Weekly I/O #58

Decoy Effect, Low Social Optionality, Make AI Charming to Learn, Experience Valuable than Database, Be Understood after being Clear

Cheng-Wei Hu | 胡程維
5 min readMar 25, 2023

Weekly I/O is a project where I share my learning Input/Output. Every Sunday, I write an email newsletter with five things I discovered and learned that week.

Sign up here and let me share a curated list of learning Input/Output with you 🎉

The below is extracted from the email sent on Mar 18, 2023

1. Decoy Effect: When we are choosing between two options, introducing a third option can cause shifts in our preference between the first two items.

Article: The decoy effect: Why you make irrational choices every day (without even knowing it)

When we are choosing between two items, introducing a third and less attractive item (the decoy) can influence our perception of the original two choices. The decoys are asymmetrically dominated, meaning they are entirely worse than one option but only partially worse than the other. A famous example of the decoy effect was The Economist subscription page a few years ago. They offered three annual subscriptions at the following prices:

  1. Website-only subscription: $59
  2. Print-only subscription: $125
  3. Print+Website subscription: $125

The decoy was the second option (print-only) because it was completely inferior to the third option (print+website). MIT Professor Dan Ariely divided his students into two groups and conducted an experiment. When he hid the second option and only offered website-only and print+website options to the first group, 32% of people chose print+website, and 68% chose website-only. However, when the second option (decoy) is offered, the ratio of people who picked print+website increases to 84%, and the ratio of people who chose website-only drops to 16% (0% chose print-only).

Why the additional decoy shifts the preference between options? It’s a lot easier to choose when we compare similar things. Our brain can quickly evaluate the differences as opposed to when options are different.

Dan Ariely explained in his talk:

“The option that was useless was useless in the sense that nobody wanted it. But it wasn’t useless in the sense that it helped people figure out what they wanted. We actually don’t know our preferences that well, and because we don’t know our preferences that well, we’re susceptible to all the influences from external forces.”

2. Approach social interactions as if we are stuck on a desert island with the other person because low social optionality situations make genuine connections.

Podcast: Low Optionality and Adult Friendships

In this episode, Taimur introduced the concept of “social optionality,” which refers to the degree of freedom we have to switch from one person/group to another during social activities. For instance, a large party with 40 people has high social optionality, a small dinner gathering of 10 people has lower social optionality, and being stranded with only one person on a desert island has the lowest social optionality.

Having lower optionality in social situations changes how we view our interactions. When meeting someone new at a large party, we might constantly assess them like, “Could I be friends with this person?” because we know we can easily walk away and never see them again. However, it is less convenient to make such a switch at a smaller dinner gathering, and we may have to wait until the dinner is over. If we are stuck on an island, we may have to stay with the other person forever without the option to leave.

Because of this lack of optionality, our mindset on the desert island turns from “Could I be friends with this person?” to “How can I connect with this person?” In low optionality situations, people are encouraged to find ways to connect genuinely, as it is for the long haul. Therefore, Taimur states, “I sometimes feel like being stuck on a desert island with someone is the only way to interact with them truly sincerely.” And he concludes beautifully:

“Wanting to connect for no other reason than because you’re two human beings on the same floating rock — this is the intention with which I’d like to approach all social interactions. Low optionality situations make this significantly easier, and I’ve started prioritizing them when choosing how to spend my time.”

3. Rather than program an AI with what we want it to know, create ones capable of learning and have users teach them. Experience is the best teacher. Make AI charming so that users will be motivated to put effort into teaching them.

Book: Exhalation: Stories

This is from Ted Chiang’s novella The Lifecycle of Software Objects, which is included in the collection Exhalation: Stories.

The story follows Blue Gamma, an AI company that sells digital AI pets called digients. The company’s philosophy of AI design is that experience is the best teacher. Instead of trying to program an AI with what we want it to know, sell ones that can learn and have the customers teach them. To motivate customers to put effort into teaching the digients, the company ensures that every aspect of the AI pets is charming, including their personalities, which the Blue Gamma’s developers are working on.

Rereading this piece, which was published in 2010, makes me reflect on ChatGPT. Looking at ChatGPT, its interaction follows a similar philosophy, with shareability playing an essential role in incentivizing users to engage with it. By sharing screenshots of the conversation with ChatGPT on social media, we get more users playing with it, which helps improve the AI’s reinforcement learning.

We probably don’t have to wait too long to see AI pet startups pop up and adopt this same “philosophy”.

4. Every quality that made a person more valuable than a database was a product of experience.

Book: Exhalation: Stories

This is also from Ted Chiang’s novella The Lifecycle of Software Objects, included in the collection Exhalation: Stories. We can use this as one of the heuristics to evaluate whether a role or function is irreplaceable by AI.

This also highlights the difference between Google Search and ChatGPT in terms of answering fact-based queries. While ChatGPT provides succinct responses, Google Search provides the context (links to websites). Although navigating the websites of reference might not necessarily make Google Search better, it offers the experience of exploration, which makes it different.

This also reminds me of how Instagram is stealing search traffic from Google among Gen Z users. When looking for recommendations on what to buy or where to go, they turn to social media influencers or friends. The journey to finding context differentiates people’s perception of search results.

5. “It is a fault to wish to be understood before we have made ourselves clear to ourselves.” — Simone Weil


That’s it. Thanks for reading and hope you enjoy it. If you would like to receive the content every Sunday, sign up below 😎



Cheng-Wei Hu | 胡程維

I've left Medium. Subscribe to chengweihu.com for new article and newsletter! 從 Medium 搬家啦,新的文章還有電子報可以在 chengweihu.com 訂閱