gm, it’s time to build


Where we’re at

Having worked in Silicon Valley for a little over 20 years now, I’ve seen my share of market cycles: from the dot com boom and bust, to the GFC, to whatever they’re calling this one. As they say: history doesn’t quite repeat itself, but it often rhymes.

After many years of easy funding and sky-high valuations, companies and people tend to get complacent (I’m guilty of this too). Large companies are only making incremental improvements which means they aren’t as willing to take risks. Startups are still being created but they tend to be a twist on what already exists “uber for this, airbnb for that”. There is always a bit of worry about when the bubble will pop, and many think they’ll be able to ride the wave until there are obvious signs. Realistically though, the systematic effects tend to be way more complex and subtle than people imagined.

What happens after a rough patch is when it really starts to get interesting: historically this has been when the greatest companies, products, and teams have been built. Complacency wanes, and there is a nervous tension in the air. This is surprisingly health and a strong motivating force for both founders and engineers alike: there is move back towards first principles thinking for what is being built.

Some more context..

The common denominator coming out of all of these periods has been that the level of abstraction upon which companies and products are built has continued to move upwards. You no longer really have to think about infrastructure — it is mostly already there. This is similar for a lot of foundational tooling, libraries, and systems.

As an example, AWS and GCP significantly lowered the bar for creating a startup. Prior to AWS, a company had to get pretty significant funding and build a lot of things from scratch to even have a product. It took years. These days, you can come up with an idea, spin up an environment in AWS or GCP, write some code, and then be serving traffic within a few weeks.

This trend will continue.. more on that below.

What’s going to matter over the next decade?

There are a couple areas that I think will be increasingly important:

  • Developer and Data tooling
  • Novel approaches to security and compliance
  • Collaboration tools
  • Operational efficiency
  • The ownership economy and web3

Developer and data tooling

With the proliferation of data and system complexity it is going to be important to have world class tooling in order to build effectively. The way engineers develop and build systems will change significantly: new abstraction layers and glue layers will be built. One thing that becomes important as you build further away from the bottom layers of the stack is observability: black boxes are bad and being able to understand what is happening with systems will become increasingly important.

The second area is automation and developer efficiency with the goal being to build software quickly and safely. The cost of bugs rises exponentially as they are found in later stages of development (or worse, production). Being able to have quick and iterations and meaningful feedback will really help development teams to level up their engineering efforts.

Next is data. We’re just going to have more and more of it — being able to operationalize it will be important. That means making sure the right people have access to it, and that they’re able to use it effectively. They’ll also need to be able to combine data from many sources and help themselves by building novel ways to interact with it. This represents quite a challenge from a tooling perspective and I think we’re currently just scratching the surface.

Security and compliance

With distributed and remote teams, the massive amounts of newly generated data, and developers building at lower levels of abstraction, the attack surface from a security and compliance perspective has widened considerably. I expect to see many new tools and systems built that try to tame the chaos here in ways that we haven’t seen before. Some general tenants are monitoring and automation but those have existed for quite a while; the new products will need to go further and have intelligence built in.

Collaboration tools

Whiteboards are no longer a thing. Development teams are either hybrid or fully remote (for the most part). The significantly changes the landscape for collaboration.

Teams are going to have to learn a new way of working (hint: it isn’t slack, and virtual whiteboards aren’t really going to cut it either). They will have to learn how to build and keep an organizational memory and learn to reduce the operational burden of collaboration. Teams will need to be more thoughtful about communication and will need the tools to make it efficient for team members to have access to as much context as possible.

I have some thoughts on what will be built here based on what I’ve seen work well in some environments, but I look forward to being amazed by the ingenuity of the teams building these tools.

Operational efficiency

I believe that operational efficiency was significantly hampered during the covid lockdowns but is currently getting better as people better understand the new ways of working. But, it’s not there yet, and probably isn’t even on par with what we had in early 2020. Automation will play a large part in this, as well as tools that augment decision making. I imagine this will be a core focus for many startups being built over the next few years.

Ownership economy and web3

After seeing what happens with large companies owning the majority of our data, I think people are starting to realize that they’d actually like to own it themselves. All of it.

While web3 has mostly been touted as a new financial system (yes, that’s important), what I really think will happen is a large push into the ownership economy where platforms allow users to own and manage their own data. The good news is that we’re no longer the product! The bad news is that we’re going to now have to find some way to pay for these services.

I don’t have a good sense of what will happen here, but there are some interesting things built in the realm of decentralized collaboration tools currently. It will be interesting to see how this manifests.

It’s time to build

There is so much we don’t know. There is no telling how long this recession will last or how it will go. What I do know is that I can already start to feel the nervous energy in the air and I know that some great things are going to be built. All we can really do is put our heads down and build something of value.

I’m incredibly excited to see what happens in the next 5–10 years, and I’m very fortunate to be able to invest in and help build it. LFG.