Sunday Snapshots (26th April, 2020)
Complexity Economics, Sonos Radio, Conversational entropy, Weird productivity tips, and Jack Ma's Alibaba
Greetings from Evanston!
I hope you and your family are safe.
Thank you for sharing Snapshots last week. We now have more than 500 readers! Over the course of writing Snapshots, I’ve realized that I’m not competing with other newsletters. I’m competing with your favorite influencer’s Instagram stories, your friend group’s juicy messages, and the latest fortune cookie wisdom on Twitter. Billions of dollars are spent on making these platform algorithmically perfect to maximize engagement. These are tough enemies to fight in the battle for attention. I’m thankful that at least 500 of you think that I’m good enough to defeat them at least once a week for a few minutes.
With that out of the way, let’s get into this week’s Snapshots in which I want to talk about:
Why does traditional economic models fail to explain our world and what are some better models to use
Why Sonos’ latest move into subscription models might be the latest case of “too little, too late”
How Gmail predicts your next words while you’re writing an email
The most absurd productivity trick that works
Alibaba and our changing relationship with China
Book of the week
Every year, I read 1-3 books that introduce truly new ways of thinking. The Origin of Wealth by Eric Beinhocker is my first such book for 2020. The premise of this book is simple – traditional economic theory is outdated and useless. We need a new framework of thinking about how the economy works. But the solution is anything but simple. In fact, we need a series of frameworks because no single one can explain everything. It’s a bit like physics in that there is no unified grand theory. There is stuff that works at the small scale (quantum mechanics) and stuff that works at the large scale (general relativity.) As Beinhocker takes us from a world of Neo-classical economics to a world of complex systems, we travel through the paradigms of dynamics, agent-based modeling, network theory, search algorithms, and behavioral psychology. The scope of this work is enormous.
Here were my takeaways from the first half of the book:
Complex adaptive systems: Beinhocker argues that an economy is a complex adaptive system. These systems are comprised of agents. These agents have goals. They achieve these goals through rules of thumb or heuristics. They improve upon these heuristics through feedback and learning.
For example, we can think of the pre-modern world in which our ancestors lived in as the system. Their goal was to survive and reproduce. A rule of thumb they might have used was that exposure to animals who make a lot of noise (like a lion or an elephant) reduces the likelihood of survive. A potential feedback that they could have gotten was that some animals like cats were less likely to be dangerous to their survival. So, then the heuristic gets updated to only animals which are large AND make a lot of noise are dangerous. Over time, this rule of thumb gets refined in service of the goal to survive and reproduce.
In the same way, the economy is a complex adaptive system. More on the implications of this next week, but the idea itself is an important departure from the “equilibrium” framework that is taught in Economics classes around the world. To understand why the equilibrium framework is favored, you have to understand that the math for a complex adaptive system is not just complicated – it is practically impossible to solve. Therefore, economists in the 19th and 20th century made a series of simplifying assumptions to resolve this math. However, at some point, introduction of assumptions that cannot be explained (like perfectly rational agents) make your model not just wrong, but useless. This is exactly what happened in the case of economics and now we’re stuck with the equilibrium theory.
The possibility-complexity tradeoff: Networks are critical to understanding complex adaptive systems since agents act not just based on their own preferences, but also take their cues from the preferences of other agents. Monkey see, monkey do. But networks introduce even more complexity and coordination problems into the picture.
For example, think about a network of employees in your organization. If there are 4 employees, there are 6 potential collaborations of 2 employees. If we increase the number of employees by 1 to 5 (a 25% increase), the number of potential collaborations goes by 4 to 10 (a 67% increase). There is a disproportionate increase in the number of potential collaborations that you have to coordinate.
This brings us to one of the central tradeoffs of complex systems. When you have more agents, you have more possibilities – you can do more with 5 employees than with 4. However, this increase in possibilities makes the coordination problem more difficult and it is a fundamental feature of complex systems as we don’t exist in isolation. We exist as part of networks. As we find policy solutions for fixing various parts of the economy, they must be scalable within the framework of this tradeoff.
Trade, Wealth, and Inequality: In 1995, researchers at the Brookings Institute built a simulation called Sugarscape. This simulated world had agents whose goal was to simply survive across a landscape where the distribution of the means of survival – sugar and spice – was random and uneven. They were allowed to trade in case they had more of one commodity than they needed and needed more of the other one to survive. These simple rules led to fascinating results.
First, when the simulation was ran with only sugar and no trade, results were expected – entities in more resource-rich areas lived and those in resource-poor areas died. But when when spice and trade were introduced, overall survivor rates and surplus commodities (or “wealth”) went up. But with this increase in the average wealth came increased inequality.
Various permutations of the game showed that this inequality was not due to any particular factor that could be fine-tuned to remove or reduce it, but simply an emergent property of the system. This has massive policy and political implications that I’ll talk about next week.
There is much more to come on The Origins of Wealth next week, so if you want to follow along, definitely check it out.
Business news of the week
Sonos Radio: Too Little, Too Late
This week, Sonos launched a radio service called Sonos Radio. In my opinion, it’s too little, too late for what the company wants it to achieve.
First, a bit of background on the company. Founded in 2002, it successfully captured the premium mediocre speaker market – not quite good enough for audiophiles, but much better than your run-of-the-mill speakers in an attractive package. Suburban homes with high household incomes flocked to buy these speakers to build home audio systems that signaled sophistication. In short, Sonos was in the differentiated hardware business.
But the introduction of smart speakers has left the company struggling to find a foothold. Most companies heavily discount their own smart speakers because they make up most of the value through their voice assistant services. So, in comparison, Sonos looks much more expensive.
Sure, it has better sound quality. But this is difficult for the average ear to pick up on. Amazon/Google have significantly more resources to deploy in terms of marketing and R&D for the voice assistants. These two players have open-sourced their assistants, so you can use both on a Sonos Speaker. But Sonos doesn’t make any money or get any accumulating competitive advantage from the services provided by these voice assistants. The only differentiated hardware company that’s done well is Apple. But Apple had its own ecosystems to lean on and and since it operates from a position of ecosystem lock-in, Sonos speakers don’t even have access to Siri.
So the company is a shell of its former self; commodified by its competitors and trying to survive in the era of services with a hardware-dominated business model.
Enter Sonos Radio.
It’s a free service that gives users access to genre stations (essentially ever-changing playlists) and local radio. They want to move to a subscription model by leveraging listener behavior and developing an ad platform. The service could have a free tier with ads and a paid one without ads. That just seems too little, too late. Playlists are commodities and most – if not all – differentiation has been captured by Spotify and Apple Music. Collaborations with upcoming artists could be impactful, but that’s a bandage solution to a problem that requires surgery.
So is Sonos doomed to be competed away at the whims of Google and Amazon that own the core part of the user experience? Or is there light at the end of the tunnel?
Here are two things that could be powerful:
A serious integration or even a merger of Spotify and Sonos, with preferred hardware status given to Sonos systems. This means better integration of Spotify and exclusive features uniquely enabled by the hardware on Sonos speakers. One of my core gripes with Apple is how superior of a product Spotify is, but given that it isn’t integrated into the Apple ecosystem, Apple Music is a better experience on the iOS ecosystem. If I were bringing an anti-trust argument against Apple, this particular case study would be my exhibit number one. This Spotify-Sonos partnerships could bring together software and hardware to build a worthy competitor to Apple.
Expansion into new product lines could be promising as well. An adjacent market with lots of competition like earphones or headphones might be particularly attractive as there are no obvious choices for the customer and they are more likely to look for and try new alternatives.
There are ways to save a sinking ship, but you have to embrace new lines of thinking. We’ll see if Sonos is able to do that and get to shore.
Concept of the week
How’s it going?
How are you dealing with the quarantine?
Hope you and your family are safe.
I didn’t just give you a starter texting kit for modern dating apps. These are examples of sentences with low entropies.
Conversational entropy can be defined by how predictable the next word is given a set of previous words. So if your sentences are predictable like the ones above, they have low conversational entropy. This feature of language – codified through mathematical frameworks like Markov dependence – is what allows Gmail or your iPhone keyboard to suggest the next word(s).
Here’s another way to think about this concept. A one-way exchange of communication with high conversational entropy is an interview. A good conversation is a two-way exchange of communication with high entropy.
May your conversations this week have high entropy.
Random corner of the week
This week’s random corner is truly random. It’s even a bit embarrassing, so I hope you appreciate the trust fall I’m taking with you.
It stems from the fact that it is tough to be productive amidst a global pandemic. New times require adoption of new systems and tactics. And my most effective tactic over the last month has been playing this video of Matthew McConaughey watching rain and drinking coffee:
When I want to start a block of work, I put this on and don’t stop working until the sound stops. The rain noise and McConaughey’s quips like ”Infinity combinations of matter and energy, with infinity possibilities. And they pick Ben Affleck to play Batman” keep the distractible part of my brain engaged and allow me to focus on the work to be done.
Embarrassing, I know. But very effective.
If you have an embarrassing productivity tip, please let me know by replying to this email.
Movie of the week
Crocodile in the Yangtze
Crises don’t corrode relationships, they simply reveal. As the post-COVID world order emerges, I doubt that the West’s relationship with China will remain the same. An important part of this is how the Chinese Internet will interact (or not interact) with the rest of the world’s internet.
Crocodile in the Yangtze is the origin story of the Chinese Internet. It is also the origin story of the pioneer of the Chinese Internet – Jack Ma of Alibaba. Created through a western perspective, it chronicles Alibaba’s rise from a dorm apartment with 17 original engineers to a half trillion dollar company at the head of the Chinese internet infrastructure.
I thought there were important lessons to be learned from the movie. This included how the idea of Chinese exceptionalism plays out in the tactics of Chinese companies, the ability of a leader to call a shot and make it, and how Davids can beat Goliaths.
That wraps up this week’s newsletter. If you want to discuss any of the ideas mentioned above or have any books/papers/links you think would be interesting to share on a future edition of Sunday Snapshots, please reach out to me by replying to this email or sending me a direct message on Twitter at @sidharthajha.
Until next Sunday,