Behavioral Architecture
TL;DR
- We think we make completely logical choices. But we’re biased
- Some popular biases: confirmation bias, endowment bias, availibility bias
- Understand what your biases are, and find people that have contrary opinions because they are going to balance that bias
- Logic is the beginning of wisdom, not the end
Full Version
We think we make completely logical choices. But actually, we are emotional humans, we’re not completely logical all the time. We’re… biased.
-
Confirmation bias: we tend to search for stuff to confirm our beliefs. Google doesn’t help with that cause it will show you results based on your bias.
-
Endowment bias: we place more value for things we own.
Your stuff > Other stuff -
Availibility bias: If you can think of an example quickly, then it is the most important one for you. Ease of recall is equivalent to importance. Deer Vs Shark, who is most dangerous? We directly think of the Shark ofc.
Micro services == Awesome,Spring > Java EE,Serverless > Docker > VM > Bare Metal -
Overconfidence: We all think we’re better than average drivers. Which comes from Halo effect, cognitive dissonance, bounded nationality and WYSIATI.
Tip: Understand what your biases are, and find people that have contrary opinions because they are going to balance that bias.
Introducing new technologies
Can start with Pilot projects, dojos hackathon. There is a risk to switching new technologies. It’s hard for big companies to move to new tech because they have a lot to lose, whereas easier for startup to go for innovation since less to lose.
Example of 401K plan: Make it easier to enroll for reitrement funds, or else people will not do it. You can apply this concept for tech as well: make it so simple for something to use an API or something to build something. It is the Nudge effect.
Fads and Bubbles: They will come and go. It applies for tech choices. There will be a cool tech framework today which will fade away, because anything has its limitation.
Logic is the beginning of wisdom, not the end
