24:28
is you know okay so there is a severe
error in reasoning and that that you
often hear by people giving you
empirical data and telling you for
example that were idiots to worry about Ebola that killed two Americans when more people slept with Kim Kardashian that year okay so for example and that
was effectively the number they said
that more people slept with Kim Kardashian in 2014 or worried about Ebola then died of Ebola for example
right so now and then sometimes you see
numbers like these and this is the kind
of thing that that if you read something
called the New York Times and you still
read that nonsense this kind of stuff
you read in it which is factually right
but bogus okay well and the kind of
thing that teaches and psychology
department that so many Americans die
from eating too many hamburgers smoking
too much alcohol now let’s think about
it in terms of tails if I tell you or
let’s say Steven Pinker give the number
that 3,000 Americans died and their bathtub every year okay 3,000 people whereas three or two have died of Ebola now let’s play
thought experiment if you read in the
papers if you go you know to Mars and
come back
and then read on you know on Google News that two billion people have died what’s more likely to have killed them diabetes
obesity is falling the rest stop
sleeping with Kim Kardashian or Ebola
Ebola there we go so you cannot compare
rule number one
thou shalt not compare a multiplicative
fat-tailed process in extremists and in
the sub-exponential class to a thin tail
process that has what we call Chernoff
bounds okay and that is in totally for
mediocre stuff simply because of the
catastrophe principle okay and then we
know for example that is very cheap to
protect yourself from Ebola you see the
probability that people dying from
smoking okay it’s multiplied by 10 next
year is 1 to the 10 – into 10 – 30 the
probability that that the the rate of
people who died from Ebola through
triples is not vastly higher you see so
you cannot compare processes thin tail
to fat tails are not comparable so this
is not empiricism this is called naive
ampere system it’s actually worse okay
so you cannot compare two processes like these by saying we worry too much about Ebola in fact we were too much about
diabetes and too little valley bola so
that’s one error of reasoning that comes
from not understanding fat tails now let
me show you the law of large numbers
everything you’ve learned in statistics
is based on a well functioning of a law
of large numbers no all right so in
other words you that as you add
observation this mean that you observe
would be very stable no and the rate is
square root of 2 square root of n number
of observations you agree all right
this is what you have on the left now on
the right for fat tail process the mean
exists but it takes much longer to
observe it
28:10
much longer my only my anymore
28:13
observations okay
Finite players play to beat the people around them. Infinite players play to be better than themselves.
I could’ve been reading an article analyzing Roger Federer and Rafael Nadal. Except I wasn’t.
ENTER: Simon Sinek
I was listening to Sinek (he was talking at Google) use game theory to describe the kinds of ‘games’ companies engage in: finite games, where the objective is to win (or cause all other participants to stop playing), or infinite games, where the objective is to continue the game as long as possible.
I first came across this line a few months ago. It has stayed with me since then. Suddenly, companies and leaders were falling into either one of these categories for me. If you are an entrepreneur, quite likely Sinek’s line speaks to you too. And you’ll start placing companies into one of these categories. I hope you will, at least.
Take Sinek’s examples as a starting point. Microsoft (under Ballmer’s leadership) executives used to tout how much better their products were compared to Apple.Apple, at the same time, however talked more about end results they were working to achieve. Microsoft, says Sinek, was playing a finite game, whereas Apple was onto an infinite game of self-improvement. Which approach is better did you ask? The business results of both Microsoft and Apple from the time speak volumes about the merits of each approach.
ENTER: Paul Graham
Understanding finite vs infinite games isn’t merely an exercise in the abstract. There is more. Lace it with investor and writer, Paul Graham’s mental model of good vs bad test, and I’d argue that we have a conceptual framework that is critical for business leaders anywhere.
Graham’s recent post about unlearning is a masterclass in understanding the merits of startup life relative to life at institutions like schools or large corporations.
The most damaging thing you learned in school wasn’t something you learned in any specific class. It was learning to get good grades.
Graham’s core idea is that it’s important to know whether you’re spending energy solving a challenge that’s directly connected to reality (studying for a good test), or a challenge that isn’t, usually imposed by an authority (bad test). Recognizing and destroying bad mental models may be even more valuable than adding new ones.
What is the litmus test for a good or bad test? Bad tests are inherently ‘hackable,’ meaning that with clever and directed energy, we can often find a shortcut to ‘scoring well’ on that test and acquiring a label of success without fully solving the underlying challenge. Good tests, on the other hand, are ‘unhackable,’ so we can either succeed at solving the challenge to a varying extent, or fail entirely.
A test in a class is supposed to measure not just how well you did on that particular test, but how much you learned in the class.
Good tests like curing cancer, making education free for anyone, or even winning a tennis match are inherently more engaging because the best scientist, entrepreneur, or player will typically win in each respective scenario.
ENTER: Elon Musk
Sinek and Graham’s models seemed familiar the first time I encountered them, and I wondered why. In filing away these new mental models, I was reminded of their neighbor in the idea world: thinking from first principles, especially as popularized by Elon Musk:
Physics teaches you to reason from first principles rather than by analogy. So I said, okay, let’s look at the first principles. What is a rocket made of? Aerospace-grade aluminum alloys, plus some titanium, copper, and carbon fiber. Then I asked, what is the value of those materials on the commodity market? It turned out that the materials cost of a rocket was around two percent of the typical price.
SpaceX went on to cut the cost of rocket launch by ~90%, while still making a profit.
Why didn’t the giant aerospace incumbents figure this out first?! In my view, the incumbents were busy playing finite games against their competitors, hacking bad tests, and thinking derivatively from their last quarterly results, rather than from first principles. Textbook opportunity for disruption.
Sinek + Graham + Musk For the Win
Good tests map beautifully to infinite games and first principles thinking. All three mental models seem to reinforce a simple message: think like a scientist.
Infinite leaders, says Sinek, filter decisions first through the unchanging values of a company. And only then, factor in the company’s dynamic interests. This may result in sub-optimal single decisions and failures along the way. However, over the years, the long string of decisions strung together will be more cohesive and, therefore, valuable (assuming the company’s values are well set up).
The political environment of a classroom, or large company, is often set up to reward those that hack bad tests and finite games, since the isolated outcome looks favorable. We don’t consider how that outcome will eventually be strung together with other outcomes in order to fully connect with reality.
The book What Have You Changed Your Mind About? chronicles painful realizations by experts playing a finite game in their area of expertise. In it, a successful hedge fund manager, Nassim Taleb (author of the excellent book Antifragile), talks of how he lost faith in probability as a guiding light for making decisions.
Good tests, infinite games and first principles thinking aren’t for everyone. For others, it’s the only way to go.
Helpfully, and devastatingly, startups afford little-to-no buffer from the real world. The company either solves the challenge, or dies. This kind of instant feedback and intolerance for ‘hacks’ forces infinite game leadership at startups – painful in the short term, but ultimately more rewarding for everyone involved. And the true test of an entrepreneurial leader? As Satya Nadella said, while transitioning Microsoft from a finite game to an infinite game, leaders must find the rose petals in a field of S#$@.
What bad tests or finite games are you putting energy into? Where have you applied first principles thinking?
Oct.31 — Nassim Nicholas Taleb, scientific advisor at Universa Investments, discusses the factors causing global fragility, hidden liabilities in global markets, and what he sees as safe trades in the current market. He speaks with Bloomberg’s Erik Schatzker on “Bloomberg Markets.”