Premature Optimization

What leads high IQ people to making poor decisions

Sometimes, when I walk around and enter a street, I often see from afar that the street is a dead end. I’ve often been right about this, but there have been some times when there would have been a small opening on to the other street, previously invisible, that I could see only as I neared the perceived dead end.

Of course, this post is not about walking well.

A lot of smart people are prone to assuming conclusions to situations before actually letting them unfold. Why?

On the surface, this seems like impatience¹. You want to get to an answer faster and hence apply some probabilistic thinking to get to a conclusion faster. But that’s not the only thing.

I think there are two reasons:

1. Lazy thinking

2. Incorrect thinking

Lazy Thinking

This is one of the biggest paradoxes I’ve realized growing up, that a lot of high IQ people are actually lazy thinkers. They think they know, without getting the required data or intuition, and choose to conclude things based on that. Want to see examples?

That friend who keeps talking about startup ideas, makes them in his head and then even talks himself out of it, all without talking to a single user.

That friend who has all the knowledge about a topic after reading everything about it, but hasn’t experienced even the slightest bit in practice.

The laziness here is basically a lack of effort to gather the required practical data / intuition to make an informed decision.

Sometimes, this laziness is also based on unwillingness to challenge previously held beliefs. Even high IQ people have a lot of biases and a lot of them are afraid of getting them broken. So they rationalise themselves out of a situation by thinking lazily.

Incorrect Thinking

The way smart people justify this is via probabilities. And because there is a lot of literature on how probabilistic thinking helps in life etc, this perpetuates this fallacy.

Let me explain.

First let’s take a situation where the odds are easily calculable.

Let’s say you are looking at an exam to give which has just 1% acceptance rate. By definition, your default chances of making it are 1%. And if you have some specific skill / experience advantage, it will be higher than 1% (but tough to point out exactly how much higher).

As you give mock tests with your peers, you will get enough data to understand where you stand and how close are you to acceptance. This is amazing.

Unfortunately, life is not as simple. But people often misapply probabilistic thinking in these domains.

E.g. Business

If you have an idea, what are the odds of that becoming successful? Well, you’d have heard from general chatter that 99% of startups fail, so may be 1%? But that’s just generic data, it takes nothing about your idea into account.

So you go into proxies that can give you some indication of the likelihood of that idea succeeding. If you dont know any better, you’d approach this like a consultant would e.g. look at if similar ideas have succeeded elsewhere and conclude based on that.

Again, this has zero feedback from real data i.e. your customers. And even if you talk to your customers, the data will be so fuzzy that you will only discover true value of your idea over time (and after some execution).

The other fallacy that people make here is assuming ergodicity.

Ergodicity is the property that the past is a good representation of the future.

You might look at history diligently and give a good thesis on why something will not work out. But often, you’ll see counter examples. E.g. Uber, Stripe, Tesla, Zepto etc.

Now, we of course have limited time and attention in life, and can’t test everything to their natural conclusions. So where do we actually apply probabilities and where not?

In static, linear domains, probabilistic thinking probably suits you well. E.g. You can reasonably predict your odds of completing a half marathon under 2 hours based on your 5k timings.

In non-linear domains, it fails. In places like art, business, etc, things seem impossible until they aren’t.

How do you think here then? This is detailed out in my post on Opportunism vs Conviction, but the answer is effectual thinking² + conviction.

Probabilistic thinking is basically based on a goal and then determining chances of achieving that goal. Since the concept of a goal itself is so fuzzy in non-linear domains, you’re better off not thinking goal first. Instead, you start with the resources you have and then use your imagination to create different paths and move on them. That way, things (and by extension, probabilities) unfold as you walk along that path.

So if it’s a dead end, you discover that, and if not, you’ll discover that also (and build conviction along the way), and that may lead to the 1% chance of greatness.

Another axis to apply in choosing the type of thinking is consequentiality. If you were to take a very high stakes decision, you’re better off not relying just on probabilities but if it’s a mostly low stakes decision, you can do 80-20 and move ahead. This ultimately comes down to judgement and personal assessment of situations.

So, the next time you’re relying on probabilistic thinking to close your decision, see if it’s because of some premature optimization and if it’s actually serving you well in that situation.

1. I think there’s a philosophical angle to this impatience as well —> we’re mostly impatient about things we don’t like a lot. And a lot of high IQ people are so goal oriented they don’t really know or stick to what they like. Hence, the impatience. Will unpack this in a future post.

2. Reading on effectual thinking or what makes entrepreneurs entrepreneurial.