How AI Can Help You Take Your Career Beyond What You Imagined

Felix Rieseberg, an engineering manager on the Claude team, recently wrote an essay called "How to be more ambitious: A tutorial for humans with models." His thoughts on ambition in the AI era really resonated with us.

He opens by explaining that, working on the Claude team, he sees people do things with Claude almost every day that he had never considered, assumed were impossible, or simply hadn't had the curiosity or time to try. Whenever he sees someone achieve something remarkable with AI, he finds himself thinking that he hasn't been ambitious enough.

We've had much the same experience. At ExplainThis, we closely follow the latest developments and approaches in AI applications. Along the way, we often come across work that leaves us wondering whether we're being too conservative with AI. Could we aim higher, go deeper, or think more broadly?

In this article, we share our reflections on his essay and some concrete ways engineers can put these ideas into practice to open up new possibilities in their careers.

Three Ways to Expand What's Possible in Your Career

In his essay, Rieseberg describes four dimensions he considers when working with AI: height, depth, breadth, and speed. These help him push his work beyond what he originally imagined.

He develops the first three in particular. Let's look at each one and explore how engineers can apply them to achieve more than they initially expected.

Height

The first dimension is height. Rieseberg suggests asking, "What do I want to happen next? What's the real purpose behind this goal?" For example, when working with AI, you could ask it to make sure a bug never happens again, taking the task further than fixing each bug as it appears.

This echoes the difference between a junior and a senior engineering mindset. Someone with a junior mindset tends to focus on solving the immediate problem. Someone with a senior mindset steps back and asks, "Why did this happen? How can we prevent it from happening again?"

Imagine a production incident: after deploying a new version, an engineer discovers that the website's login feature has stopped working. With a junior mindset, they might identify the faulty code, fix it, redeploy, and stop once login works again.

An engineer with a senior mindset would ask more questions: "Why didn't testing catch this? Why didn't we notice immediately after deployment? If it happens again, can we restore service faster?" Alongside the code fix, they might add end-to-end tests for the login flow, health checks after deployment, and an automatic rollback mechanism for failed deployments. That way, even if a similar error occurs, it has less impact on users.

The same thinking applies to AI. Limiting our expectations to building individual features or fixing isolated problems leaves a lot of potential unexplored. For each task, try taking your thinking and goals one level higher. Ask AI to address the underlying causes so the problem doesn't recur.

Depth

The second dimension is depth. Rieseberg's view is that we should ask models to use the tools, techniques, and knowledge of the best experts in a field, then keep going until the results could teach those experts something new.

He offers an example that will feel familiar to full-stack engineers. If a website takes 10 seconds to load, many engineers might aim to get it down to 5 seconds. Cutting load time in half is already a challenging goal. From the perspective of depth, though, Rieseberg argues that we should aim for 200 milliseconds.

For most full-stack engineers, the first reaction to going from 10 seconds to 200 milliseconds is probably, "That's impossible." Rieseberg argues that working with AI calls for precisely this kind of ambition: set the goal, tell the model it's entirely reasonable, and ask it to use expert tools and methods to work toward it in depth.

He also recommends asking the model questions such as: "What might be missing from this goal compared with what the very best experts in this field would aim for? What methods or tools would they use that we're overlooking?" Prompts like these can encourage more thorough problem-solving and better results. They also give us a chance to learn how leading experts approach their work and build our own skills along the way. That's a practice we'd encourage you to try.

Breadth

The third dimension is breadth: thinking about your work through the perspectives of different disciplines. Rieseberg suggests a few simple changes to help with this.

First, turn the task from a noun into a verb. For example, when defining the output of an incident review, go beyond asking for "a postmortem report." Describe the outcome you want: "Help a group of very busy people understand and trust an account of cause and effect within 30 minutes, then make a decision while some information remains uncertain." Framing the task this way may lead AI to present the work in a form you hadn't imagined.

Second, assemble your own team of experts from different disciplines. When developing a feature, a software engineer could ask AI to act as a red team and challenge the technical design, or take the perspective of a security expert to help assess vulnerabilities. These different viewpoints can make the work more thorough.

Support ExplainThis

If you found this content valuable, please consider supporting our work with a one-time donation of whatever amount feels right to you through this Buy Me a Coffee page.

Creating in-depth technical content takes significant time. Your support helps us continue producing high-quality educational content accessible to everyone.