The Attention Economy Matters for Developers, Too

What does that term even mean, the attention economy? I’ve heard it so many times in marketing and when dealing with things like Instagram or TikTok, but why do I need to care about that as a developer?

There’s actually a really important reason, and I think we’re all experiencing it. AI development, actually, if you want to be good at it, requires your attention.

This isn’t an attack on vibe coding or someone holding onto the past of reading all of the source code. There’s something very specific that’s happening here, and I want to explain it and why it matters for us as developers.

When we talk about the attention economy, this seems to be most often brought up when it comes to making you the product, or selling you something. But generally, that term really just indicates that someone’s attention is valuable. The more someone pays attention to something, the more valuable that person’s time is to the organization or the offering. If they’re paying attention, they’ll participate or they’ll buy the thing. If they’re paying attention, they’ll be activated to support your cause.

As a developer, I sort of laughed at this because I didn’t feel like I had this problem. I control my own attention. I get into flow state, and I solve problems, and I’m not distracted easily. In fact, I don’t really use a lot of social media, so why do I even care about this?

But then our automated AI agents have changed the game.

Not only are they generating tons of code, which, again, I’ll read (and sometimes I won’t.) But, they’re also making decisions for us, too. And that’s where the importance of attention actually comes in.

Once you start to trust that the code they’re generating is reasonably good, it becomes easier and easier just to keep saying, “Yes, go ahead, continue, I agree,” and not really paying attention to what the AI LLM response actually responds.

What I’ve been noticing is that the more I work with these tools, the less I want to read all of the responses. I tend to believe that, yes, it is just summarizing the code that I saw, that I did a quick review on, and therefore I don’t need to see anything else. But buried deep inside these responses are those organization/product decisions that we’ve allowed it to make.

Time and time again that the product is accurate to what the LLM decided it was going to make, but that wasn’t the product I asked for. Maybe I left some stuff out of my prompt or maybe the model responded to a temperature setting and decided to do something different. There’s all kinds of different reasons why the end product is not necessarily the thing that you envisioned.

So, just like how stakeholders in a business would have to grudgingly accept the development a programmer did, even though it didn’t match the full requirements, programmers are starting to do that now with our assistants, our agents, and our tooling.

And that’s not okay.

It was never okay that the business had to accept some subpar work just because it was more expensive or they didn’t describe the problem correctly. But, that was what happened (time constraints, budget constraints, etc). Now, though, it’s doubly not OK for programmers who usually have the time, the attention, the skill, and the ability to alter the output of AI agents to be exactly as we request to accept something that isn’t what we want.

And why is this happening? I think it’s attention.

Like I said, reading LLM responses, it gets easier and easier to just skim over them and say, “Yep, yep, yep.” Luckily, some of the LLMs have been responding with “Oh, I changed this a little bit different from what you requested.” But that’s not always the case, and sometimes it’s hard to catch, even as they’re describing what they did, what the actual requirements are and how they were used. You can’t always catch all of that in the UI either.

So, as we navigate this next direction as programmers, I think this attention economy is incredibly important. And it’s not just programmers; it’s everyone in business. You can work with these tools, and they will supercharge some of the work that you do, but you still need to pay attention. In fact, you have to pay attention even more now.

The attention economy is making developers into supervisors. A good manager or supervisor pays attention to its employees, and those are the most effective uses of the organization’s team. And that’s what we should be doing.