2017 Year-In-Review

With the way the pace of my life is moving, I feel like I haven’t had much time to reflect on longer term accomplishments and lessons.  So I thought doing an annual review is a great chance for me to reflect on the past year and figure out what I’m doing right and what I’m dong wrong.

For as long as I can remember, I have looked back at my past self and realized just how stupid I was.  Okay, probably better to put it more positively, I look back at myself and realize how little I knew.  And I think that’s a testament to my learning ethos, and it’s something that I want to make sure I continue to achieve.  So I’m hoping that this annual review will help me with becoming less stupid.  Here’s to 2018!


I’m quite proud of my accomplishments in 2017.  I’m not sure the best format to describe these things so I’ve just grouped them into buckets.


I’m quite proud of the work I did at Rubikloud in 2017.  In the past year, I’ve built up a great team of data scientists and we’ve worked together with our incredibly talented product and engineering teams to ship some really cool machine learning products, which are making a huge impact at the businesses of our retail enterprise customers.  This is the second startup that I’ve been a part of and I’m still amazed at how much value a small team of motivated individuals can create in such a short period of time (although not without a ton of hard work and chaos).  We recently raised a $37M Series B round led by Intel Capital to scale out our business, so 2018 is definitely going to be an exciting year!

Technical Blog

I’m quite proud that I’ve been able to keep up posting on my technical blog (although not so much on my main blog as you can see).  I’ve written 8 full-length posts, most of them with some sort of implementation.  This has been my primary method for learning about non-work related technical topics.  This year I finally started getting into more deep learning topics with a particular focus on a bunch of the various flavours of autoencoders.  It’s nice to have an outlet to learn about non-work related topics.  It’s very seldom that the work you do at a job aligns precisely with the topics that you’d like to learn about.  And while one of the main tangible benefits is a public record of my technical prowess, the reason I can keep it up is that I actually enjoy it a lot!  It’s one of those things that I’d keep doing even if I didn’t need to work.  Hopefully this year isn’t too crazy that I can’t keep working on it.


I’ve been taking guitar lessons for a few years now with an amazing teacher, Ryan Carr.  At the end of 2016, we started working toward one of my goals which is to sing and play at the same time.  To give some context, when I started singing I had absolutely no sense of tone.  That is, I couldn’t match pitch, tell if one note was higher than the other, or otherwise have any “natural” talent that would be helpful when singing.  Despite my lack of talent, I kept practicing (which was especially painful for my wife in the beginning because she has much better tonality than me).  It’s quite amazing how far I have progressed.  From not having a sense of tone in the beginning, to now where I have at least some vague sense of what’s going on is quite an accomplishment for me.  It’s hard to describe it because a lot of it is unconscious and through feeling.  For someone like me who is used to consciously understanding things, it was (and still is) a big challenge.  The big leap of faith is that you keep working at it, knowing that you are making progress even if you can’t see it.


I’ve been taking Chinese lessons for the past year or so at this great Chinese school.  Some background: for the longest time I’ve had very basic Mandarin skills.  Growing up, I had heard it at home and half-heartedly learned it at Saturday school when I was younger but never was able to go beyond a very basic vocabulary, which revolved around things you do around the house (eat, sleep etc.).  But after going to this school (1.5 hours every week) for a bit more than a year, I’m quite amazed at how far I have progressed.  From barely being able to read 100 or so words, I can now read stories with a vocabulary of somewhere around 1000 words (as long at they’re the right words).  And even on the speaking front, I’m much more comfortable (barely) carrying on a conversation, whereas when I started I would frequently get stuck unable to find the right words to express myself (listening was never a major problem for me though so long as it was within my vocabulary).  It’s also very gratifying to get external praise.  My wife’s family is telling me that I’ve been improving (they speak Mandarin most of the time at home) and one of the Chinese teachers was jokingly saying that I improved much faster than her expectations as well (understandably, I’d have low expectations of myself if I was her as well!).  Just like guitar, language is something that is learned unconsciously.  At some point, you’re just able to recognize more characters, speak more comfortably, and feel more confident conversing.  Of course, it’s still a lot of conscious work reviewing characters daily, memorizing stories, forcing yourself to speak when you know what you’re going to say is wrong.  It definitely is a bit of a grind but it’s amazing to see how much one can progress in just a year.


Failures are always something that I haven’t had a habit of keeping track of explicitly (actually accomplishments I don’t either unless you count my LinkedIn).  But every successful person that I’ve read about says that it’s of vital importance to reflect on them.  It shouldn’t be hidden and covered up, rather it should be explicitly stated and encouraged so that we can then improve ourselves, a kind of “de-stupification” process.  It’s kind of like finding bugs in software, it’s a behaviour we want to encourage!  Finding more bugs, means fixing more bugs, means making the software better!  So with that, let’s hope this makes my “software” better.


Since 2006 I’ve been managing my own portfolio of stocks.  Although, I’ve done better than the index, I’ve also made a huge number of mistakes.  The 2017 year in particular has been quite challenging because everything looked quite expensive so my trigger finger was getting itchy (a great recipe for disaster).  I’ve know for a long time that activity for activity’s sake yields poor performance but knowing something and applying it are two different things.  A huge advantage that individuals have over actively managed funds or trading groups at investment banks is patience.  Individuals can wait years (since they’re saving for their retirement after all), while actively managed money needs to show results in months.  I recall Charlie Munger mentioning that he sometimes waits years just sitting on cash waiting for the right time to pull the trigger.

Recently, I’ve been dabbling with options in the past year.  In general, leverage is a great way to lose all your money.  However, I’ve been reading about a particular trade that involves long dated call options (LEAPs) that is closer to long-term investing rather than gambling.  Basically, you use the same logic as picking long term investments except with a bit of leverage.  The idea is you just hold the LEAP and roll it over as it comes to expiry to maintain your position in the company.  Anyways, over the past couple of years I’ve had a bunch of successful trades with options always buying when the price of a security is way below some rough calculation of intrinsic value for companies I’ve been pretty confident in.

However, I started venturing out this strategy a bit (again because of my itchy trigger finger) to binary outcome trades (using options or otherwise).  Basically when a stock is undervalued dramatically it’s usually due to some terrible news.  It could be that the news was overblown and the market will come realize this and the price goes back to a reasonable level, or it could be that the doomsday scenario happens and it goes to zero (or actually anything in between).  So I started betting (and notice that I’m using gambling terminology here) on a few of these trades sizing my bets using the Kelly Criterion.  So long as my estimation of outcome is in the right ballpark, if I do a lot of these trades, it should average out to a positive gain.

One of trades is working out (so far), the other I’ve taken a loss.  The one that I took a loss on was Chicago Bridge & Iron Company (CBI).  My reasoning for
going into the stock was not very well thought out.  It’s been a stock that has been much talked about on the stock board I follow and it suddenly had a big drop due to poor economic outlook due to a few factors.  So I made a bet that it would rebound.  However, even at the time it was an ill advised decision because I personally had done very little research into the company.  I didn’t even read the latest annual reports or conference call transcripts, all my information came from the stock board.  This was a big failure on my part.  It’s so easy to get lazy and just buy a stock, but this is exactly what you should be avoiding (especially in a concentrated portfolio like mine).  Thankfully, I didn’t bet a big amount because I was not especially confident in the turn around and the Kelly Criterion sizes the bet smaller.  I should do well not to buy stocks on a whim but only after thorough analysis (at the very least understanding the business model and the current state of a company’s business).

Connecting with People

Being an introvert, I have a harder time than most people connecting with new people.  When meeting someone new, it feels like it requires a ton of energy.  As a result, I frequently just avoid these types of situations because, quite frankly, they are exhausting.  It’s almost as if I have to force myself to be “on” to be able to connect with new people.  It’s a bit hard to explain the feeling if you’re not an introvert.

However, this doesn’t mean it’s not worthwhile.  I think I’ve avoided these situations more in the past year (it waxes and wanes), giving myself excuses such as “I’m tired”, “I’m busy” or worse being apathetic (“I’m not interested in connecting”).  This is a failure on my part.  Of course, it is worthwhile, it just take a bit more effort.  My hypothesis is that like most things, it’s a sort of like a muscle, the more you work on it the better you get.  So this is definitely something that I’m going to consciously work in order to reap the benefits of meeting and connecting with the people around me.

Rationality vs. Relating

One of the things that I definitely need to improve upon is communicating an idea that contradicts with the listener’s point of view.  When discussing an issue, sometimes I have a very clear idea about how to approach the solution but it may contradict, negate, or otherwise mismatch with my listener’s point of view.  My usual strategy is to just explain it rationally.  Sometimes this works, other times not so much.  Most people (myself included) feel slighted when someone brings up an idea that is dissonant with their world view.  It’s difficult to suppress that feeling and rationally look at a problem.  In those cases, sometimes I still push forward with my rational explanation strategy, a failure on my part as a communicator.  I think a better alternative is to understand where the other person is coming from, start from there and then lead them to my solution.  It’s much harder because you actually have to understand the other person.  It’s not natural for me but like the one above, I’m of the belief that the more I practice the better I’m going to get at it.

In a similar vein, when giving feedback to someone, no one likes to hear the feedback straight up.  A contrived but effective gambit is the compliment
sandwich: compliment, feedback, compliment.  The only caveat is that you need to be genuine with the compliments (I’m assuming the feedback is real).  It’s very unnatural but with practice I’m told it becomes much easier (so says my wife who is great at it).  The key here I think is still to work harder at understanding the person, that’s what communicating is about in the end isn’t it?


One of the things that I’ve learned is that good habits can go a long way to getting to where you want to go.  My current model is that most big accomplishments (e.g. learning a language, being an expert in a field etc.) are not a one time big effort but rather a bunch of small, consistent steps in the right direction.  Working at something day in and day out is pretty hard!  It’s a grind sometimes.  So habits make this process much easier.  They’re the default action you take, which means that you are much more likely to do them, automatically driving you in the right direction.  Anyways, here are some good habits that I’ve been picking up.

Best Hour of the Day

Following Charlie Munger’s advice, I give the best hour of the day to myself.  Now that I have a more regular schedule going to bed at a normal time (thanks to my wife), I can wake up relatively early.  At this time before I get ready for work, I work on my own projects.  For the past couple of years, this has been my technical blog (see above).  Of course, I don’t do this every day, sometimes I go to bed late and sleep in, other times I have early morning meetings but being a habit, it’s my default behavior.  Definitely something I want to keep up going forward.

Filling in Downtime

I’ve been learning Chinese and one my strategies is to use spaced repetition on a flashcard app on my phone.  Every time I encounter a new word, I add it to the deck and everyday when I have some downtime (usually during breakfast or waiting for the subway) I work though the day’s cards.  It actually is not the most efficient way to learn because I’m just casually going through the cards so I’m not using 100% of my brain power.  However, just by the fact that I’m doing it nearly daily (and with spaced repetition), it helps over a long period of time.  The most important part being a default behavior.

In a similar vein, I’ve also been trying to practice singing during miscellaneous tasks.  Now that I can tell (mostly) if I’m singing something wrong, I practice singing while cleaning, showering, waiting for the elevator etc.  It’s not focused practice but similar to the flashcards, it definitely helps supplement the dedicated practice sessions.  Just have more practice using my voice has been a big reason why I can progress with relatively few dedicated practice sessions (of course if I had more time I would want to have more dedicated practice sessions).


Sadly this year I haven’t spent much time reading books and actually two of the books in the list below I read over the Christmas break.  Definitely something I want to change for next year.  Here’s a list of the books I read this year and some commentary.

  • Intelligent Fanatics: How Great Leaders Build Sustainable Businesses (Ian Cassel and Sean Iddings): Nice short book describing some great lesser-known business leaders (e.g. Les Schwab, Simon Marks, John Patterson, Sol Price) and how they built their businesses.  The book doesn’t spend too much time on each leader and the story telling is a bit factual for my liking.  It’s an easy read and enjoyable but nothing too insightful.
  • How To Read A Book (Mortimer J. Adler): This is a serious book!  Originally published in 1940, it’s about how to get the most out of reading by applying active reading strategies.  I actually didn’t finish the entire book, got through roughly 3/4 of it.  But my main take away, that I’ve started doing, is to always read with a pencil in my hand!  Marking up books is a great way to help retain knowledge (even though it feels a bit wrong marking up a pretty book, definitely picked up during school when the teacher would scold me for this) and it allows you to go back and re-read the important parts that you missed.  Many times in the past, I’ve passively read something and not remembered it.  Although that can be useful in its own way, wouldn’t it be better if you remembered it?  Anyways, if you want to pick up some strategies for reading, I’d definitely recommend picking this book up.
  • Rational Optimist (Matt Ridley): This book’s thesis is very simple: despite all the terrible things going on in world today, it’s getting better!  He goes through a lot of history citing evidence of how things are getting better with the main idea that trade and specialization in human societies (so called “collective intelligence”) is the reason why we have been rapidly advancing.  And further the rapid advancements are the things that are solving humanity’s problems (e.g. over population, global warming etc.).  He urges us to continue trading and specializing and reject the idea that we must slow development down because the rapid pace of development is exactly what allowed us to solve the problems in the first place!  It’s a highly convincing book with a lot of historical evidence, definitely recommend it.
  • High Output Management (Andrew S. Grove): This is a book I just got before the break from my colleague.  The author is the late and former CEO of Intel.  It’s probably one of the most useful management books that I’ve read in a while because it’s not just business book fluff.  He actually goes into specific tactics that he has applied at Intel ranging from a model of how to think about your business “output”, to specifics on how to conduct 1-on-1s and meetings, to what a manager’s main duties are (“training” and “motivation”).  He peppers the book with lots of examples and specific situations.  It’s a great read for any manager.  It’s also short, so it’s easy to go back and re-read certain sections (or look up the notes you made in the margin, thanks “How To Read A Book”!).
  • The Tao of Charlie Munger (David Clark): This book is just a bunch of quotes by Charlie Munger.  It’s not very useful and I wouldn’t recommend it because if you’re a fan of Charlie Munger you would have read all of these quotes, if you’re not a fan, it’s definitely not a good way to learn about him.  I’m just a sucker for Charlie, so I’ll buy any book that might let me learn more about him.
  • A Man for All Markets (Edward O. Thorp): This is an autobiography by the great Ed Thorpe.  He’s not as well known as Claude Shannon but he’s made a huge dent in the practical application of mathematics in the background.  He’s infamous for working out the math to beat blackjack, he also ran a statistical arbitrage hedge fund in the 70s/80s and independently (probably prior) discovered the Black-Scholes option pricing model.  His returns at his fund were insane topping more than 20% annually for more than two decades, mind you with relatively little risk.  The interesting part of his thinking is that he was never interested in the money in both cases, the problem is what he enjoyed.  The application of the mathematics in both these domains were really just about experimenting to see if his math really worked.  This is evidenced by him writing books, not long after he validated the ideas, describing in detail all his methods.  He’s incredibly interesting person, definitely genius caliber.  It’s also an easy read but you can tell his story telling is not as polished as some other biographers.  This actually adds a bit of authenticity since he probably did write most of it himself.  Definitely recommend!
  • Fearless Salary Negotiation: A Step-By-step Guide to Getting Paid What You’re Worth (Josh Doody): This is a book I saw recommended by Patrick McKenzie a while back, so I bought it and had the e-book sitting on my Kindle for a while.  I decided to do a quick read of it over the break and I was pleasantly surprised.  All the information is the book is kind of “obvious” but also incredibly useful.  Salary negotiation is a hard topic because most employees will have only done it a handful of times while the HR people do it daily.  The information in this book balances this information asymmetry giving details on how the whole thing works.  It goes further to give you a step-by-step guide on how to approach the entire process, either negotiating a salary for a new job or trying to get a raise from an existing one.  Definitely recommend, it’s a short read and worth the probable increase to your salary by applying these techniques.

The Coming Year

2018 is going to be a busy year for me.  There’s still a ton of work to do at Rubikloud, lots of things that I still want to learn in machine learning, and a few interesting side opportunities that might pop up for me as the year progresses.  It’s a bit daunting but also exciting.  Here’s to good fortune in 2018!

Two Steps Forward, One Step Back

Learning is one of those funny things that we never learn how to do — we just do it.  As children, we somehow learn to read, do math, and play sports all without thinking too much about how to learn or what it feels like to learn.  Learning is just a natural thing when you’re growing up.

As a child, you’re given lots of guidance, lots of opportunities and most importantly lots of time to learn.  Contrast this with learning as an adult:

  • You rarely have explicit guidance — you’re an adult, teachers are for kids!
  • You have fewer opportunities to learn — you’re an adult, you’re supposed to be able to identify and seize opportunities for yourself!
  • You have so much less time to learn — you’re an adult with a full schedule every day!

Which basically leads me to this fact: it’s much harder to learn as an adult!  As opposed to not being as capable to learn as an adult, which I think is a myth.

My situation was a bit different from most because I extended my schooling with graduate school.  Although everyone there was still an adult, there was still a lot of guidance, (self-directed) opportunities, and most importantly time to learn.  Now that I’m working in the “real world”, I’m finding myself bumping up against all of these challenges to learning as an adult.

But I digress, what I really wanted to talk about in this post was a phenomenon with learning that I have recently been thinking a lot about.  It’s the idea of “two steps forward, one step back” in learning.  As you learn a new skill in a subject area, at times you feel like you’re making good progress (“two steps forward”), but quickly realize you’ve hit some kind of wall and your progress grinds to a halt (“one step back”).  Incrementally you make progress but the process is grueling.

When in school, I rarely thought about this process.  You just keep moving forward, course after course, test after test.  It’s all such a whirlwind that you rarely have time to think about the actual process of learning.  More importantly, since it’s your full-time job to learn, this phenomenon may not be as apparent.  However, when your full-time job is actually your full-time job, and learning is just a hobby, you start to think about it more.

I most recently experienced this “two steps forward, one step back” phenomenon in three very different areas of learning:

  • Machine Learning: I’ve been trying to stay on top of (or catch up on?) all the recent developments in machine learning.  I’ve been pretty good at posting on my technical blog trying to explain concepts and ideas that I’ve learned.  Part of this process is really digging into the details of a particular topic including the math, the implementation and the intuition.  But the more I dug into a topic, the more apparent it was that my knowledge of the area was relatively superficial.  The math, the implementation, and even the concepts went so much deeper than I could have imagined.  Two steps forward, one step back.
  • Music: I’ve been taking guitar lessons with an amazing teacher for a while now, and recently just started vocal lessons with him too.  When I started out I was pretty terrible, I mean really terrible.  I struggled with the most fundamental skill of simply knowing if I was matching pitch, never mind actually trying to match pitch.  The analogy I like to use is that I was completely in the dark.  As I progressed, I gradually got better at telling if I was on pitch and correspondingly matching pitch.  I went from completely blind to having just really bad eye sight.  However, as I was able to “see” more, the distance between two notes that initially felt so small, felt like they had much more space in between.  But I soon after came the realization that there was yet another layer of detail that I was missing.  Two steps forward, one step back.
  • Chinese: I recently decided to take Mandarin lessons at a great Chinese school in the evening after work.  Part of my inspiration came from Lee Kuan Yew, also of Chinese descent, who learned Mandarin in his thirties.  So I thought: if one of the greatest minds of the last century could do it, why couldn’t I?  Obviously, it was a lot harder than I anticipated.  One particular thing I was proud of early on, was how many Chinese characters I learned in a short period.  There is this great little Chinese learning app called Plecko that has a flashcard program built-in.  Every time I hear a new word, I add it to the program.  At one point, I hit 1000 flashcard entries and it felt like a big achievement!  The next lesson, I got put back in my place.  I realized just how little I knew.  In addition to a flurry of new vocabulary that I had never seen, there was a whole slew of new grammatical structures and phrases that were foreign to me.  Two steps forward, one step back.

This idea of making progress, but feel like you’re not, is nothing new.  It’s part of learning, especially in a time-constrained way where you are bound to feel it more.  As soon as I realized this pattern, I started to feel a bit better about it.  Learning is rarely easy, there’s always a grind, especially if you’re trying to do it in a short period of time.  Among the many quotes available, here’s one by Confucius:

Real knowledge is to know the extent of one’s ignorance.
— Confucius

So the fact that I feel like I’m taking one step back is actually a sign of gaining more knowledge.  Life is full of these little paradoxes.  Now if someone would just tell me where I’m going to die so I won’t go there.

Memorization and Learning

It’s funny how after so many years of being in school, there’s always something new to learn — especially about learning.  You would think that after more than 20 years of schooling somehow I would have learned every tip and trick, but nope, my ignorance seems to know no bounds.  Recently I rediscovered a technique that has greatly improved my learning: memorization.

Memorization and I have a very complicated history.  When I was but a wee lad, I was frequently forced to learn things using memorization such as: Chinese, piano and of course, the usual memorization-laden subjects in school: history, geography, biology, etc. Even though I was not particularly adept at memorizing things, I will say that I got through these formative years with relatively high grades by having a lot of assiduity (as Charlie Munger likes to say “sitting on your ass until you do it”) but these experiences left a sour taste in my mouth for memorization.  The subjects I much greatly preferred required less rote memorization and more thinking (just my kind of problems!).  Subjects like math, computers, English 1, physics and chemistry2 were really my cup of tea.  You just have to understand a single concept or technique and all of a sudden you can answer hundreds of questions, talk about efficiency!  Of course, it still requires a lot of practice but it was a huge departure from trying to memorize a list of seemingly random facts.

As with many things in life, I feel like I’ve come full circle.  Recently, I’ve started taking Chinese classes again as well as music lessons (mostly guitar, a bit of vocal).  One thing that is crystal clear is that memorization is hugely beneficial… in certain contexts.  For example, my Chinese speaking has always been lacking.  I was a shy kid and didn’t really make much of an effort to speak in Chinese.  However by memorizing and reciting the textbook lessons, character by character, I’ve been told my Chinese has improved quite a bit.  There’s something about actually moving your mouth, activating your vocal chords, and having sound come out that gets your neurons connecting properly.  Another more obvious explanation: if I wanted to get better at speaking, I should practice speaking more!  However, I will point out that the opportunities to memorize a piece of text and say it out loud are much more abundant than speaking to an actual human.  In any case, a very useful application of memorization.

The other domain where I’ve found memorization very useful is in music.  When I was young, I thought music was about memorizing the correct order of notes to play.  In my case, it was keys on a piano.  I just had to figure out the right order and timing and I was done!  Unfortunately, piano was not conducive to my learning style at the time and I quit after only a few months.  However, while learning guitar as an adult, I realized that playing an instrument was so much more than just hitting the right notes.  There are so many more abstract, higher level concepts like the subtleties of rhythm or the attack on the guitar.  If you’re still thinking about which note to play, you’re definitely not thinking about these higher level concepts.  Similarly, when trying to do two things at once, like singing and playing, there are so many things going on that you can’t be thinking about all of them at once.  Enter memorization.  By memorizing the “easy” parts, like the notes and lyrics, you can start focusing on the important parts of playing.  In the case of a B.B. King solo, getting every little lick to sound just like the king, signature vibrato and all.  Or when singing and playing, knowing the lyrics down pat, and playing the notes without thinking so you can concentrate on actually getting them to work together.  Again, another important application of memorization.

My conclusion (which in hindsight is quite obvious) is that memorization has (at least) two important purposes.  First, use memorization to physically enable you to learn something.  That is, getting your muscle memory working and all the related neurons in your brain.  For example, speaking a language, or playing a musical instrument.  There’s really no other way around learning these things except by rote practice — that’s how our bodies work.  Second, use memorization to be able to ignore the “easy” parts so you can concentrate on higher level parts.  For example, memorizing the lyrics to a song so you can focus on the actual singing, or memorizing the multiplication table so you can focus on algebra.  In this application, memorization is a means to end to learn some higher level concept that is too difficult if you don’t have the basics down.

I’m quite excited by my new (obvious) discovery on the usefulness of memorization.  Although I probably should have come to this conclusion sooner, I’m quite happy to be a little bit less ignorant.  One of my goals in life is to be less ignorant, an ignorance removing machine if you will3.  After all, you just have to be a bit better at ignorance removal and you won’t get eaten by a bear.


  1. Essay writing requires coming up with a structured argument/flow. []
  2. We really didn’t do a lot of organic chemistry in high school so most of the subjects was explaining concepts or doing a bit of math. []
  3. That’s just a cute way of saying that I like to learn. []

Comments Off on Memorization and Learning

Categories Posts

Consistency and Commitment

If you’ve read my last post, a good model for human behavior is that of a moist robot.  We’re nothing more than mere automatons (or automata) except that we’re wet and squishy.  This post is about another common behavior among us squishy automatons: consistency and commitment.

Influence by Cialdini describes this tendency in detail as one of the big six principles for how others manipulate our squishy brains into doing things that we otherwise wouldn’t want to do.  The funny thing about these psychological tendencies is that we’re mostly blind to them — psychology researchers included!  A recent blog post by Andrew Gelman discusses how scientific research in (certain parts of) the psychology community is quite lacking (particularly from a statistical point of view):

The short story is that Cuddy, Norton, and Fiske made a bunch of data errors—which is too bad, but such things happen—and then when the errors were pointed out to them, they refused to reconsider anything. Their substantive theory is so open-ended that it can explain just about any result, any interaction in any direction.

The interesting thing is that one of the main authors, Fiske, is a well-respected researcher in her field.  Gelman is commenting about an open letter she sent criticizing social media and all the negative comments her research has received.  She asserts that the comments should be done in a “moderated” (read: research journal) where an editor can filter inappropriate comments.  While in theory it sounds good, practically, these psychology researchers are falling for consistency and commitment fallacy.  They believe that (implicitly) just because their research is published and they are well-respected then it has to be good (or at least not bad) even in the face of obvious errors, from Gelman’s post (emphasis mine):

She’s implicitly following what I’ve sometimes called the research incumbency rule: that, once an article is published in some approved venue, it should be taken as truth. I’ve written elsewhere on my problems with this attitude—in short, (a) many published papers are clearly in error, which can often be seen just by internal examination of the claims and which becomes even clearer following unsuccessful replication, and (b) publication itself is such a crapshoot that it’s a statistical error to draw a bright line between published and unpublished work.

Ironically from a psychological point of view, it all makes sense.  If your worldview is different from reality, what’s easier to believe: the nice rose-tinted glasses of your existing worldview or the unpleasant scent of cognitive dissonance?  From Gelman’s post:

If you’d been deeply invested in the old system, it must be pretty upsetting to think about change. Fiske is in the position of someone who owns stock in a failing enterprise, so no wonder she wants to talk it up. The analogy’s not perfect, though, because there’s no one for her to sell her shares to. What Fiske should really do is cut her losses, admit that she and her colleagues were making a lot of mistakes, and move on. She’s got tenure and she’s got the keys to PPNAS, so she could do it. Short term, though, I guess it’s a lot more comfortable for her to rant about replication terrorists and all that.

Funny enough, this is not the only ironic situation that I’ve come across.  If you’ve hung around enough programmers and data scientists, I won’t have to tell you that they’re (at least they say they are) all about “logical reasoning” and “rational thought”.   But from what I’ve seen, they’re some of the most wet and squishy amongst us all.  Just ask them what’s wrong with X, where X is any of the most popular programming languages, just be ready to sit through a half an hour rant that is heavily rooted in opinion over facts.

There are worse things than highly opinionated people though.  The worst case scenario is where consistency and commitment tendencies block the person from seeing their own mistakes, from Gelman’s post:

We learn from our mistakes, but only if we recognize that they are mistakes. Debugging is a collaborative process. If you approve some code and I find a bug in it, I’m not an adversary, I’m a collaborator. If you try to paint me as an “adversary” in order to avoid having to correct the bug, that’s your problem.

Being open to collaborators who point out our flaws (hopefully in a congenial way) is really the first step in learning.  Think about all the teachers you’ve had, probably half of the learning comes from them pointing out mistakes (the other half from them teaching you a better way of solving the problem).  Whether you’re a toddler or a famous researcher, everyone needs help finding their mistakes, and more importantly everyone can learn from them.  It’s easier said than done though because, after all, we’re just moist robots.

Moist Robots

Scott Adams has this really memorable term for how to think about people: moist robots.  Moist because — well, we’re wet and squishy.  Robots because there are certain predictable behaviors that we repeat.  In situations like A, most people do f(A).  In situation; in situations like B, most people will do f(B).  Sounds very robot-like to me1.

Now I’m sure you can remember a time where this rings true, everyone has that friend who thinks their going to win big at the casino despite what hundreds of years of math suggest.  This is great example of irrational behavior that has been programmed (in one way or another) into many of us moist robots.  Most of the time our squishy brains are great at detecting these problems but only when it’s not about us.  It’s easy to see your friend has no idea about probability, but much harder to see why you’re such a sucker for Instagram.

As you read this sentence, you can imagine how this story can play out in any number of ways.  Here’s a very common story, you see someone who is super successful, perhaps someone famous you saw on the interweb.  Naturally, you wonder how you can become super successful.  Your parents will tell you that all you need is hard work.  You’re not so sure since they also told you that the Easter Bunny was real, so you also read Ms. Super Successful’s biography.  The book tells you it was her passion that drove her to eventually become CEO of company X.  You’re mostly convinced but to be sure you listen to her commencement speech and any interviews you can get a hold of, and they tell you the same thing: hard work and passion got Ms. Super Successful to where she is, and so can you2!

But let’s not forget about one very important thing: we’re all still moist robots.  A very common pattern (or cognitive bias as the psychologists call it) with moist robots is that of self-serving bias: we have the tendency to perceive oneself in an overly favorable manner.  So when Mrs. Super Successful is attributing her success to something, she overemphasizes the aspects that she has control over, namely: hard work and passion.  Never mind the fact that her parents were in the 99th percentile of wealth, had connections with the trustees at an Ivy League school, and had an abundance of opportunities to learn the skills she needed to be successful — all she really needed was hard work and passion.  Which leads to another pattern called survivorship bias where us moist robots mistakenly draw conclusions about what worked based on who survived (e.g. successful people) instead of the entire population (e.g. all people, successes and failures).  If you looked at the people who worked hard and have passion, you would see a surprising number of moist robots who weren’t successful3.

And so a successful moist robot tells another young moist robot, how to be successful and we see a predictable pattern of behavior again (you’re a young moist robot in this situation).  You have multiple authority figures (parents, teachers, successful people) telling you to one thing (authority bias), these people are generally very likable (liking bias), and you see that all successful people have worked hard and followed their passion (suvivorship bias and bias from mere association), it’s no wonder we’re programmed to think that hard work and passion alone are enough to be successful.  We’re just moist robots after all.

Having said all of that, I’m still a big fan of working hard and following your passion.  It’s a pretty good program as far as us moist robots go.  I would add another crucial subroutine though: luck (or more aptly randomness).  Everything that happens is so dependent on our surrounding environment, which we rarely have control over, hence luck.  If you’re not factoring in luck then you’re not factoring in reality.  The trick is that your program should seek out and maximize the situations where luck is on your side and correspondingly move out of positions where it isn’t.  After all, what good is a (moist) robot if you can’t program it?


  1. If you want to learn more about moist robots, then you might be interested in getting Scott Adams’ book: How to Fail at Almost Everything and Still Win Big: Kind of the Story of My Life []
  2. Phrase intended, check out Stephen Colbert’s book []
  3. Although, you rarely see a successful person who doesn’t work hard and have passion.  It’s kind of a necessary but not sufficient condition []