Blog
A CEO’s Take on the Amazon Shareholder Letter
Andy Jassy recently published his annual letter to shareholders, recapping 2023 and laying out his vision for the future of the company. It’s hard to ignore the fact that Amazon.com (30 years old) is well positioned to pass Aramco (90 years old) and maybe even Walmart (62 years old) as the largest company in the world by revenue.
Here are the highlights from that letter and my take on them:
Building Primitives
First attempts are usually flawed, sometimes disastrously—the 80/20 rule applies to the effort. By building in primitives, those failed attempts leave behind tools instead of artifacts. As tools accumulate, flywheels begin to emerge almost automatically.
Using those tools, Amazon is able to course correct and even build new features and services it did not previously imagine. As Amazon Web Services (AWS) customers, we also get to build on these primitives in delightfully creative ways. We benefit from the primitives, the flywheels that emerge from AWS, and the flywheels we build for ourselves. At Rhythmic, our passion is thinking in primitives and using our creativity to find unique solutions to customer problems.
Being Closer to the Customer
In 2012, Jeff Bezos said in an interview with Charlie Rose that he wasn’t sure same day shipping was viable—but that they were still trying like hell anyway. Needless to say, it was viable, and even if it wasn’t, the pursuit of that goal permanently changed the supply chain and fulfillment industries.
Bezos also famously talked about being customer obsessed instead of competitor obsessed in that interview. Turning that notion inside out, we should also think about pushing things closer to the customer instead of pulling them closer to us. Keeping things close to the vest, nestled in a single region, in a single VPC, in a single account that we lovingly care for is the ultimate illusion of control.
Let go of the notion that regions are for increasing resilience. Stop thinking every app needs its own discrete database. Your data probably doesn’t need to all be in the same place even within a single service. You actually still can access, protect and operationalize your data even if you distribute it, and there are a litany of services that will make this easier than ever for you. Push data and services closer to your customers for them. You’ll be more resilient, more innovative, and more pleasing to your customers.
Being Efficient
AWS growing 10% despite leaning hard into customers optimizing their spend is the ultimate flex (and yes, AWS really did lean hard into this last year). They worked fearlessly in 2023 to help their customers spend less with them and still grew. Perhaps this puts a dent in the narrative that Peak AWS has come and gone. Know your business, question assumptions, and invest in efficiencies. In particular, look for opportunities to re-baseline on primitives and services that did not yet exist when you built what you currently have.
Cloud-native is an elusive pot of gold worthy of pursuit. There is no destination, but rather a journey to learn and reinvent. Experiment often in the name of efficiency. Never treat an infrastructure problem as solved—it is merely temporarily improved. Constantly revisit architectural patterns and always question assumptions.
Being All In on GenAI
Amazon dedicated over 15% of the shareholder letter to the potential of GenAI. As Andy astutely observes, while AWS is working hard on creating new primitives and services for GenAI, the real innovation is going to come from customers. Normally, seismic shifts in technology take decades to unfold. Not so long ago, early adopters of smartphones were accused of being nerds who didn’t properly value battery life and signal strength.
GenAI won’t take decades to unfold. It is universally available and inherently accessible. This shouldn’t be a surprise. After all, GenAI works like us, and we’ve been interacting with humans for millennia. It is effectively a REST API to control an army of compliant, inexpensive, and imminently capable humans. Why wouldn’t that change everything?
AWS is taking a lot of heat for being late to the party, but this is a shallow take. AWS took a different route, recognizing that it would never develop a superior or easier to use model. Instead, they positioned themselves as a platform for building AI solutions, with a model-agnostic interface and an incredible set of tools to build, refine, and operationalize GenAI ideas. No one has this breadth of capability. Outside of edge cases, there is little reason to build a GenAI solution anywhere but AWS.
Amazon’s success is not just about technology, timing, and market fit—it’s been about mindset. By embracing a culture of building primitives, staying close to customers, relentlessly pursuing efficiency, and boldly exploring transformative technologies like GenAI, they have positioned themselves for sustainable growth and innovation. There are many lessons we can learn and apply together with our customers.