Professor M

2020-12-29 12:01:02 Insane professor – Part 2

I could say that I learned the quick-and-dirty 1/0 method for terminating scripts from a Ph.D. classmate who picked it up from quant portfolio managers researching algorithmic trading strategies at multi-billion-dollar hedge funds. But this would be an argument from authority. And we’ve talked about before how such arguments are lame because they fail to address the why.

What’s the root cause of disagreement between the smart, skillful programmers in that chat and the Insane Professor? (By the way, I’m sort-of-seriously (but not really) considering changing the channel’s name from Professor M to Insane Professor :)

The programmers and the Insane Professor are in the business of selling different things. The programmers sell code and, for that reason, want the code to be perfect. The Insane Professor sells ideas, which he tests using code. In other words, coding is a means to an end for me, while for the programmers, it is an end in itself.

That’s why my satisfaction from deliberately-not-caring-about-the-code-being-perfect infuriates those for who code is the raison d'être.

Insane professor – Part 1 Open / Comment

2020-12-28 12:01:01 Insane professor – Part 1

If you’re doing something innocuous and this makes others angry, you might be on to something profound.

Two months ago, I forwarded my entry “Simplicity is the ultimate sophistication” (about dividing by zero to terminate a script’s execution) to the admin of Data Science by ODS.ai. It’s a great channel, by the way; you should subscribe if you’re in this field.

The admin forwarded this entry to the channel’s 40K+ subscribers. To say that it created a stir is to say nothing. The entry became the most discussed one in the channel’s chat. And when I say discussed I don’t mean praised. I mean—criticized.

My favorite comment is this one:
— I almost considered leaving the channel when I saw this message. Terrible advice. INSANE PROFESSOR. Is this a meme or what? Completely absurd.

How do I reconcile these reactions with my conviction that this has been one of the most insightful entries in my channel?

Wait for Part 2. Coming soon. Open / Comment

2020-12-21 12:00:59 Where does the truth come from?

Imagine you observed a Nobel Prize Winner abstain from dinner to enhance the mental sharpness needed to deliver an after-dinner lecture. Is this behavior worth emulating?

Two thoughts. First, be cautious when using a sample of size one to guide your decision-making. Billions of samples of that size collectively point in all possible directions. Collect more observations; don’t just settle on one data point.

Second, if your one data point has the Nobel Prize, it doesn’t matter. Nobel Laureates aren’t immune to mistakes even when it comes to their areas of expertise, let alone when they comment on topics outside of their realm. It’s flawed to insist that something must be true just because an eminent person said so. An appeal to authority is a cheap ruse often deployed amid failure to understand why a person of eminence said so.

So how do we make progress in discovering the truth? Two paths: Empirics and Theory.

Empiricists would ideally conduct a randomized controlled trial to investigate the causal impact of food input on intellectual output. In the absence of that, they would collect data from gazillion people on their eating habits and life outcomes. Afterward, they would use a clever identification trick to circumvent the issue of correlation not implying causation.

Theorists would be satisfied with why an eminent person said so. The Nobel Prize Winner did explain that digestion results in the outflow of blood from the brain, impeding one’s ability to be at their intellectual best. Everyone is familiar with this feeling: We have all been drowsy after a heavy meal. And behind this feeling is a solid theory, confirmed in countless physiological studies showing how the human body works.

When empirical work overwhelmingly confirms a theory’s predictions, it becomes a good theory. You may even choose to trust the not-directly-tested predictions of that theory. Not because you respect the person who made that suggestion, but because you understand the idea behind it.

Aristotle’s theory was that the heart was the seat of intelligence, and the brain was a cooling mechanism for the blood. His other theories have stood better against the test of empirical times. Open / Comment

2020-12-11 12:01:02 What do Nobel Laureates have for dinner?

During my Ph.D. studies at Harvard, I attended a group dinner with the famous financial economist Robert Merton. In 1997, together with Myron Scholes, Merton won the Nobel Prize for a method to determine the value of options and other derivatives.

Dinner, first. Merton’s one-hour speech and Q&A, second. Ten years later, I recall at best 5% of his rich-in-detail speech. However, I still remember 100% of the food he consumed that night. What did he have?

Nothing. Not a morsel of anything. Not even a piece of bread or salad.

“Have you eaten already?” I asked. He answered: “No, I haven’t.” Eating launches digestion, he explained, and some other physiological processes which inevitably result in the outflow of blood from the brain, impeding one’s ability to be at their intellectual best. For that reason, he chose to delay eating until after he was through with the speech and Q&A.

If I hadn’t eaten anything that night either, my long-term memory might have captured a higher fraction of his talk. Open / Comment

2020-11-16 12:01:02 Crowdsourcing Telegram channels

Instant messaging apps connect their users into networks. The more people have an app installed on their smartphones, the more value the app may give to each user. As such, incentives to spread the word about an app arise. Useful apps may take off even with zero marketing budgets (like Telegram allegedly did).

But Telegram is also a two-sided network (see an earlier entry for more details). On the one side are users who text and read, like all of us here. On the other side are content creators who run channels, like yours truly. More users stimulate more and better channels; more and better channels attract more users. And the infinite feedback loop continues.

To make Telegram’s two-sided network stronger, I want to spread the word about the channels my subscribers could find interesting. And I want to crowdsource this information.

To recommend a channel, please send me a link to it in a personal message. What other English language channels do you follow?

If some are worth-recommending because of their potential appeal to the intellectual, creative, and hard-working audience gathered here, please text me the link. If some are not worth-recommending, maybe you should unsubscribe from them :)

As I go through these messages and review the channels, I’ll be recommending them here on an ongoing basis. Open / Comment

2020-11-12 12:01:01 Two-sided markets and networks

A two-sided market features two distinct types of participants, and the involvement of each is necessary for the market to function effectively.

The business of debit and credit card payments, operated by American Express, MasterCard, and Visa, is a traditional two-sided market. For the payment instrument to be useful, on the one hand, consumers must carry a card in their wallet to be ready to use it at checkout. On the other hand, merchants must have terminals in place to accept a card instead of cash.

The more consumers and merchants use a specific card, the more willing are others to adopt this card. A network effect! In this case, it’s a two-sided network.

To encourage the adoption of payment cards by consumers and propel the network forward, banks and payment providers may choose to issue cards at no cost. With so many consumers enticed to sign up for that card, merchants would be happy to tap into this demand, obtain a terminal, and pay the processing fee for each transaction.

Two-sided markets are everywhere. Facebook is a two-sided market with users (not paying for the product with money) and advertisers. So is your local newspaper distributed for free. So is eBay, and Alibaba, and dating platforms. So are shopping malls.

A previous entry introduced my research paper, which, among other things, emphasized the importance of anchor tenants in shopping malls. An anchor tenant (i.e., a large store) shutting down makes the mall far less attractive to customers. Shopping is better elsewhere. This may launch a domino effect and push other stores to abandon this mall. Numerous dead malls illustrate that two-sided networks need care.

We engage in various networks (both regular and two-sided) all the time. Do what you can to make them stronger. Others will benefit, and so will you. Open / Comment

2020-11-07 01:05:11 Coding and kettlebells

Kettlebell sport is about endurance and strength. In competitions, athletes have 10 minutes for as many reps as possible without placing the kettlebell on the ground. When executing a one-arm lift, such as the snatch, you can change arms. Once.

When training solo, it’s hard to keep track of reps. For best performance, you want to occupy your mind with feeling the body, the technique, the breathing—and not with numbers. Imagine running and counting steps to track the distance? Neither fun nor useful.

Previously, playing squash led to my learning Python techniques to interact with webpages. I used Python to reserve a court immediately once it became available online.

What’s the new frontier? Learn to analyze video and count the numbers of reps algorithmically.

P.S. For info on kettlebells, check out the 14-time world champion Ivan Denisov (YT, IG, TG). The visuals are helpful even if you don’t speak Russian. But he does plan on producing more content in English. Open / Comment

2020-10-29 11:01:02 The Endowment Effect

An ordinary mug weighs around 400 grams. This measurement doesn’t change whether it belongs to you or someone else. The weight is an object’s fundamental property. But what about an object’s value?

In a classic behavioral economics experiment, the experimenter shows an ordinary mug to the participants and asks half of them to imagine that they own it. “At what price do you value this mug?” the experimenter asks, “What’s the lowest price at which you are ready to sell it?”

The other half, who did not get an imaginary mug, answer a similar question: “At what price do you value this mug? What’s the highest price at which you’d be willing to purchase it?”

Amazingly, the (randomly assigned) owners assign a significantly higher value to the mug. The Nobel Prize-winning economist Richard Thaler coined this phenomenon the endowment effect.

Reminiscent of the sunk cost fallacy (see an earlier entry about it), the endowment effect depends only on the fact of possession and not on how much it took to obtain an object. If you feel reluctant to sell something because of how much effort and money went into getting it, that’s the sunk cost fallacy.

If this reluctance stems from the ownership-induced sense of attachment, that’s the endowment effect speaking. If you associate with an object, getting rid of it is equivalent to purging a part of your soul. Not a good deal.

A time-tested sales technique offers customers a full refund if they aren’t satisfied with a product and return it within a month. The ability to use a product effectively for free for a month lowers customers’ defenses, and more people end up buying. If a product is halfway decent, however, relatively few customers would return it. What’s one of the reasons?

The endowment effect. Having become owners, customers no longer find the original price fair. Open / Comment

2020-10-18 11:01:02 Engineering serendipity

Artemy Lebedev’s (TG) design studio offers an online tool called “The Idea Matrix.” It collides random words to eventually produce a phrase leading to insights.

Similarly, designers of office spaces may choose a building layout to maximize the incidence of chance encounters and interactions between workers and to foster productive conversations and, ultimately, insights.

Now, allow me to digress for a minute. The main reason why I read books mostly on Kindle is because of how easy it is 1) to highlight sentences and paragraphs and 2) to export the highlights into a text file or a spreadsheet. This allows me to review daily a few dozen highlights pulled randomly from a database of ~20K entries, the crème de la crème of the hundreds of books in my library.

Reviewing important ideas just from one book is useful in and of itself—repetition is the mother of learning. But I’ve recently realized another value of this practice. It exposes you to potentially insightful cross-pollination across books.

One day, for example, a quote from Carol Dweck’s “Mindset”—which builds a case for the growth mindset (i.e., abilities aren’t fixed but can be developed)—appeared next to the following quote from Andrew Roberts’s biography of Napoleon:
— Napoleon was capable of compartmentalizing his life, so that one set of concerns never spilled over into another. ‘Different subjects and different affairs are arranged in my head as in a cupboard,’ he once said. ‘When I wish to interrupt one train of thought, I shut that drawer and open another. Do I wish to sleep? I simply close all the drawers, and there I am asleep.’

Attention residue notwithstanding, no need to be jealous of Napoleon’s innate ability to switch from commanding an artillery attack to drafting legislation. We, too, can mold our minds into a cupboard if we adopt the growth mindset.

Proactively expose yourself to as many ideas and their combinations. Open / Comment