Before desktop publishing, creating a magazine page could involve a graphic designer, a typesetter, and a paste-up artist.
Then one person could do all of this using software on their computer.
Technology changes how much work fits inside one role. I think AI is accelerating this again.
Here’s how I’m thinking about startup teams, whether you’re building one or figuring out what to learn next.
1. Roles are getting smashed together.
As technology makes work easier, the expected scope increases.
But it shouldn’t feel like more work. Through automation and better tools, you’re able to own more of the outcome.
We’re seeing “design engineer” pop off. More designers can now implement what they design, especially with AI.
I’ve even interviewed designers who do high-fidelity designs straight in code. For them, it feels easier. They can build the interaction, try it, and change it themselves.
Their scope has expanded. The handoffs have shrunk.
I think we’ll see more of this.
2. What’s hard is changing.
Before spreadsheets, updating a financial forecast meant laborious recalculation. Speed and accuracy mattered.
Spreadsheets automated that. More of the value shifted toward choosing sensible assumptions, testing scenarios, and interpreting results.
I think something similar is happening with code.
Getting a first version working is becoming easier. Deciding what to build, making it intuitive, and knowing when it’s actually good remain hard.
Generating code is getting cheaper. Deep infrastructure, security, reliability, and difficult technical problems still need expertise.
More of the value sits in knowing how to architect a good solution—and checking that it actually works.
So I’m leaning toward smaller, design-heavy teams. Broader roles. Agents handling more implementation. Deep engineering where we need it.
The question I keep asking: as the tools make your current work easier, what can you take responsibility for next?
Until next time,
Ajay
