Solve Big Problems
Premature optimization is the root of all evil
— Donald Knuth
Every time I board a plane, I'm annoyed by how long it takes. One person stops to put their bag overhead. Everyone waits for them to finish.
The mom with the three little kids. Yes, they're cute, but why did it take them 3 minutes to sit down?
Yesterday, I boarded my plane hoping to see the above youtube video play out in practice. That video shows the "optimal" way to board a plane.
Instead, I watched a lady walk all the way to the back of the plane, realize her seat wasn't there, and then walk all the way back up to the front of the plane. She held up every single person in line for 3 minutes.
I stood there, a little annoyed, a little amused, but more so thinking about how much better things would be if airlines implemented the optimal boarding protocol.
But with 3 minutes to process it I realized something: it doesn't matter.
Not just because those 3 wasted minutes don't matter. But because those 3 minutes weren't actually wasted.
Does a plane boarding faster help the plane take off sooner? Not really.
Planes take off when air traffic control says they can take off. In the case of my flight yesterday, that was 30 minutes after everyone sat down.
Boarding is not the bottleneck. Air traffic control is the bottleneck.
It's intuitive to understand that speeding up a non-bottleneck doesn't speed up the entire process, but when you're stuck in a local environment, like the 15 minutes it takes to board a plane, it's pretty easy to spot areas to improve.
If only we could shave those 15 minutes down to 10 minutes then everyone would be to their destination sooner...
This type of thinking produces a variety of problems, including:
- armchair quarterbacking
- people optimizing at local minima
- complexity
Armchair Quarterbacks
I was an armchair quarterback. These damn airlines are so stupid. Why can't they make onboarding more efficient? They can. It just doesn't matter.
From the outside, you'll always seem stupid for not working to optimize everything.
People optimizing at local minima
I'm sure there's someone (probably a whole team) at Delta or American whose job is to optimize onboarding. Those people have a dashboard to tell them the average boarding time for all flights. If that number goes down, they win. If that number goes up, they lose. But what does that number do for Delta's actual goal of making money? Probably nothing.
People waste resources optimizing at local minima.
Complexity
The onboarding team thinks they need to reinvent the boarding group system. The marketing team is worried that the new system will mess with their perks. They compromise. The new system makes no sense.
Every optimization generates complexity. Be careful where you spend complexity.
The Solution
Invention, not automation, will be the greatest contribution of superintelligence.
— Mark Zuckerberg
Work on big problems. Don't optimize at local minima.
Why are all these people automating airline operations anyways? If fast transportation is the goal, why don't they work to invent teleportation?
An obviously naive example, but that's why I'm building Denta Insurance.
There's no need to spend time on insurance operations when Denta immediately teleports the money straight into the practice's bank.
Nothing like this currently exists. Denta is not optimizing the current dental insurance infrastructure. Denta is building it from the ground up.
Solve big problems. This is tenet #1 of Denta.
Insurance at Lightspeed | dentainsurance.com