more powerful language reduces your need for working memory because the more you can see on the screen, the less you need to remember.

How To Reach Top Bloggers

2011-05-31T11:15:01-04:00

Other YC startups keep asking us for advice on reaching bloggers, so I figured I'd write down what we do. This isn't the definitive answer by any means, but it's worked reasonably well for us. What follows is basically our step-by-step method:

0) Use Google alerts and twitter search to find the bloggers already talking about you. The ROI from increasing the engagement of a user who already cares is 10x greater than convincing a new person to care.

Find bloggers and tweeters already talking about the problem you're trying to solve. For example, I know a doctor in Brooklyn who has his alerts set up to let him know whenever any twitter user within 15 blocks of his office has a cold. He then tells them he hopes they feel better, but if not they should come down to his office and he'll take a look at them for free. This is the only lead gen he needs. (Best tools for this: HootSuite and InboxQ.)

1) Figure out which bloggers you want to reach. To do this you first need to decide what you want out of this— beta testers, page rank, page views, sales leads, etc.

If you are solving an esoteric problem then there are only so many page views you're going to get from the blogosphere, so looking for mainstream adopters isn't always the right answer. In fact I'd generally recommend against it until you know that your engine of virality is yielding good results.

2) Meet users in your target group. There simply aren't enough influential bloggers than you can get away with sending out emails that don't convert. And the best way to figure out if you're in the right ballpark is to take a bunch of your target bloggers out for dinner and ask them for feedback on your proposed email, website copy, website design, features, whether they'd use the product, whether they'd blog about it, etc.

3) Find the relevant bloggers you want to target. The tools meant for PR agencies and larger companies include Vocus, Cision, eCairn, and PR Matchpoint. These products vary wildly in both cost and quality, and the efficacy of any given one depends largely on your particular needs. In many cases these tools are not only overpriced, but actually work less well than just doing things by hand. That said finding the right bloggers is extremely time intensive, and it's always a good idea to trade money for time where possible.

The tools meant for end users include Klout, Twitaholic, Technorati, PostRank, PeerIndex, Alltop, SocMetrics, FoodBuzz, BlogDash, and Loadedweb. Of these SocMetrics seems like the best bet for most purposes, albeit I haven't yet had time to fully explore all of them thoroughly. Each of the major blogging platforms also has a list of the best bloggers on that platform, e.g. WordPress, Typepad, Tumblr, Blogger, etc.

Expect that it's going to take you 15-20 minutes per blogger to build a well-targeted contact list. This includes the three bloggers that aren't a good fit for every one you find that is, and also the time it takes to get their email address, read their pitch guidelines if they have them, figure out their posting frequency, monthly page views, RSS subscribers, page rank, etc.

It may also be helpful to think about the types of events that your target bloggers would likely attend. For example, if you're trying to get mommy bloggers to write about your fitness product then try looking through the attendees list for events like BlogHer, FitnessBloggersConference, HealthyLivingSummit, etc. (Or else recreate the attendees list by searching for everyone who blogged or tweeted about the event.) Right now Lanyrd has a pretty good list of blogging-related events. Don't forget Meetup.

4) Once you have a decent list of bloggers, try to quickly figure out what motivates each of them to blog. E.g. MIke Arrington likes breaking new stories, Seth Godin likes prodding people toward doing great work, Scoble likes to be involved with everything that's going on, etc.

For a lot of them you probably won't be able to get this specific, so just try to mentally place them into one or two of the following categories: physical, mental, spiritual, recreational, family, career, social, and financial. Every goal falls into one of these eight buckets, and taking thirty seconds to scan over their recent posts and come up with a decent approximation can greatly increase your yield.

I'll explain more about this below, but for now here's a funny clip of one mom blogger explaining what she hopes to get out of Swagapalooza. Most people know that you need to sell on benefits and not features, but often forget that you can't sell on benefits without understanding the person's needs and desires.

5) Be authentic. What you say should be a reflection of who you are. If writing is like talking on paper, then a good email is like looking someone in the eyes on paper. People naturally want to help others, but they also tend to be apathetic and jaded, so engage with them at a human level. (See also: http://bit.ly/3D3LWN)

A personal observation: the best way to make mommy bloggers feel special and appreciated is to treat them like tech bloggers, and the best way to make tech bloggers feel special and appreciated is to treat them like mommy bloggers. (Kind of like how the reason the movie Field of Dreams was so successful is that it's basically a chick flick for guys.)

6) Have something of value the bloggers can offer their readers. Some examples:

Offer to do an interview or a video interview.
Have a manifesto. The goal of a manifesto is to first sell the readers on the underlying ideology behind your product, and then to show them how your product fits into this new worldview in a way that can help them. Examples of startups that played heavily on this are Facebook, Google, and Wikipedia.
Write an eBook -- This is a more detailed how-to guide that's targeted at the people already interested in your service. There are a lot of people who are super competent and willing to work, but they don't like thinking creatively. So give them a bunch of pre-packaged ideas about how to use your site better. Startups like Heroku and AirBnB would benefit from this. For an example, see http://bit.ly/hFohV

While you should aim to give the bloggers as much social capital as possible, avoid giving them too much of actual monetary value. If you start giving them things that cost lots of actual money then you stop being the underdog that they are rooting for and trying to help out, and they quickly start becoming entitled and apathetic. Obviously you don't want to be the little guy forever, but it's going to work against you if you start mouthing off and trying to throw your weight around *until* you're ready to absolutely crush it.

7) Convince them that you're not going to make them look like an asshole if they blog about you. The way you make them look like an asshole is if they go to bat for you, and then you don't ship. So you need to make it look like you're going to ship BIG TIME, and your success is inevitable. The way you do this is by having a track record of shipping great things, and having lots of social proof. The latter is why having a few famous advisors is critical, especially when you're just starting out. Otherwise it's easy to get caught in a trap where you can't ship without certain resources, but you don't have access to those resources because you don't have a track record of shipping. Good blog posts eventually become social proof in and of themselves, and have the magical effect of not only generating leads, but also of bringing all sorts of people out of the woodwork to help you with whatever you need.

The best way to lower this requisite social proof quota is to meet with the bloggers in person. Just spending half an hour grabbing coffee will make any blogger vastly more likely to write about you than if you had only connected via email or phone.

8) The more blog posts you have about you, the more each blog post is worth. This is because blog readers need to hear about your company several times before they buy the product. So the first ten blog posts you get are worth almost nothing, but having 100 blog posts about you could easily be worth a million dollars. It's what technical founders would call one of the n(n-1)/2 deals.

The same is also true for using blog posts as social proof for investors and potential hires. Any half-assed startup can get ten blog posts written about them, but if someone sees 100 blog posts about you they instantly assume you must actually have something of value.

9) Below I'm going to post a copy of the generic version of our invitation email to Swagapalooza. After the email I'm going to analyze the salient features of what makes this work:

[fname],

I'm Alex Krupp, director of Swagapalooza (www.swagapalooza.com), the first invitation-only event for only the most-followed bloggers and twitter users from across the country. 85 of the most-followed bloggers and tweeters from Chicago are coming to learn about new and interesting products. I'd love to have you attend the event, which takes place the evening of June 20th.

If you're able to come please fill out this form below so that I have you in the system:

https://swagapalooza.wufoo.com/forms/swagapalooza-rsvp/

If you have any questions feel free to call; I hope to see you there!

--
Alex Krupp
Director
Swagapalooza
Cell: (607) 351 2671
Web: www.swagapalooza.com
Twitter: @swagapalooza

=====================================

About 40% of the bloggers we send this email to end up RSVPing, and of those about 60% end up showing up. (Normally for events getting even 2% of your email list to show up would be considered quite good.)

So while this email is by no means perfect, it's still been good enough to get people to drive 3 hours from Boston to New York, fly back from vacation in Singapore a day early, fly up from North Carolina, etc.

What makes this email work?

- Five sentences or less. No first email to someone should be longer than five sentences for any reason, ever.

- Appeals to both social exclusivity and novelty seeking behavior.

- Looks like a personal email and was sent from my personal GMail account on a Sunday afternoon, i.e. sent in a way that's consistent with how personal emails are sent. And really that's because they are; I do my best to answer all incoming questions within five or ten minutes, as per http://bit.ly/azdIrn.

- Has my cell phone number. This is the most important part because it shows that this email isn't spam and that you value the other person and their time. (Counterintuitively, adding your phone number to emails is the best way to ensure that people read the FAQ before wasting your time with dumb questions.)

- Email signature = looks like a legit company that's actually going to deliver on their promises and not just some kid in their basement. This is also why some people advocate that every startup should always say they're hiring even if they're not, and it's why sites with 800 numbers convert better even when almost no one calls them.

- Linked URL in parenthesis makes you look like a trustworthy PR manager, not like one of those unsavory IT folks.

- For many bloggers we customized the first sentence or second sentence of this email to match what we perceived as being their motivations for blogging, as described above. E.g. if we were writing Paul Graham we would probably change it to something like,

"I'm Alex Krupp, director of Swagapalooza (www.swagapalooza.com), the first invitation-only event for only the most-followed bloggers and twitter users from across the country. Many highly followed bloggers from NYC are coming to learn about new and interesting products, and I think you would find the event to be intellectually interesting."

10) Obviously big events don't make sense for every startup, but at the very least try hosting a dinner with five or six local bloggers to get feedback on your website and the emails you're going to send out. It doesn't make sense to start burning your leads until you know that your conversion rate is going to be sufficiently high, and the only way to really know how if your emails are good enough and your site is ready is to sit down with a group of bloggers and let them bounce ideas off each other.

11) Make sure you have a professionally typeset one-pager (media sheet) ready to go as a PDF. The bloggers won't need this, but once you start getting blog coverage you'll start getting leads, partnership offers, media enquiries, etc. so you will want something you can send out at a moment's notice. Also make sure you have professionally shot high-resolution photos of the product and team. These will be the first thing that every trade journal, newsletter, and magazine asks for, and many won't even talk to you without them.

I'd also recommend having a Squidoo page about the product / company. In addition to being good SEO, they allow you to speak in a different voice to a different group of people without seeming inauthentic. They also resonate well with a wide variety of people because they are designed to incorporate mixed media. Some people are best sold by paragraphs of text, others by bullet points, photographs, YouTube videos, audio interviews, etc. We've discovered that they're not good as the first thing to send someone, but they work great for closing the sale after you've already emailed the person and talked with them on the phone. (Basically they are good for showcasing social proof, which makes it easier for someone to pull the trigger and write you a check.) Check out what we did with ours here: http://www.squidoo.com/swagapalooza.

12) Once you start getting coverage, go back and review step zero.

13) Hopefully this email lays out a viable framework for getting some blog coverage without having to resort to making a 14-year-old kid your CEO or acting like Dennis Rodman.

Basically just create a well-targeted list, treat them with respect, and be conscious of the social idioms you're drawing from so you're not just accidentally subcommunicating stuff at random. Then run your stuff by a few bloggers first in person and you should be good to go!

How I got into YC as a non-technical single founder

2011-04-25T14:35:34-04:00

When bloggers apply to attend Swagapalooza, we ask them one question:

"What makes you a top digital influencer?"

I've read through thousands of applications, but at the end of the day there's really only one good answer: Google me.

What follows is an explanation of how I got into Y Combinator by adopting this posture toward life, and how I've earned the dubious honor of being the only non-technical single founder that Y Combinator has ever funded. As a free bonus, I'll throw in a couple YC interview tips that I haven't seen posted anywhere else on the web.

tl;dr What makes me so unique and special? Nothing. I just shipped.

I started this blog the day after I got rejected from Y Combinator.

Why? Because of a story my college rowing coach told me about his own experience as an undergrad. Even though he had been one of the fastest rowers the fall of his freshman year, he wasn't put in the boat for the big race of the season. Apparently they wrote everyone's names on popsicle sticks and moved them around as the way of setting lineups, and when his popsicle stick fell under the table no one had noticed. By the time anyone found out it was too late, and he simply didn't get to race for that season.

Then and there he made a vow that no one would ever lose his popsicle stick again, and he doubled down on his training until he practically was the boat. This was my inspiration. I would start the greatest blog ever, and Paul Graham would never lose my popsicle stick again.

So I started my blog, and within mere weeks my writing was regularly being featured on the front page of HN, Reddit, Delicious, StumbleUpon, etc. I even made the front page of Digg, which even then was virtually impossible.

When I reapplied the first time, I did so with someone who was widely recognized as one of the smartest technical folks on HN. We made the interview round, but then got rejected again.

PGWTFBBQ??

So I went back to school, finished my degree, and largely forgot about YC proper. After college I participated in Seth Godin's alternative MBA, an experience that practically deserves its own book. As part of this program we each had to start our own business, which is where Swagapalooza came from. This was never meant to be a real business, it was basically something I started because the domain name was available and I thought it would be funny.

The first event was in September 2009, and essentially everything that could have possibly gone wrong did. We even forgot a name tag for Peter Shankman, our keynote speaker.

Imagine my surprise when then next day I discovered a couple dozen blog posts and articles about how much people loved the event. So I quickly came up with a plan to scale this into a real business, and reapplied one last time.

Now you'd probably expect that if Y Combinator accepted me as a non-technical single founder, it must be because I have some sort of extraordinary sales ability, and that I completely killed it during the interview. Nothing could be further from the truth.

Halfway through the interview I discovered that no one actually understood what I was trying to make. I started getting more and more nervous, until eventually I basically just lost my ability to talk and was curled up in a little ball in the interview chair. The ten minutes finished and everyone in the room still had very little understanding of what I was working on, let alone why it was going to be big. They told me the interview was over and I asked if I could play a short video clip of a blogger talking about how much she loved the event. They said no. I turned up the volume and started playing it anyway as I slowly backed away. Jessica looked disappointed.

Later that I night I got a phone call from pg saying, "You probably weren't expecting this call, but..." and the rest is history.

What happened?

The coach of the Yale rowing team has a saying about high school recruiting: "7:20 2,000m time, 720 SATs. 6:20 2,000m time, 620 SATs."

So far as I can tell, that's essentially what happened here. If you have a track record of shipping, nothing else matters. You can (apparently) flub the interview, have all sorts of stuff go wrong, and still be just fine. If I hadn't had 25+ blog posts and testimonials from people talking about how much they loved the event then I would have been dead in the water, even if I had the best cofounder in the world.

That said, as poorly as my interview went, there are still a few tips I want to share:

When you get into the interview room, start by explaining what it is you do even if they don't ask you. The YC partners interview so many people that by the time they get to you, everything is a blur. However, perhaps because it's somewhat socially awkward, they're not always the best at admitting this. Even if one of them starts by asking a specific question, it's best to start by explaining the whole idea from scratch. Just pretend it's like The Bourne Identity, and assume that they have zero memory of everything on the application.
Have a mindmap with the answers to all the questions they might ask. Having a word document with notes is no good, because as soon as you start scrolling through they'll get impatient and just ask another question. Whereas with a good mindmap you can answer every conceivable question in real time, even if you're drawing a complete blank. Here is an example of the mindmap I made for my interview, with the caveat that I've removed most of the answers: alexkrupp.com/yc.html. It's obviously much better if you're able to completely nail every answer without having to glance down at your computer, but it's also really easy to forget two of the four ways you're going to get new users when you're under pressure.
Basically answer these three questions, regardless of whether or not you get asked them: What does your startup do? How does it scale? How do you get new users? The second question is especially important, because it's what they're looking for in every startup they fund. By the end of the ten minutes they had no idea what I was making, but they sure as hell knew how it was going to scale.

In one of his essays, Paul Graham says that the secret of building a successful startup is getting someone like Einstein and having them design refrigerators. My strategy going into the interview was to convince YC that, despite the fact that I might be completely unqualified to start most of the businesses they work with, I was in fact massively overqualified to start and scale this particular business.

So, how is that business going today? Here are two blog posts written within the last week. Or, better yet, Google it.

Get Fast Fast: A Project Management Approach To Lightweight Rowing

2011-04-17T18:08:56-04:00

This guide was written primarily for lightweight male rowers, but may useful for others as well. C.f. the section below on assumptions, caveats, and prerequisites.

You’re reading this because you want to get fast. Clearly. Otherwise you wouldn’t be here.

What follows is a comprehensive guide on going from being a complete couch potato to an Olympic-level athlete in six months. How is this possible? After all, normally this would take several years. The secret is focusing on the fundamentals. Consider the following passage, written about Japanese baseball:

"In every confrontation with a real American professional team it seems that what we need to learn from them, besides their technique of course, is how uniformly faithful their players are to the fundamentals. Faithfulness to the fundamentals seems to be a common thread linking professionalism in all areas." –T. Kageyama

The interesting thing is that Kageyama was actually a professional go player, and the quote above comes from a book about go. But as it turns out, what’s good for baseball players and go players is good for rowers as well.

Here are the basic principles behind this guide:

There are three components to fitness: cardio, strength, and flexibility. If you want to have the fastest 2k, then you need to have the most cardio, the most strength, and the most flexibility.
Getting fast at rowing is a project, just like writing software or building a house. The best way to improve strength, cardio, and flexibility as fast as possible is to use the same research-based methodology behind all other project management.
Each of the three components of fitness is broken into various subtasks. Each workout is designed to maximize the measurability of one’s progress in each subtask over time.

Most other training guides start out with a 2k or a series of max lifts, and then attempt to divine a new best workout each day based on initial performance. These top-down programs never work for two reasons:

They don’t have you doing the most efficient workout on any given day.
They fail to generate actionable data.

These ‘systems’ are mostly designed to keep the kids from getting bored, and to make the person who created them seem smart.

In contrast, by holding our workouts constant we can easily track our rate of progress and make data-driven decisions quickly. This allows us tweak our workouts to keep our rate of improvement in each exercise linear for as long as possible, rather than having it prematurely reaching the point of diminishing returns. I will follow up with several real world examples of this below, as well as some raw data from my own personal experience.

The ultimate goal is to go from your current level of fitness (or lack thereof) to being able to pull a sub-6:20 2k in six months. In order to break 6:20, here are the benchmarks you need:

Aerobic cardio: sub-1:55 for 2 x 45’ @ 155hr / 19spm
Anaerobic cardio: >2900m avg. for 3 x 10’ @ 24spm (4 min), 26spm (3 min), 28spm (2 min), 30spm (1 min)
Squats: 320-360 lbs max
Pull-ups: 35

If you can achieve each of these goals then you are pretty much guaranteed to have a sub-6:20 2k. Notice how we are breaking the 2k down into its component fitness parts, each of which we can directly measure against the data generated from our daily workouts. This way there is never any guessing. Each day you know exactly how much progress you’ve made toward each goal, and exactly how many days it will take to reach each goal at your current rate of improvement.

So what does this actually look like in practice? It’s extremely simple. Here is the plan for the first three months:

Months 1 - 3

Monday – Saturday: Erg 2 x 45min @ 19spm, 75% max HR
Monday – Saturday: Stretching, core strength
Monday & Thursday: 6 x max pull-ups
Tuesday & Friday: 6 x 10 squats
Wednesday & Saturday: 4 x 10 back extension
Sunday: Rest

Erg

2 x 45 min @ 19spm, 75% max HR (around 155)
Take a water break every 15 minutes. The standard half-strength Gatorade advice is good here, i.e. one scoop of Gatorade powder for a 32 ounce Nalgene. Limit the water breaks to fifteen seconds so as not to mess up your numbers; this should be enough time to down at least 8 ounces of fluids.
Between the two sets eat something like a banana or a power bar. This helps prevent chronic glycogen depletion, the cause of most overtraining syndrome.
Take 15 or so min in between sets to through a full routine of stretches. After 45 min of erging your muscles will be nicely warmed up, so this is the ideal time to gain flexibility.
Keep the drag factor around 100 for the 90 minute pieces.
While spending 20 minutes between sets probably isn’t ideal from a purely cardio point of view, taking the time to properly stretch, eat, and hydrate goes a long way toward preventing repetitive strain injuries. And since pretty much the only thing that will keep you from breaking 6:20 if you follow this plan is injury, this is time well spent in the long run.
I'd recommend taking one extra day off from the 90 minute erg pieces every other week, and then cross training for one day during the weeks that you're not taking an extra day off. This means that you should be doing your 90 minute erg pieces 22 days out of the 30, with an additional two 90 minute pieces per month of cross training. However, don't erg any less than 22 days per month or else your cardio will take a dramatic hit.
Never skip a strength workout. Even one missed day will set you back at least a week. In economics terms, the bulk of the improvement comes from the marginal effort you put in near the end. That is, the first 80% of the effort gets you only 20% of the benefits.
Similarly, never skip a workout for an erg test. I’ve seen way too many rowers try to cheat the system this way, which makes no sense. By training through every piece you ensure that even if you’re a little slower in fall, you’ll be miles ahead in spring. This takes a certain amount of self-confidence and assuredness in your ability to make the boat, but the dividends here are enormous. Go easy the day before your last 2k of the season, but train through everything right up until then.

Pull-ups

6 sets of as many pull-ups as you can do.
Record how many you get on each set in your notebook, but in excel record only the number you get on the first set and the total for all six sets.

Squats

Start with the maximum amount of weight you know you can safely do for 6 sets of ten reps. Each time you can successfully completely the 6 x 10 two workouts in a row for a given weight, increase the weight by 10 lbs for the next session. Always use a squat belt.
If you complete the lift just barely both times then keep with that weight for another session or two, especially as the weight gets heavier. Again, if you follow this workout plan then pretty much the only thing that can stop you from pulling a 6:20 is getting injured, so don’t fucking get injured. If there is ever a question about safety then immediately reduce the weight.
All squats should be full squats, i.e. the Olympic-style squats described here.
It generally helps to put a 5 or 10lb plate under each heal. This lets you push with your entire foot while keeping good form.
Once the squats start getting very heavy it’s good to lie on the ground between sets, and with your legs on elevated (e.g. resting on a chair or on the erg) do crossovers with 5 or 10lb weights. Start with your arms lying on the ground completely extended, as if you were making a snow angel. Then slowly bring each arm across to the other side of your chest with the weight in each hand. The point of this isn’t to get a workout, it’s just to help drain the lactate and other cruft out of your legs. Doing this for a minute or so between sets dramatically improves recovery.

The method of lifting described above is extremely taxing. Especially as you get better, it’s not uncommon for your heart rate to be close to its maximum after each set. Because of this it’s ok to wait a few minutes between sets, especially in hot weather. Remember, the reason we lift this way is that it’s the fastest way to improve our strength. Don’t try to rush through the lifts to gain cardio; this vastly slows down one’s rate of strength improvement, a tradeoff that makes little sense considering that strength takes so much longer to gain than cardio. Even worse, rushing through lifts gives only a nominal additional cardio benefit if you’re already doing the 2 x 45 ergs, resulting in a huge overall net loss in rate of improvement.

Similarly, the best way to complete these workouts is to do the lifts before the cardio. The thinking is that even if you can’t erg as fast at the 155 heart rate after lifting, you’ll still be getting roughly the same cardio benefit. Whereas if you’re lifting less weight each week because you’re lifting after cardio, then ultimately it’s going to take much longer to gain strength.

Months 3 – 4.5

Monday – Saturday: Erg 2 x 45min @ 19spm, 77.5% max HR (around 160)
Monday – Saturday: Stretching, core strength
Monday & Thursday: 6 x max pull-ups
Tuesday & Friday: 6 x 8 squats
Wednesday & Saturday: 4 x 10 back extension
Sunday: Rest

Monday & Thursday: 3 x 10’ @ 22spm (4 min), 24spm (3 min), 26spm (2 min), 28spm
Tuesday & Friday: 2 x 1250m @ 32spm
Tuesday & Friday: Extra 45 min of cross training cardio.
Wednesday & Saturday: Varies

I’d recommend the lifts and 2 x 45 erg in the morning, and then the harder stuff in the afternoons. Do the Thursday and Saturday pieces on the erg. The rest of the hard pieces should be done on the water, unless your goal is purely to get a fast erg time. Do at least 30 min cardio during the warm up with a few power 10s or 20s, and do an active recovery between pieces.

Wednesday and Saturday is your choice of anaerobic cardio. Good options are 2 x 30’, 3 x 20’, 10k, 5k, etc. The point is to do at least one hard piece every day. After all, once you know how to pull hard, being in shape is just icing on the cake.

The cross training should be done on the elliptical, the bike, or running. Do elliptical at the same heart rate you would erg at for the 2 x 45, do the bike at 10bpm lower, and for running do 10bpm higher.

Months 4.5 - 6

Monday – Saturday: Erg 2 x 45min @ 19spm, 77.5% max HR (around 160)
Monday – Saturday: Stretching, core strength
Monday & Thursday: 6 x max pull-ups
Tuesday & Friday: 6 x 6 squats; superset each with 3 jumpies
Wednesday & Saturday: 4 x 10 back extension
Sunday: Rest

Monday & Thursday: 3 x 10’ @ 24spm (4 min), 26spm (3 min), 28spm (2 min), 30spm
Tuesday & Friday: 2 x 1250m @ 34spm
Tuesday & Friday: Extra 45 min of cross training cardio.
Wednesday & Saturday: Varies

At the end of the six months you should be easily hitting the goals outlined at the beginning. If so, then you either should be able to break 6:20 on pure cardio, or else be very close. If you’re hitting the goals but you’re not quite there yet, then here is what to do:

Months 6 – 7.5

Monday – Saturday: Erg 2 x 45min @ 19spm, 77.5% max HR (around 160)
Monday – Saturday: Stretching, core strength
Monday & Thursday: 6 x max pull-ups
Tuesday & Friday: Plyometrics for legs (i.e. jumpies)
Wednesday & Saturday: 4 x 10 back extension
Sunday: Rest

Monday: 3 x 10’ @ 24spm (4 min), 26spm (3 min), 28spm (2 min), 30spm
Tuesday & Friday: 2 x 1250m @ 34spm
Tuesday & Friday: Extra 45 min of cross training cardio.
Wednesday, Thursday & Saturday: Varies

For the plyometrics, the simplest program is to work your way up to 3 x 100 jumpies. For example, start with something like 6 x 20 your first week, and then slowly work your way up.

On Wednesdays, Thursdays, and Sundays, do shorter anaerobic sprint pieces. Good examples are 5 x 2000, 5 x 5’, and 6 x 2’.

Complete at least two 8 x 500 ergs before your next 2k. Do the pieces at the following stroke rates: 26, 28, 30, 32, 29, 31, open, open. Rest for 1:45 after each piece, and 8 minutes after the fourth. If you can break a 1:35 average for all eight pieces then you should now be able to break 6:20 on the 2k.

(Alternate version: 6 x 500m open. Rest for 1 min after each piece. Drop the fastest, drop the slowest, and the average of the middle four is your 2k split.)

The most important thing to remember here is never skip the 2 x 45’ in favor of the sprint pieces. Approximately 70% of your 2k speed comes from base cardio, and the sprint workouts contribute only a small fraction to the remaining 30%. At the elite level about 100k per week is the break-even point. That is, if you’re doing more than 100k per week you’re gaining cardio, and if you’re doing less than 100k per week you’re losing cardio. That’s not to say that doing sprint work isn’t extremely important, but as it gets closer to sprint season a lot of people start doing less cardio and only focus on the sprint pieces. This is a huge mistake. Never do sprint workouts unless you’re already doing at least 100k per week.

Assumptions, caveats, and prerequisites

You are a lightweight male between the ages of 16 and 38. If not you can always modify goals and workouts as long as they remain consistent with the principles outlined above.
This guide is written for people trying to go from anywhere under 7:00 to just under 6:20. It's not intended for beginners trying to go from 9:00 to 7:00, nor is it meant for those trying to go from 6:18 to 6:08.
At least 3+ years of general fitness, including basic experience with strength training, cardio, stretching, and competition. You don’t need to currently be in great shape, but you do need to know what you’re doing and have some basic muscle memory in place.
Getting fast is either the first or second most important thing in your life. Meaning that if your number one priority is school/work, then training has to take priority over hanging out with friends and significant others, partying, staying out late, surfing the web, etc. This system will only work if you have a love of sport and a strong intrinsic motivation to be the best.
You know how to erg, row, and lift weights safely. If not visit Row2K to find a learn-to-row program in your area.
This guide assumes you have at least a minimum of non rowing-specific strength. You don’t need the perfect beach body by any means, but you should hopefully be able to do 20 dips and bench 155. If not then spend six weeks or so working on this before you get started, perhaps concurrently with a longer version of the two-week cardio plan outlined below. The reason is that for the next six months you’ll be focusing almost exclusively on rowing muscles, so you want to make sure your antagonistic muscles (pecs, triceps, hamstrings) are strong enough that you don’t injure yourself. Some of the book recommendations below may be highly useful for creating an appropriate program.
Similarly, you'll want to go back to training your antagonistic muscles, as well as doing core and cross training, for 2.5 - 3 months once this program is over before repeating this program next year. Otherwise you won't have enough base strength in your non-rowing muscles to avoid injury.
If you’re completely out of shape when starting this, then you might want to take at least two weeks to ease into the erging. Try something like 15 minutes the first day, 25 minutes for the next two days, 35 minutes for the next three days, 45 minutes the next four days, and 2 x 35 min for 5 days. Hopefully you are steady stating at least under 2:30 by the time you start the full workout plan.
The biggest caveat here is with the pull-ups. For me I found that I gained strength very quickly using this method, but then eventually topped out around 35. While this is really all you need to break 6:20, I feel like there are probably ways to improve upon this component after the first three months. Again, the reason you are plotting your progress in excel is so you can see if your improvement is starting to level off. If it does, figure out why and modify the workouts as needed.

Suggested variations

Consider substituting some of the pull-ups for bench pull after the first three months, especially if you are already close to hitting your pull-ups goal. The only reason I didn’t include this in the standard plan is that most people don’t have access to the necessary equipment. Basically the suggested workout would become:

Monday: 6 x max weighted pull-ups
Thursday: 6 x 20 bench pull

Once you’re able to do over 30 pull-ups on your first set without weight, add 10 lbs using a dip belt. Then add another 5lbs every time you’re able to do 20 or more on your first set for two workouts in a row.
Start on plyometrics early if you’re already close to hitting your goal for squats after the first three months.
If you reach a point where you’re no longer making acceptable improvements in your cardio, trying alternating light and heavy weeks. I find that doing over 144k per week on average is unsustainable, but if you’re already pushing this limit and struggling to improve then try alternating between doing 185k one week and then only 100k the next. This seems to work well, but I’d recommend against doing it more than two cycles in a row.

Stretching

If you want to be the fastest then you need to have the most strength, the most cardio, and the most flexibility. Unfortunately all too often stretching takes the back seat behind the other two components of fitness even though it’s equally important. Therefore I’m taking the position that the best stretching routine is the one that you do every day. And the best way to do this is to create a routine that you actually enjoy because it feels good. To that end I’d recommend doing the stretches below in the order listed. Doing the wrong stretches or doing them in the wrong order can be really tedious. The idea here is to create something that’s aesthetically appealing while at the same time hitting all the muscle groups required to keep you safe and maximize your 2k.

Aerobic Flexibility - Stretches #1, #3, and #2
Wrist flexors stretch.
Shoulder stretch.
The hurdler stretch. First with each leg extended, then with both legs extended. The goal here is to get your chest as close as possible to your knees.
Quads stretch.
The thoracic stretch.
The butterfly stretch.
The cobra stretch.
The Achilles stretch. If you prefer this can also be done on all fours (pushup position) instead of against the wall.

I’d recommend stretching between the 2 x 45’ erg pieces. This helps keep your muscles operating at their peak going into the second piece and may also offer some protection against injury. Hold each position for at least 30 seconds. If you want to add in any dynamic stretches then do them after the static ones. I’d also recommend stretching again after the 2nd 45’ piece.

If it’s a lifting day then warm up for at least ten minutes and stretch out fully before lifting. I find that for that quad stretch flexibility gains tend to be limited after the first couple months, but it’s still important to always do this before and after squats so that your muscles fibers don’t shorten up as they heal.

If it’s a race day then warm up as usual but only stretch lightly. This is because heavy stretching makes your muscles significantly weaker for the next few hours. This has only a minimal effect on your steady state cardio, but it makes a big difference when it comes to racing.

Core Strength

Core strength is popularly believed to prevent chronic use injuries. While the scientific evidence is somewhat equivocal I think it’s safe to assume that at least a solid baseline level of fitness here is necessary, if not a p90x beach body physique. Here are three exercises below:

Pillar bridges – Get into pushup position. Hold for 10 seconds. Raise and extend right arm. Hold for 10 seconds. Switch arms. Hold for 10 seconds. Raise and extend right leg. Hold for 10 seconds. Switch Legs. Hold for 10 seconds. Raise and extend right leg and left arm. Hold for 10 seconds. Switch arms and legs. Hold for ten seconds. Hold pushup position for 10 seconds.

The way you get better than this is by increasing the amount of time you can hold each position. Try to get to the point where you can hold each position for 25 seconds. Be careful not to let your lower back sag because this can cause damage.
Cherry Pickers - Remember, the goal isn’t just to move your arms back and forth, but rather to isolate your core muscles and use them to twist your entire trunk from side to side.
Leg Lifts - Lie flat on back with head and shoulder blades off ground. Lift both legs so heels are six inches off the ground and knees are straight. Hold for one minute.

This above shouldn’t be taken as a complete routine, but rather as just a few exercises to get started. You should supplement the above by layering in other core exercises as necessary.

Food

I’d recommend picking up Nancy Clark’s book on sports nutrition. This book is invaluable, especially when cutting weight for race season. No special diet is needed to complete these workouts, just remember to eat something in between 45-minute pieces to prevent chronic glycogen depletion.

My only other tip is to try the Seth Godin breakfast. It’s healthy, low calorie, and delicious. Try it with cilantro, sprouts, and grape tomatoes.

Learning to pull hard

Knowing how to pull hard is the most important thing in rowing. However it’s also something this particular guide doesn’t put a lot of emphasis on. If this is something you need to work on then try adding in the following sprint workouts:

Erg: 2 sets of 8 min @ 14 spm
Erg or Water: 4 sets of 10 on 10 off 10 times. If you do this in the boat try dragging the bucket on the first and third set.
Gamut: 1min on, 1min off, 1 min on. The original gamut erg is an excellent tool for learning how to pull hard because there is a lot of resistance at the catch and very little at the finish. Because of this it rewards the out-of-control energy that makes the 8+ win championships.
Erg: 100 meter sprints
Erg or water: 500m sprints. (Or quarter mile sprints if rowing the 1x.)

Gimmicky Ideas

None of the stuff below is necessary, but here is a list of ideas for if you have some extra time and want to experiment.

Supplements - The only supplements I'd recommend are the ones that are generally recommended anyway for health reasons: a basic multivitamin like this or this, Omega-3, and Co Q10. That said, if you want to try something crazy, try drinking a Red Bull and a couple teaspoons of honey 30-45 minutes before a big erg test. It sounds disgusting, and it kind of is, but the raw energy surge and pulling aggression you get from this is completely ineffable. (But don’t have more than one because WADA and the NCAA have strict limits on acceptable caffeine use.)
Hypoxic tent – This is a low oxygen tent that you sleep in to increase your red blood cell count. This allows your blood to hold more oxygen, which improves athletic performance.
Power Lung – Supposedly strengthens lung strength. The claim is that A) the feeling of gasping for breath you get at the end of the race happens because your lung muscles are exhausted, not because you’re not getting enough air. And B) that gasping for breath wastes an enormous amount of energy that should be going into your rowing.
Jump training - If you’re more ambitious with your plyometrics you can purchase one of the vertical jump workouts designed for basketball. There are several readily available that you can find via a simple Google search, although I can’t personally vouch for the safety or efficacy of any individual program.
Logging workouts - In addition to keeping a journal with all of my workout data, I also like logging my workouts in more creative ways as a more visceral reminder of how much I’ve accomplished. One example of this is eating one power bar for every 90-minute piece you do, and then saving the empty boxes. The greatest idea I ever came up with was switching showers after each cardio workout. The gym where I went to college had 72 showerheads, so with an average of roughly 22,000 meters per 90-minute piece it took well over 1.5 million meters to work my way around the facility.
Positive visualization - There are several books about applying positive visualization to sports. Brad Alan Lewis's book Assault On Lake Casitas actually has a whole chapter about this.

Other training guides Although I haven’t actually talked much about project management explicitly, hopefully it should be clear how this program was inspired by and designed to be compatible with project management tools like iterative development, total quality management, PERT, critical path method, etc. If not, try diffing this guide against some of the other popular training methods below:

Concept2 Training Guide for indoor rowing.
Rowing Faster - A book about rowing training by Volker Nolte.
Lactate Threshold Training - As the reviewers point out, this book is poorly edited, difficult to understand, and even contradictory in places. But there are still some good ideas if you're willing to slog through it.
Super Sport Systems - ($488 per year)

I actually think reading through these guides would be hugely beneficial to anyone willing to take the time, but I wouldn't actually follow any of them outright for the reasons outlined in the introduction.

Books and movies about rowing:

Sports training books and heart rate monitors

Hacker News user PaulHoule has an excellent set of book recommendations for sports training:

"I'm skeptical of anybody who claims to have a simple answer to success at weight training. The more I learn about weight training, the more ignorant I feel.

A few books I've enjoyed lately are 'Starting Strength' by Rippetoe and Killgore and Brookfield's 'Mastery of Hand Strength'. The latter book is a real eye-opener: every page is stuffed with information that makes sense, but you never would have thought of. (Hand development, of course, is important for computer keyboardists.)

Although it's not a complete weightlifting manual, I like the attitude of Bruce Lee's 'Art of Expressing the Human Body', which turned me on to circuit training. Some other good books which are more focused on bodyweight training are Cook's 'Athletic Body in Balance' and Boyle's 'Functional Training For Sports'."

Men's Health has a similar set of book recommendations. It's important to note though that these books aren't specific for rowing, so many are geared toward training entirely different muscular aptitudes. Rowers are generally trying to maximize their average power per unit muscle mass over a 240 stroke race, so it would be an enormous mistake to simply follow a training guide designed to bulk you up as fast as possible.

As far as heart rate monitors go, you don't need anything fancy. You just need something that will be able to display your current heart rate, as well as your average heart rate at the end of your workout. Currently the two best heart rate monitors that fulfill these needs seem to be the Polar FS3C and the Timex T5G971.

My personal data

As promised, I want to share some of my personal data about how this program worked for me. When I started this program I had already rowed previously for many years at a reasonably high level, before taking a 1.5 year hiatus. During this time I was almost completely sedentary, except for things like walking around campus between classes. As such, I was completely out of shape when I started training again and could barely keep my erg avereage below 2:30 at a 155 heart rate. I took 2 - 3 weeks to work my way into the program, as described above, and then followed the program diligently for the next 4 months.

At this point I unfortunately had to take 2.5 months off due to a medical problem that was being exacerbated by training. After these 2.5 months I trained lightly again for a month, attending team practices but not erging or lifting at all on my own. After these few weeks back I was able to pull a 6:23.3 on the erg, even though I was nowhere near in the physical shape I had been before. Thus it stands to reason that even if I wasn't in good enough shape to have broken 6:20 before this break, I probably would have come very very close.

More importantly, I had easily achieved my training targets for pull-ups and squats, and was less than a month away from achieving the erg targets as well. So, as it stands, while I'm not the fastest collegiate lightweight rower ever to have rowed, you'd be hard pressed to find anyone who got to my level of fitness faster than I did. Not only this, but the data suggests that I still had a ton of potential left as my rate of improvement hadn't yet started to significantly decrease. But don't take my word for it, download my workout data for the first four months and see for yourself.

Final thoughts

If you make it through the above then you should have little difficulty in ultimately breaking 6:20. The real trick is staying motivated.

So far as I can tell there are really only two ways to go about this. The first is to regularly visualize yourself winning and achieving your goals. The idea here is to motivate yourself by the thought of getting to go out and fuck up someone else’s day. If you’re the kind of person who enjoys this then it’s theoretically possible to make it work almost indefinitely.

However, the better longterm strategy I think is to adopt what Hinduism calls the sanyasa-tyaga mindset. That is, take the right actions, but then renounce the consequences. If you can learn to enjoy the process for its own sake then this is the true path to sustainable fastness.

Lastly, don’t be afraid to make changes. But only make changes that are based on data, and that will generate more actionable data. Good decisions are based on patterns, so plan ahead and create workouts that make these patterns easy to spot. As Yogi Berra once said, “You can observe a lot just by watching.”

How writing creates value

2010-06-13T15:18:53-04:00

What makes something insightful, informative, interesting, or funny?

This is one of the most interesting questions in all of linguistics.

Why?

Because having a good answer would allow us to much better understand how writing creates value. What’s more, it would enable us to create a framework for analyzing whether or not any specific piece of text does so.

In this essay I’ll attempt to answer that question. What’s more, I’ll show how each of these four rhetorical features of language was shaped through biological evolution. I’ll do this by mapping these features onto a Piaget-inspired theory of learning.

Once we can objectively say what makes something insightful, interesting, informative, or funny, we should be able to, among other things:

Create a search engine that returns results structured in the way that we think.
Write textbooks that leverage our intrinsic motivation to learn, making learning more fun and effective.
Discover new ideas by understanding the properties of the structures that underlie all ideas.

Exciting? I think so at least.

Let’s begin by taking a look at what makes something interesting.

What makes something interesting?

The word interesting is traditionally defined as “arousing or holding attention.”

 The problem with this definition is that it isn’t actionable because we can’t predict whether a piece of text will hold someone’s attention until they’ve already read it. This isn’t very useful.

Therefore I would like to propose a new term: something is objectively interesting when it violates an expectancy, or when it suggests the existence of a new pattern.

The key here is understanding what is meant by an expectancy violation. To explain, consider the methodology of Baillargeon’s now classic experiment in cognitive development:

Infants' well-documented tendency to look longer at novel than it familiar events suggested [a] method for investigating infants' beliefs about objects. In a typical experiment, infants are presented with two test events: a possible and an impossible event. The possible event is consistent with the expectation or belief examined in the experiment; the impossible event, in contrast, violates this expectation. The rationale is that if infants possess the belief being tested, they will perceive impossible events as more novel or surprising than the possible event, and will therefore look reliably longer at the impossible than at the possible event.

Using this violation-of-expectation method, investigators have demonstrated that even very young infants possess many of the same fundamental beliefs about objects as adults do. For example, infants aged 2.5 to 3.5 months are aware that objects continue to exist when masked by other objects, that objects cannot remain stable without support, that objects move along spatially continuous paths, and that objects cannot move through the space occupied by other objects.[1]

In each of these experiments the scientists created an illusion whereby an expectation was violated. For example, one of the scenarios involved a toy car rolling down a track and seemingly driving through a solid wooden block and continuing along the track on the other side. What the scientists found was that the infants looked at the impossible scenarios for considerably longer than the possible scenarios. Because our expectations come from our schemas, these experiments were a way have showing that infants do have the basic schemas needed to reason about physical objects.

The bigger point though is this: these expectancy violations made the impossible scenarios objectively interesting relative to the given schema. That is, something is objectively interesting relative to a given schema if it violates an expectation that schema logically creates.

Of course the traditional, subjective definition of interesting still stands as well. What’s more, there is a clear relationship between the two. If a phenomenon is objectively interesting relative to a given expectation, and there exists an individual who holds this expectation, then we can predict that this phenomenon will create a subjective feeling of interest that arouses or holds this individual’s attention. At least for a while. Once the novelty is gone this subjective feeling of interest will fade for that individual, though the scenario can still be said to fall into the ‘interesting’ bucket in the objective sense.

From now on, I will use ‘interesting’ strictly in the objective sense unless otherwise noted.

The role of ‘interesting’

The reason why expectancy violations feel subjectively interesting is because they are the best places to look for new insights, information, and humor.

Let me offer an example.

My college biology teacher sophomore year liked to tell a story about putting a terrarium in the freezer for a few months to see what would happen. His plan was to eventually take the terrarium out of the freezer, and then compare the growth rate of plants in that soil to the growth rate of plants in soil collected on the same day from the same location, but which had not been frozen.

When he unfroze the first terrarium though and dug through the soil he found a little bug that was still alive. This surprised him, and made him suspect that there must be an interesting story behind this. In other words, how was the bug still alive after several months in the freezer? The implication being that if he had been willing to investigate further, there was probably some novel scientific discovery waiting to be found.

The moral here is that interesting stories are valuable because they are often a source of insight and useful information. Had my professor pursued the question further he almost certainly would have discovered (or at least learned) something new.

What makes something insightful?

A piece of text is objectively insightful if it creates a new schema or contextualizes an existing schema. Since one can’t do this without highlighting breakages in how we currently reason about a given phenomenon, anything insightful could be said to be interesting as well. However, in practice one wouldn’t describe an insight as interesting for the same reason one wouldn’t describe a senator as a congressman, even though the senate is a branch of congress. That is, because insights are so much more valuable than writing that is merely interesting, describing insights as being interesting wouldn’t do them justice.

In the same way that not everything interesting in the objective sense is interesting subjectively, so too does not every insight feel insightful. A piece of text containing an insight will only feel subjectively insightful if the connection described is novel to the person reading it.

For example, we all have a schema ‘writing’ that connects pens with paper, but this does not feel insightful because we already know how pens and paper work.

Giving an example of a paragraph that actually feels insightful is harder because, as per our definition, what feels insightful varies considerably depending on our prior knowledge. But here are a couple paragraphs from the DailyKos user Vinifera that might fit the bill:

I have several friends (and a spouse) with college degrees that are irrelevant to the work they are now doing. And yet, salaries would be considerably lower without that degree on paper, unrelated as it may be. Basically the degree means you get an entry pass into the middle class.

What is also scary is that many of these degreed folks aren't actually producing anything. Meetings and project plans and reorganization do not constitute real work-- it's more like sound and fury signifying nothing. It feels like a hoax economy, set up to distribute money but without involving any actual production: you buy your pass, and you get entered in the desk jobs raffle and hope to win a high-paying imaginary career. It's extremely frightening.

The reason this paragraph probably feels subjectively insightful is because the author is creating a series of novel and meaningful connections between our schemas for skills, degrees, income, and career.

For more thinking on the common patterns underlying insightful writing, c.f. my previous essay How to blog insightful.

What makes something informative?

"Knowing the right answer requires no decisions, carries no risks, and makes no demands. It is automatic. It is thoughtless." —Eleanor Duckworth

Not all writing with information is informative. A piece of text is only informative if the information provided is meaningful within the context of the reader’s pre-existing schemas. Information without meaning is not informative. And if the text goes above-and-beyond this and changes our schemas themselves or the way they are structured and ranked, then we would describe that text as being insightful instead.

To give an example of something that’s informative, we must start with an existing schema. For example, let’s say we believe the best way to judge the utility of a purchase is to weigh the costs of the purchase against the other possible uses of those resources. If one has this belief, then an article about what else we could be doing with the resources going into Iraq may seem highly informative. Of course, the level of subjective informative-ness the reader would feel would be proportionate to how novel the information was to them, and how well they could incorporate the information into their pre-existing schemas.

At its best, informative writing can cause someone to change their beliefs and opinions, without having to go to the trouble of actually reprogramming their schemas or worldview. (A person’s worldview is roughly their collection of schemas and their relative importance compared to each other.) Unfortunately, as we said earlier, not all writing with information is informative.

For an example of how writing can contain information without actually being informative, consider the typical textbooks employed by schools and universities. While reading through my high school history textbook I always got the feeling we were only ‘fake learning’; that is, learning without really learning. Whenever I’d read a ‘real book’, I’d generally remember what I had read, but with my textbooks the information was always gone after a week or two.

The reason for this, though I could not articulate it at the time, was that the information in my textbooks was not connected to any schemas. I’ve always said that good writing changes the way you see something, and great writing changes the way you see everything. My textbooks changed the way I see nothing. That is, we’d read these paragraphs stuffed with names and dates, and yet our understanding of the world remained completely untouched.

An effective history textbook would organize each chapter around a different belief the author wanted the student to hold about the world. This is, of course, not allowed. It’s been banned due to the fact that real teaching, and real learning, is perhaps the truest form of subversion. Which leaves us with the shadow game of modern schooling.

But enough of that tangent.

Information that’s informative is always objectively interesting, because it shows that our previous beliefs were either mistaken or else incomplete. The former generally provides a place to look for new insights, and thus informative writing is interesting even if it doesn’t violate an expectancy directly.

The following Reddit comment by user Pseudotype provides a good example of a post that is both objectively and subjectively informative:

Ahmadenijad is fundamentally linked to the Basij, and the Basij are a national tradgedy for Iranians that is very difficult to reconcile for Western minds. During their war with Iraq, Iran needed massive numbers of expendable troops for human wave attacks into heavily mined no-man's-lands between fortified front lines (think WWI, but in the desert, and with more crazy). So a nationalist campaign was started to get every patriotic family to give one son to the Basij militia. Needless to say, it's easier to get children to do stupid things than adults. Thousands of these kids were slaughtered on the front lines, and some stories of how they were employed are truly sickening...civilized countries like Stalin's USSR used dogs for the kinds of suicide missions in which the Basij specialized. So, now the sentiment among those families, who sent their hundreds of thousands of children into battle, is that the Basij are a quasi-religious patriotic nationalist front that is incorruptable. For many Iranians, accepting that Basij are evil would mean accepting that they sent their own adolescent children to die in a meaningless act. That is why the Basij are so powerful and untouchable in modern Iran.

Ahmadenijad publically acknowledges that he was a military intelligence officer during their war with Iraq, and it is widely believed that he was a commander and trainer of Basij forces. He has been entitled to massive Basij uprisings every time people speak out against him. Through this mechanism, Ahmadenijad is able to subordinate atrocities such as we are seeing today without using official Iranian military or police forces.

Notice that none of our expectancies are directly violated, although the post is written in such a way that may lead us to create new insights on our own. In this way a writer can use information in hopes of getting us to change our own worldview, without having to go in and directly challenge any of our beliefs himself.

What makes something funny?

People usually say that what makes something funny is when it has an element of ‘absurdity’. The problem with this definition is that it doesn’t actually mean anything. That is, what does it mean for something to be absurd? And how do we know if something is or not?

According to the academic literature there are two types of humor: novelty humor, and breakage humor. That is, things are funny either because they’re new or else because they’re broken. But clearly this isn’t the whole story, because not everything that’s new or broken is funny.

Rather, there seems to be a meta-cognitive condition involved, whereby things are funny only if they’re broken or new in a way that seems like something shouldn’t be broken or new. This means that for something to be funny it must not only violate our expectations, but it must do so in a way that’s discordant with the reasoning behind those expectations. This is because behind each expectancy is a set of reasons for that expectancy based on our schemas, and ultimately these schemas create meta-expectancies, that is, expectations for our expectations. So when something is new or broken for an individual in a way that violates the meta-expectancies of that individual, then they will tend to find that novel or broken stimulus to be subjectively funny.

For example, with the Baillargeon toy car experiment, we might expect the car to continue rolling down the track until it hit the wooden block and stopped. But this might not happen, for instance the car might stop on its own before hitting the block or even fall off the track. These scenarios would violate our expectancy, but they because they still fall within our ‘expectations for expectations’ they would be merely ‘interesting’, rather than funny.

So why does humor exist? Because it’s our checksum on reality.

Humor focuses our attention on the validity of our schemas or the lack thereof, thus giving us the chance to confirm or improve them. In this sense, humor is sort of the halfway point between finding something interesting and creating an insight.

A good example of this, and another good example of humor, is the “You’ve got AIDS” sketch from Family Guy. The segment is funny because it not only violates our expectations for how the doctors will share the news of a terminal illness, but also our expectations of how this sort of news should be shared. The reason this sketch is so funny is that it not only violates our expectations, but also the schemas underlying those expectations. And the fact that these underlying schemas are quite valid makes the sketch even more funny. (Perhaps one could even say absurd.)

In addition to validating or casting doubt on the schemas of an individual, here are some other important evolutionary functions of humor:

Humor creates and strengthens shared beliefs. It does this because the meta-cognitive element is inherently leaky; that is, we can’t make a joke without revealing our actual beliefs about a subject. For example, let’s say we are making a breakage joke. We are instantly revealing that we believe that something should not actually be broken in this way.

The importance of this largely depends on whether the joke-teller is a comedian or a humorist. The difference is that a comedian creates absurdities arbitrarily. For example, think of Dane Cooking talking about flicking cashews off of his penis into his mouth. A humorist, on the other hand, talks finds and highlights the absurdities already present in society. For example, George Carlin talking about the way we use euphemisms. In the latter case, Carlin uses breakage humor to spread his ideas about what the ideal society should look like.
Humor is a way of putting the social acceptance of an idea up to vote. Laughter (or lack thereof) is a way of either validating or shooting down the ideas of the joke teller.
Of course, others often decide whether or not to laugh not because of the joke itself, but because of who is telling the joke. Thus in addition to creating and strengthening the consensus reality, humor can also establish the social hierarchy.
Humor can be used as a way of testing whether or not someone is a member of the community. When someone gets the humor of a community it’s a strong social signal that they also understand the underlying ideas and values of the community.

Thus, the reason why humor subjectively feels good is probably because when others laugh at your jokes it shows that they respect you as a person and agree with your ideas.

The Big Picture

If what you see tomorrow entirely fits with your experience today, you’re not paying attention. —Robert Anton Wilson

So how do we know when writing creates value? I’d offer this as a general litmus test:

Good writing makes us see something differently. Great writing makes us see everything differently.

I don't know how Slashdot did it, but they seem to have discovered exactly the four qualities that can make a piece of text good: being insightful, informative, interesting, or funny. If you either added a bucket or took a bucket away, the signal to noise ratio would go down.

In the long term, our understanding of the rhetorical structures of language will help us design learning systems vastly superior to the tools of today. But what can this knowledge help us do right now?

One thing, I think, is that it can help us to avoid much of the ‘fake learning’ that comprises so much of modern schooling and the media. These institutions have learned to exploit the mechanisms that make something feel subjectively interesting or informative, even when there is ultimately no insight to be gained.

Reading material like this is often triggers your pleasure centers in the moment, but there is no lasting value. Before you click that link, ask yourself whether you’re going to see something differently after reading. Or is this just another shock story designed to suggest the existence of a new trend, one that probably isn’t even there.

In addition to helping us avoid bad material, this theory can of course also be used to help us identify good material. While more work needs to be done before a fully automated system can be created, I believe we have enough theory already to enable efficient human-mediated systems to filter social news sites in a much more intelligent way than is done currently. Imagine a site like Reddit or Hacker News, but where articles were algorithmically curated to maximize the return on time spent. Not only would the articles be better, but the discussions would be better too because so many more people would actually be grappling and engaging with the material.

Impossible?

Only until someone goes out and builds it.

How to blog insightful

2010-03-28T13:13:20-04:00

This is a guide on how to blog insightful. Not insightfully. Because insightful isn't a writing style, it's a type of idea.

There are hundreds of guides to writing like a writer. This is a guide to thinking like a writer.

I suspect that most of us view insightfulness as having this magical and unpredictable character. This is not the case. Insights aren't something you have, they're something you create.

Let me explain.

When mathematicians are proving conjectures, they rely on a standard collection of tools and methods. Like mathematicians, writers also have a standard bag of tricks to draw from. What follows are a few tricks that writers use to create insight.

Specifically, this essay will explain both what makes something insightful, and also where insights come from.

I’m writing not from the perspective of a great writer, but rather from the perspective of someone who knows what creates value for him as a reader.

===What Makes Something Insightful?===

The Oxford American Dictionary defines insight as “the capacity to gain an accurate and deep intuitive understanding of a person or thing.” If we accept this definition, then insights seem to be tools of sorts. Very special tools that help us make sense of the world by illuminating the patterns and relationships between things.

And as tools, insights have the same properties as a hammer or a hacksaw. Let’s brainstorm some ways of writing about tools that would create value for the reader:

We can give readers a new tool.
We can teach them how to use an existing tool more effectively.
We can teach them when and when not to use a tool.
We can replace an old tool with a better tool.
We can replace a complicated tool with a simpler tool that still does the same job.

And so on.

Each insight is like a new tool. And just like a table saw, we must first make the reader aware of this tool, and then teach them how to use it, when to use it, why to use it, what to use it for, etc.

So how do we do this? Through the creation and contextualization of schemas.

Like insights, schemas are models for helping us comprehend phenomena. However, not all writing that deals with schemas is insightful. For example, when we see a pen we know that it is a tool for writing on paper. We know this because we have a schema that connects pen with paper for the act of writing. However, there is nothing insightful here because we are simply using an old schema. Insightful writing is only that which creates value by offering new schemas or helping us better understand existing schemas.

The good news is that there are some common patterns in the ways that schemas are created and connected.

===Patterns in Schema Creation===

Define a phenomenon

A person defines a phenomenon by finding a pattern, describing it, and giving it a name. This creates value by giving others the ability to collectively think about the phenomenon.

A good example of a recently defined phenomenon is Yak Shaving. Yak Shaving is the technical term for when you find yourself at least eight levels deep in a stack of jobs. For example:

I needed to file my taxes, but first I needed turn on my computer, but then I noticed the pillow on my chair was missing, so I had to find it, and when I found it I realized that the filling was coming out the sides, so I had to fly to Bolivia to shave a yak to stuff the pillow.

All this just so I could finish my taxes.

Describing this phenomenon and giving it a label allows others to think about it and discuss it. We can recognize when we’re engaging in the behavior. We can ask questions. For example, how much of our day do we spend yak shaving? Why? And is yak shaving a form of procrastination or a prerequisite to productivity? Giving this pattern a name allows societies to share common schemas—that is, mental models of the way the world works.

Naming the pattern makes it actionable at the level of society, accessible to teachers and policy makers and therapists and clergymen. Once this pattern has been named we give others the ability to recognize it as a negative habit and take action accordingly. By articulating something that many people thought previously but were unable to express we have created value.

In general, the more encompassing and more actionable the new model, the more value created. In situations like this where we already intuitively accept the model before someone articulates it for us, the meme is prone to spreading quickly and becoming part of common parlance.

There are countless examples of people who have created value by finding and naming patterns, but here are a couple of my favorites:

Seth Godin on eBooks

AJAX: A New Approach to Web Applications

Create a Hypothesis

While naming something merely identifies and labels a pattern, a hypothesis attempts to explain causality.

Anyone who watches Fox News knows that trailer parks are frequently ravaged by tornadoes. I have a friend who’s convinced that trailer parks are actually responsible for creating them.

This is a perfect example of a scientific hypothesis. It is eminently testable, because as soon as we observe a non trailer-park-generated tornadoes then we are able to falsify the hypothesis.

Compare this to a non-scientific hypothesis, such intelligent design. While intelligent design could be proven true under the right circumstances, it can never be falsified. Falsifiability is what separates a scientific hypothesis from hypotheses in general. All hypotheses can be evaluated using tools from philosophy, but only scientific hypotheses can be evaluated using the methods particular to science. Thus, while all types hypotheses are potentially useful, new scientific hypotheses tend to be more valuable since we have more tools we can apply to evaluate their truthfulness and create understanding.

Split One Schema into Two

Sometimes we find ourselves working with schemas that clearly have elements of truth, but that for some reason don’t seem to hold up well empirically. Often times this is because the schema we have in our heads is more broadly defined than the underlying phenomenon.

A good example of this is the phenomenon of prodigies. We know there are some people in society who are exceptionally talented in certain areas, and we call these people prodigies. We then have certain schemas that we apply to these prodigies in our quest for sensemaking.

But even though prodigies clearly exist, our schemas often seem to not hold up so well. For example, studies have shown that child prodigies are often not significantly more successful than the rest of us when they grow up. And similarly, many prodigious adults were completely unremarkable as children. Why is this? How is it possible for such exceptional children not to make anything of themselves, and for such exceptional adults to have been completely average as children?

Malcolm Gladwell observes that the reason for this is because when we describe child prodigies, we are describing people who are gifted at learning. Whereas when we describe adult prodigies, we are actually describing people are gifted at doing.

Because we are applying one set of sensemaking tools to both groups, our schemas tend to not hold up so well even though they are based on an underlying truth. The solution to this is to create one set of schemas for understanding and dealing with child prodigies, and another set of schemas for understanding and dealing with adult prodigies.

There are often areas where we engage in fuzzy thinking and apply one toolset to multiple distinct phenomena. As writers we can create enormous value by identifying distinct phenomena, and giving suggestions for how to think about each one.

Examples:

The Myth of Prodigy and Why it Matters

How to Make Wealth

Combine Two Schemas into One

In the same way that the clarity of our thinking suffers when we treat two distinct phenomena as one, the opposite danger also exists. That is, we can mistakenly hold two completely different sets of schemas about apparently distinct phenomena that are actually the same—the intellectual equivalent of a double standard.

For example, support for slavery was ubiquitous in the antebellum south with many reasoning that the practice was morally acceptable since blacks were not actually people. The slave narratives of the mid-19th century changed this. The authors created value by humanizing blacks, their memoirs personifying the emotions and indignities behind the statistics. Even as late as 1861, the year Incidents in the Life of a Slave Girl was published, black women were not yet considered to be women. Harriet Jacobs argued otherwise, persuading America that black women were just as capable as white women of possessing the feminine virtues – chastity, humility, loyalty, submissiveness, etc.

The insight that made the slave narratives so compelling was that blacks and whites both share the same basic set of characteristics that make us human—thoughts, feelings, hopes, dreams… And so by extension blacks and whites should be treated as equals under the law.

This same false differentiation can also be found at the root of most economic bubbles. Francis Wheen, in his book Top 10 Modern Delusions, writes, “Financial sophisticates in the 21st century smile at the madness of the South Sea Bubble or the absurdity of the Dutch tulip craze. Yet only a few years ago they scrambled and jostled to buy shares in dotcom companies which had no earnings at all nor any prospect of ever turning a profit. To justify this apparent insanity, they maintained that such a revolutionary business as the Internet required a new business model in which balance sheets were irrelevant. In short, they thought they had repealed the laws of financial gravity - until they came crashing down to earth.”

Writers can predictably create insight by showing that a phenomenon that appears to be qualitatively new, like dotcom economics, can actually be accounted for by existing models, albeit perhaps with a few twists.

Think On A Higher Order

One of the principle ways for a writer to create value is by making complex ideas easier to understand. There are several techniques for doing this; one of the most important, described here, is called black boxing. Black boxing is the term we use to describe the process of encapsulating a low-level phenomenon and expressing it at a higher level to hide complexity.

For a good example of thinking on a higher order, think back to your introduction to statistics class from college. Unless you were a math major, you probably didn’t learn the advanced calculus that underlies the common statistical tests. This is fortunate, because the math needed to derive these models is very difficult. But because your only goal in stats 101 is to learn how to use statistical models to make sense of data, the task is much easier. In this example the underlying calculus has been black boxed, and so we have the luxury of thinking on a higher order.

Are there even more abstracted ways of looking at statistics? I would argue yes. For example, what if you don’t need to know how to do statistics? What if you just want to learn how to interpret them? Taking a full-year college course seems overkill. Thus, we can create further value by producing a guide that only covers the higher-level skill of reading and interpreting statistics, not the lower-level skills associated with actually performing statistical tests.

We can see many examples of this in everyday life. Of all the people you see driving cars today, how many do you think could design one? Just about none. Why? Driving has been black boxed.

Similarly, you don’t need to understand cell biology to care for a pet. Nor do you need to understand chemistry to know about cell biology. At least up until a certain point…

Think On A Lower Order

Just as we can create value as writers by cutting out what’s not important and focusing on only the higher order details, we can also create value by looking at things on a lower order. For example, ecologists can look at the environment from the perspective of a microbiologist. Biologists can look at organisms from the perspective of a chemist. Chemists can look at molecules from the perspective of a physicist. And physicists can view the natural world from the perspective of a mathematician.

Why should we want to do this? One reason is the law of leaky abstractions. To quote Joel Spolsky:

Abstractions fail. Sometimes a little, sometimes a lot. There's leakage. Things go wrong. It happens all over the place when you have abstractions […] and the only way to deal with the leaks competently is to learn about how the abstractions work and what they are abstracting.

Going to driving school is a great way to learn how to operate a car, but that only works until the timing belt snaps or a spark plug goes bad. When that happens, no matter how good of a driver you are you can’t get the car to work until you open the hood and fix the problem or take it to someone who can.

Thus, in the same way we can create value by black boxing a problem and teaching something on a higher order, we can also create value by teaching the underlying mechanics (no pun intended).

If we can find no explanations for a phenomenon at one level of the stack, we can always look for meaning at a lower level. For example, a psychiatrist finding no explanations for a given behavior at the level of conscious might try searching for an answer at the level of DNA.

Individuals frequently fall into the trap of looking for solutions to problems only at the level of the stack they feel most comfortable working with. As writers we can create novel insights for our readers by learning to recognize when a problem can be more appropriately addressed by looking for answers on a higher or lower order of abstraction.

X is a subset of Y

Often there are times when we find ourselves working with a smaller model that is really part of a bigger model, even though this isn’t apparent at first. When situations like these occur there is an opportunity to create insight.

For example, the concept of ‘weight’ is a subset of the concept of ‘mass’. That is, we can use weight to accurately compare how much matter two objects have as long as we’re on earth. But as soon as we’re in space we need to use the concept of mass instead. By explaining to students how weight is a very specific instance of mass that works to compare two objects as long as we’re on earth we have created insight.

As another example, consider environmentalism. The goal of an environmentalist is to figure out how to equitably divide finite natural resources among multiple stakeholders who want to use these resources for competing purposes. For example, for any given national forest some might want to use the resource for renewable lumber, others might want to use it to open a ski mountain, and still others might want to use it to graze cattle. The goal of an environmentalist is to ethically divide this resource among the competing stakeholders who all want the resource for their own usage. By realizing that environmentalism (unlike conservationism) is really a subset of ethics, we can show that, among other things, environmental problems are best by using the toolsets that normative ethicists have developed over the millennia.

X and Y are both instances of Z

The reason insights are powerful is because they give readers new tools for creating ideas, tools that can later be used whenever the reader wants. Insights are like trees laden with fruit. Our job as writers is to build the reader a tree and teach them how to bang on it to knock some coconuts down whenever they’re hungry.

In his essay Hackers and Painters, Paul Graham’s central insight is that hacking and painting are similar because they are both acts of creating. While you’d have to ask him to be sure, I’d suspect that once he figured this out the rest of the ideas just sort of fell into place. That is, once we know that hacking falls into the same category as painting, it stands to reason that hacking and painting have many elements in common. For example

The reputations of both have a large random component introduced by fashion.
In both hacking and painting, it’s best to start by sketching.
Both software and paintings are intended for a human audience, so both hackers and painters must have empathy to do good work.

Once Graham sells us on the idea that both hacking and painting have commonalities as acts of making, we can draw upon the rich history of painting to gain a better understanding hacking. When we ask ourselves questions such as what makes a good hacker, or how hackers create value, it then becomes logical to look to painting for possible analogs.

Simplify a more complicated model

Writers can create value by simplifying a more complicated model. As explained in the introduction, a model is the set of relationships that explain any given system. (And a schema is a model that is simple enough for us to reason with intuitively.) However, the same system can often be explained through one or more alternative sets of relationships that are logically equivalent, but vary in their complexity.

For example, the Ptolemy’s geocentric model of the solar system technically did work, but it was very complicated. Simulating the solar system required a mechanical device with hundreds of widgets. This gave Copernicus an opportunity to create enormous value by changing the point of reference from the earth to the sun. The insight that the relationships between the heavenly bodies could be conveyed in a simpler way was the basis of the simplified heliocentric model, one of the most important advances in the history of astronomy.

How does one simplify a model? One way is by choosing better assumptions. As I’ve written previously,

The problem with assumptions is that they're usually correct. For certain people, at a certain times, in certain places.

The danger isn't that you'll sometimes be wrong. The danger is that you'll always be right.

That is, your assumptions about human nature will be true, but less useful than those of your competitors.

I’d posit that most models have one or more core assumptions that the rest of the properties of the system are logically derived from. Once one changes the core assumption, the properties of the entire system change. And if the model becomes simpler as a result while maintaining its accuracy, it’s safe to say that we’ve created value.

X is not mutually exclusive with Y

Writers can always create value by correcting false schemas. One cognitive mistake we see often is the belief that two things are mutually exclusive when in fact they are not. Because false beliefs in this category are so common, writers can consistently write insightful articles along the lines of:

It is possible to support the troops but be against the war.
Corporations can be environmentally friendly and still make a profit.
Atheism is not mutually exclusive with morality

And so on. It turns out that outside of mathematics two things are rarely ever mutually exclusive, at least when we’re talking about abstract concepts. Thus, when we hear people making statements that suggest otherwise this is a good area to look for opportunity to create insight.

===Conclusion===

Once you understand the formula above, it should become trivially easy to create an indefinite number of insights. Especially if you're using my KWL method for generating ideas, and thinking about those ideas hierarchically by using software like FreeMind.

What might be less obvious though is that this idea can make you a much better reader as well. I've always said that good writing changes the way you see something, and great writing changes the way you see everything. Insightful writing is great writing, especially if those insights are counterintuitive. (As a general rule of thumb, I figure if it's intuitive I already understand it, so I only bother to seek out books and articles that are counterintuitive.) If you keep this idea of insightfulness in mind when reading through the newspaper, you can skip over 98% of the stories and still be automagically smarter than just about everyone else.

Have better ideas using KWL charts

2009-06-14T12:36:39-04:00

KWL charts aid reading comprehension. That's what they were designed for at least.

What's a KWL chart? Let me explain.

Before reading a book each student writes down what they already know and what they want to know. Then afterward they write down what they've learned. They do this for two reasons. First, by forming questions they activate prior knowledge, which makes it easier them to learn. Second, it's easy for both the teacher and student to see exactly which ideas the student is and is not getting from the text.

That's the theory anyway. In practice, every kid hates KWL charts. Or at least I did.

Why?

They're completely useless. That's how they seemed at least. Plus they were annoying. "I already know know how to read, I already get the story, why are you making me fill out this dumb chart??"

But ultimately all this kvetching may have been a mistake.

Why?

As it happens, KWL charts are probably the single most important tool to improve creativity.

The reason is this. Every brilliant idea starts with a question. For example, let's say I want to know more about how it is that institutions and organizations use social status to exploit people. Chances are if I'm asking myself this question then this isn't my only question. Chances are I have a bunch of related questions. Like, how can one use the promise of social status in business to get employees to accept lower wages? How is social signaling corrupting our schools and universities? And so on.

The point is that when thinking about any sufficiently large topic you'll probably have one main question and then a whole bunch of sub-questions.

The problem is that when you start researching these questions you begin to learn. That doesn't sound like a problem, but it is.

What ends up happening is that you acquire all these new models that sorta answer your questions. For example, maybe you read Rosabeth Moss Kanter's academic papers on why people join cults like scientology. This all seems really insightful and intellectually gratifying. The problem is that this new knowledge displaces your original questions, even if they aren't fully answered. And what happens is that your original questions often don't seem as relevant or important in light of your new learning.

Even when they are.

Breakthrough insights always come from thinking about the space in between established knowledge and a good set of questions. And unless you write down your questions and theories in advance, what always happens is that you read a few books and then forget your original theories about how things worked. Which is bad, because they're often at least partially correct. So you had this brilliant idea, or at least the start of a brilliant idea, but thanks to your research it's been lost forever.

Even if you haven't completely forgotten your original theories, your new insights become inextricably bound with background knowledge in a way that makes it impossible to communicate your learning to the outside world. What happens is that you end up sounding like Shulgin trying to explain his experience with mescalin: "I understood that our entire universe is contained in the mind and the spirit. We may choose not to find access to it, we may even deny its existence, but it is indeed there inside us." What the fuck.

Here's how to avoid this problem.

Whenever you're thinking about a big problem, write down all your questions and background knowledge in advance. Don't just write down your main question, write down every question whose answer could conceivably be insightful or useful to your intended audience. Then write down all your background knowledge. Not just a paragraph or two, but write down all your subject knowledge and all your theories of how you think things are working. (Preferably in a mindmap.) Don't do any research until this is done. Not even a Google search.

What ends up happening is that when you have both your questions and your background knowledge written down, it becomes ten times easier to think clearly about whether your reading is truly answering your questions. It's infinitely easier to come up with and recognize new ideas. And what's more, you still have your original questions written down so you have a clear framework for expressing your ideas to others.

The takeaway is this. You don't need a new method for coming up with brilliant ideas. You already have them. But sadly as you continue to learn these brilliant ideas are often lost. By using the KWL method you not only preserve your original insights, but you can use your research to expand upon them as well. By setting up a purposeful system that allows us to diff our background knowledge against existing models we can generate far bigger insights than would otherwise be possible.

How intellectual pollution has crippled America's children

2009-06-07T12:05:02-04:00

Some countries are smarter than others.

You can tell it in a million ways. By the books people are buying. By the conversations they're having in their classrooms. By the politicians they elect. By the stories in their media. By the laws they enact, their rates of crime, and how they treat their prisoners.

Some countries seem capable of creating sensible policy through intellectually honest debate. Whereas in America and Great Britain, this hasn't happened for decades. Rather, our laws emerge from a haze of ignorance, apathy, and misdirection.

Why?

It all goes back to Dr. Spock and the industrial revolution. Let me explain.

In 1946, Dr. Benjamin Spock wrote a book that would change the course of history. His book, Baby and Child Care, became internationally famous for promoting the idea that when it comes to raising children, "you know more than you think you do." In other words, when it comes to raising your children you should follow your intuition. This idea proved revolutionary, and remains the dominant advice from pediatricians to this day.

The only problem is that it's completely wrong.

Think about it. When deciding how much calcium to give your kids, would you just eyeball the container of Tums and pick an amount that feels right? Of course not. And yet there are many other parenting best-practices that have been determined in exactly this way.

It's not that parents weren't raising their children this way before. They were, but since the 40's the wealth of scientific best practices have been largely ignored thanks to Spock's influence.

Which brings us to the industrial revolution.

As it turns out, a person's work environment affects much more than just their satisfaction at work. It alters their intuition and perception, their rhythms and routines.

Parents naturally raise their children according to their intuition, and as farmers left the fields for the factories their intuition about how best to raise their children dramatically changed. That is, factory workers began to raise their children using methods that echoed the way they were being managed in the workplace.

Sociologists have long suspected that, for example, children of factory workers would be implicitly taught that the best way to succeed was to keep your head down and obey authority. And while this might in fact make them more likely to succeed in a factory, it would also stifle their chances of upward social mobility.

There is no doubt that the worldview parents impart on their children vastly affects their chances of future success. But this is where the damage was thought to end. Experts assumed that because the effect of parenting styles on cognitive, physical, and emotional development is less salient than with the example of calcium above, parenting styles didn't really matter. It was all just a matter of preference. And that while the children of factory workers may lack certain advantages, this could be largely ameliorated later given enough school intervention.

But thanks to modern early childhood development research, we now know that the effects of factory-style parenting are much more toxic than previously assumed.

For example, the research of Hart & Risley has shown that over half of the variance in a child’s vocabulary at age 3 can be attributed to the ways in which a parent talks to their child, and the way parents talk to their children varies dramatically depending on socioeconomic status. Specifically, they estimate that by age 4 the children of professional parents have been exposed to about 45 million words, whereas the children of welfare parents have been exposed to only about 13 million words. Because of this by time children are 3 years old, parents in less economically favored circumstances have said fewer different words in their cumulative monthly vocabularies than the children in the most economically advantaged families in the same period of time.

"Even if we have overestimated by half the differences between children in amounts of cumulative experience the gap is so great by age 4 that the best that can be expected from education or intervention is to keep children from falling still farther behind. For an intervention to keep an average welfare child's experience equal in amount to that of an average work-class child would require that the chid be in substitute care comparable to the average professional home for 40 hours per week from birth onward." That is, already by pre-school the low-SES children are so far behind in language development that it is impossible for them to catch up.

Similarly, Hart & Risley found that the average child in a professional family was accumulating 32 affirmatives and 5 prohibitions per hour, a ratio of 6 encouragements to 1 discouragement. The average child in a working-class family was accumulating 12 affirmatives and 7 prohibitions per hour, a ratio of 2 encouragement to 1 discouragement. The average child in in a welfare family, though, was accumulating 5 affirmatives and 11 prohibitions per hour, a ratio of 1 encouragement to 2 discouragements.

(For more on how language exposure and the ratio of encouragements to prohibitions affects childhood outcomes, c.f. Hart & Risley's excellent book Meaningful Differences in the Everday Experience of Young American Children.)

There's no especially good reason for low-SES parents to talk less with their children and use more prohibitions. They're just following Dr. Spock's advice and raising their children according to their intuition. Which, as it happens, is to manage their children the way their employers manage them at work. Raising children like employees has benefits that are immediate and hugely salient, whereas the harms created are subtle and visible only in aggregate through statistical analysis.

These harms are longitudinal; even as societies transition toward knowledge work, these flaws in parenting remain and reproduce themselves in future generations. When figuring out how to raise their children, parents look to the way their parents raised them. Parenting styles which caught on overnight may take hundreds of years to be displaced, because in each case the benefits are more salient than the drawbacks.

What's more, as white-collar jobs have replaced factory work an entirely new set of parenting flaws has arisen. Once again parents are raising their children according to their intuition, intuition which has been shaped and molded by their experience in the workplace. Success as a white-collar worker largely means developing the skill of being chosen: for the soccer team, for college, for internships, employment, promotions, and so on. This is the mindset underlying the methodology that high-SES parents adopt when raising their children.

Are these parenting practices good for a high schooler? Perhaps. But when applied to infants they're empirically damaging.

Many high-SES parents over-schedule and overstimulate their pre-k children in an attempt to impart an early educational advantage. By depriving their children of time for self-directed play they may be undermining executive function, the best predictor of future success. source
By forcing their children to read before they're ready, they may well be robbing their children of intrinsic motivation to read and contributing to future alliteracy. source
By letting their infants watch baby videos they are at worst contributing to future cognitive and behavioral problems, and at best dramatically decreasing their children's rate of learning. source

Again, these parenting practices stem largely from the intuitions parents pick up from their workplaces. Compared with using intuition to decide how much calcium to give a child, using intuition to make decisions like how much TV to allow an infant to watch is equally absurd and equally damaging, but largely ignored because the intangibility makes these decisions less salient and seemingly less important.

You can't measure bad parenting in parts per million. But the effects are just as real as lead poising, obesity, or thalidomide. And as with factory workers, these parenting mistakes will be passed on long after these working environments are gone.

I don't know why it is that certain countries seem so incapable of setting rational and coherent policy. I'm sure there are dozens of reasons. But I suspect a good percentage of the problem stems from a series of specific parenting flaws largely attributable to parents raising their children with intuition acquired in the workplace. The reason then that these social problems are most pervasive in countries like America and Great Britain is because these countries were the earliest and most extensively industrialized.

Every time we buy a cell phone, a flat screen TV, or anything produced in a factory, the damages are more than just the environmental contaminants and individual suffering. Rather, we're imposing a cultural shadow on ourselves, one which cripples our children and our children's children. This is a negative externality of the way we choose to structure society, a form of intellectual pollution far more harmful than anyone could have predicted.

This suggests a market opportunity for teaching parents to separate their parenting styles from the way they spend their day. But what's more, it suggests new modes of civic responsibility, and new answers to the age-old question of what it means to be a good person.

Is College Obsolete?

2009-04-14T20:47:54-04:00

Many people say that our education system is broken. It's not. Our system of education is obsolete. What may have made sense one hundred years ago no longer makes sense today.

One hundred years ago college didn't matter. Maybe for bragging rights, but not for getting a job. People lived in the same town their entire lives. Everyone knew everyone else and anyone could vouch for you. It was a network; the most advanced form of social organization.

With the advent of modern transportation, people were no longer fixed. After the first world war, people could live anywhere they wanted, and increasingly they did. No one knew anyone and so no one could vouch for you. Transportation and surpassed our social support systems. Colleges saw this and stepped up to plate as the trusted middleman, and so we regressed from networks to hierarchies.

This wasn't necessarily bad in and of itself. Hierarchies can be efficient too. Corporations are usually hierarchies, and they are one of our most adaptive social systems.

Things got bad when credentialism surpassed education as the primary function of college. If you want evidence of this, consider the criteria for the US News & World Report college rankings: Peer assessment, student selectivity, faculty resources, graduation and retention rate, financial resources, alumni giving, and graduation rate performance. Notice anything missing?

Are colleges not ranked in order of how much students learn because that would be impossible to measure? Or is it because no one cares?

As the old saying goes, "What gets measured gets done." According to the National Adult Literacy Survey, 71% of college graduates are unable to read proficiently. What exactly is getting done?

Colleges have failed. Credentialism has failed. Hierarchies have failed.

And they haven't just failed abstractly or in general. They've failed me personally.

The good news is that, thanks to the Internet, degrees may be on their way out. The Internet, as the name suggests, is a global network. Everyone is connected to everyone else. I'm connected to every fortune 500 CEO within one degree. So are you. Sounds a lot like pre-WWI society doesn't it?

Recruiters rely on degrees because they are the easiest way to grok what someone is all about. The problem is that relying on credentialism is an act of faith, comparable to closed source eVoting without a paper trail. And we all know what that got us. The Internet is changing this. Every year our lives become more transparent as our words and actions become increasingly digitized and searchable. What credentialism was to the twentieth century, Google will be to the 21st.

In the 20th century your references were something you put on your resume. In the 21st century your references ARE your resume.

Thanks to the Internet I am connected with almost every CEO within two degrees. It's as if we all live in a small village again, where everyone knows everyone and anyone can vouch for you. Hierarchies are no longer necessary, because hyperlinks have subverted them.

As technology multiplies, new social systems and metaphors will emerge as others obsolesce. Only humanity remains untouched.

Smelting Gold In The Twitter Ghetto

2009-03-25T10:34:37-04:00

Twitter is a ghetto, and most individual tweets are worthless. Yet on any given day there are dozens of really interesting conversations worth following.

Even if the value of individual tweets is fleeting, the value of these conversations and contexts is permanent. The problem is that there's no good way to read through past conversations after the fact. For example, during #sesny today there were ten or fifteen really interesting conversations. But of course for every insightful tweet that contributed something to the conversation, there were twenty that were retweets, spam, nonsense, or just offtopic. That's not so bad when you're reading the new tweets in real time, but trying to grok the conversation later it's completely unbearable.

While twitter as a whole is probably best described as a ghetto, there are a ton of thought-provoking conversations being generated that deserve to be preserved and curated — the same as a museum would do for works of art worth saving for future generations.

As an example of what I mean, check out my lens on the Hacking Education conference that took place the other week. That lens still might not be perfect, but it's infinitely easier to read than scrolling through the 65 pages of mostly-garbage tweets that make up the official #hackedu twitter search.

Every event needs a twitter hashtag, and every hashtag needs a page like this. Otherwise the vast amounts of knowledge and culture created daily will be as ephemeral as the tweets themselves.