Deep.

The Ceiling Is Rising. The Floor Isn’t.

I’ve been thinking a lot about what technology has actually done to the economy—not in theory, but in terms of how people experience it day to day.

On one hand, the upside is very real. I’ve seen people completely change their trajectory because they were able to access the right skills and plug into the right systems. Careers that didn’t exist a generation ago now offer real financial independence, global mobility, and a level of leverage that used to be out of reach for most people. Technology has made that possible, and it’s hard to argue that this hasn’t been a net positive.

But that’s only part of the picture.

The more I look at it, the more it feels like we’ve raised the ceiling much faster than we’ve raised the floor. If you’re already in a position to take advantage of what the technology economy offers, things can move very quickly. But if you’re not, it’s not clear that anything has fundamentally changed. Access is still uneven. Opportunity is still gated. And in many cases, the gap between those two realities is getting wider, not smaller.

What’s been bothering me more recently is how we talk about progress in this context. There’s a tendency to assume that building more advanced technology is the same as creating broader economic value. It’s not. And yet, if you look at where time, attention, and capital are going, we’re heavily skewed toward things that are new, ambitious, and often unproven.

I understand why. Frontier innovation is exciting. It signals momentum. It creates the sense that we’re pushing boundaries. But at the same time, there are very real, very solvable problems around access—education, pathways into the workforce, economic inclusion—that don’t get nearly the same level of focus.

And I don’t think it’s because those problems are harder. It’s because they’re less interesting to build around.

That’s the tension I keep coming back to. Technology is incredibly effective at accelerating people who are already on a certain path. It’s much less effective at creating that path in the first place. And if we’re honest about it, a lot of what we’re building today continues to reinforce that dynamic.

None of this makes me pessimistic about technology. But it does make me think we need to be more deliberate about what we choose to prioritize. Because if we continue to define progress only by what’s new and cutting-edge, we’re going to keep missing the more important question—who is this actually working for?

And if we don’t start answering that more directly, we shouldn’t be surprised when the benefits of all this progress continue to concentrate in the same places, with the same people, over and over again.

What DevOps Taught Me About AI

I've spent my entire career working for software startups, especially those in San Francisco. Many of these startups were really just a product of a bygone era of innovation. I moved to San Francisco at 25 after working in international affairs for a year, and landed right in the middle of the SaaS movement. That wave created hundreds of copycat companies and, in many cases, very little real change.

While I was pretty disappointed by the overall ethos of “Silicon Valley” when I first arrived, I did eventually come across something that stuck with me: DevOps. What I found interesting was that it wasn’t just a set of tools or practices—it was a way of thinking about how systems should work. It was rooted in open source, shaped by a strong community, and grounded in best practices that had actually been tested over time.

At its core, DevOps is about reducing friction. It focuses on identifying the parts of the development process that are repetitive, manual, and error-prone—and then systematically removing that burden through automation. Things like build pipelines, testing, and deployments used to require constant coordination. When done well, they now largely take care of themselves.

That idea always felt intuitive to me. Humans naturally optimize the things they do repeatedly. What feels difficult at first eventually becomes second nature, and then eventually something you don’t even think about. DevOps just scales that instinct across teams. It turns individual efficiency into organizational efficiency.

But the real value isn’t just in saving time—it’s in what that time gets used for. When engineers aren’t stuck managing fragile processes or doing the same manual work over and over again, they can focus on higher-leverage problems. Better systems, better products, better decisions. The goal isn’t just speed—it’s depth.

Lately, I’ve been thinking about how closely this maps to what we’re seeing with AI.

A lot of the conversation around AI right now is about automation, but that framing feels incomplete. What’s actually happening is closer to what DevOps enabled years ago—we’re starting to remove entire layers of cognitive and operational overhead. With agent-based systems especially, tasks that used to require constant human involvement are becoming more self-directed.

But if DevOps taught me anything, it’s that automation doesn’t just eliminate work—it changes it. It raises expectations. Once the baseline becomes easier, the standard for what matters moves up. You’re no longer judged on whether something works—you’re judged on how well it works, how thoughtfully it’s designed, how it scales.

I think AI will follow the same path. It will reduce a lot of the manual effort that fills up our time today. But it will also shift where we’re expected to contribute. More judgment, more creativity, more responsibility for outcomes rather than inputs.

Which is why I don’t think the most important question is how much work AI can take over. The more important question is what we choose to do with the space it creates.

Because we’ve seen this pattern before. DevOps didn’t just make teams faster—it gave them the ability to work on better problems. AI has the potential to do the same at a much larger scale. But that outcome isn’t guaranteed. If we treat it purely as a way to increase output, we’ll likely end up in the same place—just moving faster without really changing what we’re building or why.

And that feels like a missed opportunity.